[{"data":1,"prerenderedAt":2915},["ShallowReactive",2],{"blog-posts-page-4":3},{"posts":4,"total":2914},[5,695,1163,1684,2245],{"id":6,"title":7,"author":8,"body":9,"coverImg":681,"description":682,"extension":683,"head":8,"meta":684,"navigation":685,"path":686,"publishedAt":687,"schemaOrg":8,"seo":688,"sitemap":689,"stem":693,"updatedAt":687,"__hash__":694},"blog\u002Fblog\u002F2026\u002F10\u002Fai-stock-analysis-tools-compared-what-each-one-reads.md","AI Stock Analysis Tools Compared: What Each One Reads",null,{"type":10,"value":11,"toc":662},"minimark",[12,16,19,22,27,30,119,122,126,129,268,271,275,278,290,297,300,350,353,360,364,373,376,379,385,391,395,398,401,406,409,413,416,419,424,429,433,436,439,447,452,457,461,464,467,472,477,481,484,487,492,495,499,502,505,519,523,526,534,538,546,549,568,571,575,584,593,599,603,606,639,643,646,649,652,656,659],[13,14,15],"p",{},"Seven tabs open. Seven AI stock analysis tools, and every one of them says its AI is the clever one. So which is best?",[13,17,18],{},"Wrong question, it turns out. These tools read completely different things, from hundreds of indicators across the whole market to a single photo of a chart. They hand back completely different things too, from a score of 1 to 10 to a paragraph in plain English. Pick by the question you bring and the choice gets easy.",[13,20,21],{},"Below is the verdict first, then what each tool reads, what it hands back and the catch in each. One of the seven, Primo Trade, is our own product. It gets the same treatment, gaps included.",[23,24,26],"h2",{"id":25},"the-best-ai-for-stock-analysis-depends-on-your-question","The Best AI for Stock Analysis Depends on Your Question",[13,28,29],{},"Find your question in the left column. That is most of the decision.",[31,32,33,49],"table",{},[34,35,36],"thead",{},[37,38,39,43,46],"tr",{},[40,41,42],"th",{},"Your question",[40,44,45],{},"Start with",[40,47,48],{},"Because it",[50,51,52,64,75,86,97,108],"tbody",{},[37,53,54,58,61],{},[55,56,57],"td",{},"\"Which of thousands of stocks should I even look at?\"",[55,59,60],{},"Danelfin or Kavout",[55,62,63],{},"Ranks the whole market by a score",[37,65,66,69,72],{},[55,67,68],{},"\"Where is this chart pattern showing up?\"",[55,70,71],{},"Tickeron or TrendSpider",[55,73,74],{},"Searches charts across many symbols",[37,76,77,80,83],{},[55,78,79],{},"\"What is moving in the next hour?\"",[55,81,82],{},"Trade Ideas",[55,84,85],{},"Scans live during the session",[37,87,88,91,94],{},[55,89,90],{},"\"Can I get a signal feed on my phone?\"",[55,92,93],{},"Prospero.ai",[55,95,96],{},"Is a phone app built around signals",[37,98,99,102,105],{},[55,100,101],{},"\"What is going on with this one stock or chart?\"",[55,103,104],{},"Primo Trade",[55,106,107],{},"Reads one ticker or one chart photo and explains it",[37,109,110,113,116],{},[55,111,112],{},"\"Can I just ask anything, my way?\"",[55,114,115],{},"A general chatbot",[55,117,118],{},"Takes any question, but you design it and check it",[13,120,121],{},"Plenty of people use two of these: a ranker or scanner to find candidates, then a single-stock read to look closer before deciding for themselves.",[23,123,125],{"id":124},"ai-stock-analysis-tools-at-a-glance","AI Stock Analysis Tools at a Glance",[13,127,128],{},"Every detail below comes from each tool's own pages, checked from the United States on October 8, 2026, so treat it as dated to that day. Prices are left out; they change too often.",[31,130,131,150],{},[34,132,133],{},[37,134,135,138,141,144,147],{},[40,136,137],{},"Tool",[40,139,140],{},"What the AI reads",[40,142,143],{},"What you get back",[40,145,146],{},"Horizon",[40,148,149],{},"The catch",[50,151,152,169,186,203,220,236,252],{},[37,153,154,157,160,163,166],{},[55,155,156],{},"Danelfin",[55,158,159],{},"900+ technical, fundamental and sentiment indicators",[55,161,162],{},"AI Score 1 to 10, four subscores",[55,164,165],{},"3 months",[55,167,168],{},"Probabilities, not certainties, in its own words",[37,170,171,174,177,180,183],{},[55,172,173],{},"Kavout",[55,175,176],{},"Fundamental, technical and sentiment data on 9,000+ US stocks",[55,178,179],{},"Kai Score, Stock Rank, Technical Rating, strategy lists",[55,181,182],{},"Not stated (rankings refresh daily)",[55,184,185],{},"The picker sits behind the Pro tier",[37,187,188,191,194,197,200],{},[55,189,190],{},"TrendSpider",[55,192,193],{},"Your charts, market data, fundamentals, filings, news",[55,195,196],{},"Pattern recognition, an AI assistant, backtests, bots",[55,198,199],{},"Whatever you set",[55,201,202],{},"A lot of platform for one question",[37,204,205,208,211,214,217],{},[55,206,207],{},"Tickeron",[55,209,210],{},"Price patterns and trends",[55,212,213],{},"Pattern matches with confidence levels, signals, robots",[55,215,216],{},"Intraday to swing",[55,218,219],{},"Confidence describes past setups",[37,221,222,224,227,230,233],{},[55,223,82],{},[55,225,226],{},"Live market data",[55,228,229],{},"Holly buy and sell signals, scans",[55,231,232],{},"Intraday",[55,234,235],{},"The homepage examples are a highlight reel",[37,237,238,240,243,246,249],{},[55,239,93],{},[55,241,242],{},"Institutional data and market movements",[55,244,245],{},"Key signals, cumulative scores",[55,247,248],{},"Short and long term",[55,250,251],{},"Its fine print says the results are simulated",[37,253,254,256,259,262,265],{},[55,255,104],{},[55,257,258],{},"One ticker's prices, news and financials, or one chart photo",[55,260,261],{},"Signal, Primo Score, strategy idea, Key Insights",[55,263,264],{},"6 to 12 months for the Forecast Value",[55,266,267],{},"One stock at a time, no scans or alerts",[13,269,270],{},"Primo Trade gives AI-generated analysis for information and education. It is not financial advice. Markets carry risk; do your own research before you trade.",[23,272,274],{"id":273},"one-monday-three-honest-answers","One Monday, Three Honest Answers",[13,276,277],{},"Before the tools, one dated example of why the horizon column matters more than any brand name.",[13,279,280,281,289],{},"Monday, January 27, 2025. Nvidia closed at $118.42, down from $142.62 the Friday before, a drop of about 17% in one session, according to Yahoo Finance's daily ",[282,283,288],"a",{"href":284,"target":285,"rel":286},"https:\u002F\u002Ffinance.yahoo.com\u002Fquote\u002FNVDA\u002Fhistory\u002F?period1=1737676800&period2=1754006400","_blank",[287],"noopener","price history",". Reuters tied the sell-off to a low-cost AI model from the Chinese startup DeepSeek and called it a record one-day loss in market value for a Wall Street stock.",[291,292],"img",{"src":293,"alt":294,"width":295,"height":296},"https:\u002F\u002Fmedia.primotrade.app\u002Fblog\u002Fai-stock-analysis-tools-compared-what-each-one-reads\u002Flong-red-candle-macro-panel.jpg","Stylised chart art of one long red candlestick in sharp focus among smaller green and amber candles.",1067,600,[13,298,299],{},"Here is what followed, read long after the fact from the same Yahoo Finance data:",[31,301,302,315],{},[34,303,304],{},[37,305,306,309,312],{},[40,307,308],{},"Checkpoint after the January 27 close",[40,310,311],{},"Close",[40,313,314],{},"Versus January 27",[50,316,317,328,339],{},[37,318,319,322,325],{},[55,320,321],{},"Next day, January 28, 2025",[55,323,324],{},"$128.99",[55,326,327],{},"+8.9%",[37,329,330,333,336],{},[55,331,332],{},"Three months, April 28, 2025",[55,334,335],{},"$108.73",[55,337,338],{},"-8.2%",[37,340,341,344,347],{},[55,342,343],{},"Six months, July 28, 2025",[55,345,346],{},"$176.75",[55,348,349],{},"+49.3%",[13,351,352],{},"A same-day signal, a three-month score and a 6 to 12 month estimate were answering three different questions that Monday. Each could look wrong at one checkpoint and fine at the next. Nothing on the chart that day said which path came next, and no model could have known it either.",[13,354,355,359],{},[356,357,358],"strong",{},"So the first thing to match is not the AI but the horizon."," A tool built for the next hour and a tool built for the next quarter can both be honest and still disagree.",[23,361,363],{"id":362},"danelfin-one-number-per-stock","Danelfin: One Number per Stock",[13,365,366,367,372],{},"Danelfin boils each stock down to a single AI Score from 1 to 10. According to its ",[282,368,371],{"href":369,"target":285,"rel":370},"https:\u002F\u002Fdanelfin.com\u002Fhow-it-works",[287],"own explanation",", a higher score means a higher probability that the stock outperforms the market over the next three months. For US stocks and ETFs the benchmark is the S&P 500 total return index; for European stocks it is the STOXX Europe 600.",[13,374,375],{},"The inputs run long: 10,000 features per stock per day, built from more than 600 technical, 150 fundamental and 150 sentiment indicators. It covers US stocks, Europe's main markets and US-listed ETFs. Beside the headline score sit Fundamental, Technical, Sentiment and Low Risk subscores, and an explanation window shows which features pushed the score up (green) or down (red).",[13,377,378],{},"That window is the good bit. You can see that a high score leans on momentum rather than earnings, and decide whether you believe the reason. Danelfin also publishes how each score level has done historically, with the universe, date range and sample size printed beside the chart: more homework than most tools show.",[13,380,381,384],{},[356,382,383],{},"The catch:"," Danelfin says it plainly. Its AI \"calculates probabilities, not certainties,\" and the scores are \"useful but not infallible information.\" Some stock pages also show an AI forecast for set time frames, so check which horizon a number means. It runs on the web and in free-to-download iOS and Android apps with paid plans inside.",[13,386,387,390],{},[356,388,389],{},"Best for:"," sorting a big universe of stocks by one comparable number, with the reasons on show.",[23,392,394],{"id":393},"kavout-rankings-by-strategy","Kavout: Rankings by Strategy",[13,396,397],{},"Kavout's AI Stock Picker ranks more than 9,000 US stocks every day with three measures: the Kai Score, Stock Rank and a Technical Rating. Kavout says they combine fundamental, technical and sentiment data, and the rankings refresh daily.",[13,399,400],{},"What you browse is a menu of strategies rather than one score: Greenblatt Magic Formula, Lynch Growth Momentum, Dual Analysis and more, each with its own top list for large caps, small caps, the S&P 500 and so on. Around the picker sit tools named InvestGPT, Financial Agents, Smart Signals and Market Movers.",[13,402,403,405],{},[356,404,383],{}," a top-ten list tells you which stocks a model prefers under its rules today. It does not tell you why one of them suits you, so the research on each name starts after the list. The picker itself sits behind the Pro tier, and Kavout's disclaimer says nothing on the site \"constitutes investment, financial, trading, or any other form of advice.\"",[13,407,408],{},"If you like screening by a named strategy and want a ranked shortlist to dig into, this is your tool.",[23,410,412],{"id":411},"trendspider-the-whole-cockpit","TrendSpider: The Whole Cockpit",[13,414,415],{},"TrendSpider is the multi-tool of the group. It is a full charting and research platform with automated pattern recognition, scanners, a no-code strategy tester that runs over decades of market data, and an ML Quant Lab that trains a machine learning model on inputs you choose.",[13,417,418],{},"The newest layer is Sidekick, its AI assistant. TrendSpider says Sidekick can see your charts, pull market data, read SEC filings, insider transactions, news and earnings call transcripts, and work as an agent: building scanners, creating alerts across a watch list and researching several symbols from one prompt. You choose which large language model answers, from Google, Anthropic or OpenAI, and even a personality, from long-term investor to number-crunching quant.",[13,420,421,423],{},[356,422,383],{}," breadth. TrendSpider rewards people who already chart every day and know what they want to test. For one quick read on one stock, it is a lot of cockpit. Its no-code bots can also place trades for you, a bigger commitment than reading a chart. Its own footer says nothing in the platform \"should be construed as investment advice of any kind.\"",[13,425,426,428],{},[356,427,389],{}," active traders who want charting, scanning, backtesting and an AI assistant under one roof.",[23,430,432],{"id":431},"tickeron-pattern-hunting-at-scale","Tickeron: Pattern Hunting at Scale",[13,434,435],{},"Tickeron builds its AI around chart shapes. Its AI Pattern Search Engine finds end-of-day chart patterns with breakout and target prices, confidence levels and how the pattern behaved in the past. AI Real Time Patterns does the same during the session, and the AI Trend Prediction Engine gives, in Tickeron's words, \"AI's opinion on whether a stock may go up or down.\"",[13,437,438],{},"If you would rather not run the tools yourself, there are Daily Buy\u002FSell Signals aimed at swing traders, delivered end of day, and AI robots that generate several trade ideas a day, which its robots page sorts into 5, 15 and 60 minute timeframes. Several of the tools offer a free try, and there are apps for iOS and Android.",[13,440,441,442,446],{},"If you already spot ",[282,443,445],{"href":444},"https:\u002F\u002Fprimotrade.app\u002Fblog\u002Fbullish-and-bearish-chart-patterns","bullish and bearish chart patterns"," by eye, Tickeron is a way to hunt for them across many stocks at once.",[13,448,449,451],{},[356,450,383],{}," a \"confidence level\" on a pattern describes how similar setups behaved before, not what this one will do. Same shape, new market.",[13,453,454,456],{},[356,455,389],{}," traders who think in patterns and want a machine to do the hunting.",[23,458,460],{"id":459},"trade-ideas-built-for-the-next-hour","Trade Ideas: Built for the Next Hour",[13,462,463],{},"Trade Ideas is the most caffeinated tool on this list. Its AI, called Holly, produces buy and sell signals with entry and exit points during the session. Around it sit Money Machine for momentum setups, TI Wave for market pattern recognition, a scanner, backtesting and simulated trading on live data. Orders can go straight to the market through its TradeStation link.",[13,465,466],{},"There is a free account to explore the scans and a free weekly email of AI setups. The AI signals themselves are part of the paid plans.",[13,468,469,471],{},[356,470,383],{}," on October 8, 2026 its homepage led with four recent Holly buy signals and the big moves that followed each one. That is a highlight reel, not a track record. A list of the best calls says nothing about the calls that did not work.",[13,473,474,476],{},[356,475,389],{}," day traders who want live scans and signals and already have a plan for acting on them.",[23,478,480],{"id":479},"prosperoai-signals-in-your-pocket","Prospero.ai: Signals in Your Pocket",[13,482,483],{},"Prospero.ai is a phone-first app on iOS and Android, and it markets itself as free. Its homepage says it combines millions of data points into key signals, drawing on institutional data and market movements.",[13,485,486],{},"In the app, cumulative scores show which stocks are gaining strength or losing steam, and an in-app guide explains how to read the signals.",[13,488,489,491],{},[356,490,383],{}," the headline is confident; the fine print is careful. The homepage invites you to \"identify short-term and long-term winners\" and publishes performance figures for its newsletters. Its own disclaimer, on the same page, says those results relate to signals from its models, \"are not actual trading results,\" and that simulated results \"do not represent actual trading.\" Read that part first, here and on any tool's page.",[13,493,494],{},"It suits investors who want a signal feed on their phone and treat it as one input among several.",[23,496,498],{"id":497},"what-about-just-asking-a-chatbot","What About Just Asking a Chatbot?",[13,500,501],{},"Fair question. You can paste a chart screenshot into a general assistant such as ChatGPT and ask what it sees; OpenAI says image inputs work on both free and paid plans.",[13,503,504],{},"What it reads is the picture plus your question. What comes back is whatever shape you ask for, which is both the appeal and the problem: you are designing the analysis as you go, and nothing fixes which data it checks or how it scores anything.",[13,506,507,508,513,514,518],{},"OpenAI's own ",[282,509,512],{"href":510,"target":285,"rel":511},"https:\u002F\u002Fhelp.openai.com\u002Fen\u002Farticles\u002F8400551-image-inputs-for-chatgpt-faq",[287],"image inputs FAQ"," is frank about the weak spots. It says the model may struggle with graphs where colors or line styles vary, struggles with tasks that need precise positions, and \"may generate incorrect descriptions.\" A ",[282,515,517],{"href":516},"https:\u002F\u002Fprimotrade.app\u002Fblog\u002Fhow-to-read-candlestick-charts","candlestick chart"," is mostly colors, line styles and positions.",[23,520,522],{"id":521},"primo-trade-one-stock-or-one-chart-at-a-time","Primo Trade: One Stock or One Chart at a Time",[13,524,525],{},"Our own app, so weigh this section accordingly. It reads one thing at a time, through one of two doors.",[13,527,528,529,533],{},"Search a US-listed stock or ADR ticker, from about 16,000 symbols, and the AI reads its prices, recent news and financial metrics. Or photograph or screenshot any chart, and the AI reads the chart itself: trend, volatility, volume, sentiment and the pattern it sees, which our ",[282,530,532],{"href":531},"https:\u002F\u002Fprimotrade.app\u002Fchart-patterns-cheat-sheet","chart patterns cheat sheet"," lets you look up. If it spots a US ticker on the image, the full ticker analysis runs instead.",[291,535],{"src":536,"alt":537,"width":295,"height":296},"https:\u002F\u002Fmedia.primotrade.app\u002Fblog\u002Fai-stock-analysis-tools-compared-what-each-one-reads\u002Fphone-photographing-chart-2.jpg","Seen from behind, a person holds a phone up to photograph a sideways candlestick chart on a desk monitor in the evening.",[13,539,540,541,545],{},"Every result carries a Signal from Strong Buy to Strong Sell as the AI's directional read, a Primo Score radar of five scores from 0 to 100, a Trading Strategy idea with a risk level, and up to three Key Insights. Ticker analyses add a price chart, Recent News and a Forecast Value, the model's 6 to 12 month estimate. A chart photo has no ticker data behind it, so it gets none of those three. The ",[282,542,544],{"href":543},"https:\u002F\u002Fprimotrade.app\u002Fblog\u002Fhow-ai-reads-a-stock-chart","way AI reads a chart"," has real limits, and they apply here.",[13,547,548],{},"Its gaps, as plainly as the others':",[550,551,552,556,559,562,565],"ul",{},[553,554,555],"li",{},"No market-wide scan or ranking; you bring the stock.",[553,557,558],{},"No alerts, watchlist, portfolio, backtests, bots or order placement.",[553,560,561],{},"No real-time data, and only US-listed stocks and ADRs.",[553,563,564],{},"Mobile only, on iOS and Android.",[553,566,567],{},"A crop with no axis or timeframe limits what the AI can read.",[13,569,570],{},"It is free to download, with free analyses to start, and it fits best as a second read on one stock or chart you already have open.",[23,572,574],{"id":573},"red-flags-regulators-have-named","Red Flags Regulators Have Named",[13,576,577,578,583],{},"\"AI\" makes a handy sticker, and regulators have noticed. In March 2024 the SEC announced ",[282,579,582],{"href":580,"target":285,"rel":581},"https:\u002F\u002Fwww.sec.gov\u002Fnewsroom\u002Fpress-releases\u002F2024-36",[287],"settled charges against"," two investment advisers, Delphia and Global Predictions, for false and misleading statements about their use of AI. The agency called the practice \"AI washing,\" and the firms agreed to pay civil penalties of $225,000 and $175,000.",[13,585,586,587,592],{},"FINRA, the SEC and NASAA also published a joint alert on ",[282,588,591],{"href":589,"target":285,"rel":590},"https:\u002F\u002Fwww.finra.org\u002Finvestors\u002Finsights\u002Fartificial-intelligence-and-investment-fraud",[287],"AI and fraud",". It warns about platforms claiming their AI \"can't lose,\" and cautions that AI-generated information might rest on data that is inaccurate, incomplete or out of date. Even when the input is correct, it says, AI output \"can be faulty, or even completely made up.\"",[13,594,595,596],{},"None of the tools in this guide make \"can't lose\" claims, and several carry careful disclaimers. The alert is still a good filter for anything new you find. ",[356,597,598],{},"If a tool sells certainty, the problem is the tool, not your skepticism.",[23,600,602],{"id":601},"five-checks-for-any-ai-stock-analysis-app","Five Checks for Any AI Stock Analysis App",[13,604,605],{},"Once you know which kind of tool fits, these five questions sort a useful one from a confusing one:",[550,607,608,615,621,627,633],{},[553,609,610,614],{},[611,612,613],"em",{},"Can you see why?"," A score with no visible reasons is hard to act on. Danelfin's explanation window is one example of a tool showing its work.",[553,616,617,620],{},[611,618,619],{},"Which horizon does it mean?"," A three-month probability, a 6 to 12 month estimate and an intraday signal answer different questions, as the Nvidia Monday shows.",[553,622,623,626],{},[611,624,625],{},"What does it cover?"," US only, US and Europe, ETFs or not. Check before you search for a stock it never rates.",[553,628,629,632],{},[611,630,631],{},"How are past results framed?"," Backtests, simulated results and homepage highlight reels are not a record of real trades. The better tools say so on the same page.",[553,634,635,638],{},[611,636,637],{},"Does it fit your week?"," A platform you open once a month will not repay the hours a full charting suite asks for.",[23,640,642],{"id":641},"free-ai-stock-analysis-what-you-can-try-first","Free AI Stock Analysis: What You Can Try First",[13,644,645],{},"Almost everything here lets you look before you pay, though what is free differs a lot.",[13,647,648],{},"Prospero.ai presents its app as free. Danelfin is free to download with paid plans inside, and Primo Trade is free to download, with free analyses to start. Tickeron offers free tries of several tools, Trade Ideas has a free account and a free weekly email of AI setups, and TrendSpider includes a small monthly allowance of free Sidekick messages. Kavout keeps the AI Stock Picker in its Pro tier.",[13,650,651],{},"A free tier is a fair way to judge one thing: whether the output makes sense to you. Run it on a stock you already know well. If the AI's read matches what you understand about the company, or points out something you missed and shows why, that is a good sign. If it gives a firm signal with no visible reason, the free trial has already told you what you needed to know.",[23,653,655],{"id":654},"the-bottom-line","The Bottom Line",[13,657,658],{},"AI stock analysis tools are not one product. Rankers score the whole market, scanners hunt for setups, signal apps push a feed, chatbots answer whatever you ask, and single-stock readers explain one name or one chart. Each reads different data and answers a different question.",[13,660,661],{},"Compare them on four things: what the AI reads, what it hands back, the horizon it means and how honestly it describes its own limits. Treat every score, signal and backtest as one input to your own research, and be wary of any tool that promises more than that.",{"title":663,"searchDepth":664,"depth":664,"links":665},"",2,[666,667,668,669,670,671,672,673,674,675,676,677,678,679,680],{"id":25,"depth":664,"text":26},{"id":124,"depth":664,"text":125},{"id":273,"depth":664,"text":274},{"id":362,"depth":664,"text":363},{"id":393,"depth":664,"text":394},{"id":411,"depth":664,"text":412},{"id":431,"depth":664,"text":432},{"id":459,"depth":664,"text":460},{"id":479,"depth":664,"text":480},{"id":497,"depth":664,"text":498},{"id":521,"depth":664,"text":522},{"id":573,"depth":664,"text":574},{"id":601,"depth":664,"text":602},{"id":641,"depth":664,"text":642},{"id":654,"depth":664,"text":655},"https:\u002F\u002Fmedia.primotrade.app\u002Fblog\u002Fai-stock-analysis-tools-compared-what-each-one-reads\u002Fcover-panel.jpg","AI stock analysis tools compared by what the AI reads and hands back: scores, scanners, signal feeds and chart-photo reads, each with its catch named.","md",{},true,"\u002Fblog\u002F2026\u002F10\u002Fai-stock-analysis-tools-compared-what-each-one-reads","2026-10-08",{"title":7,"description":682},{"loc":686,"images":690,"lastmod":687},[691,692],{"loc":293},{"loc":536},"blog\u002F2026\u002F10\u002Fai-stock-analysis-tools-compared-what-each-one-reads","F1hIQYIwP49AnQ55GTdHotncKBnMRSjEVdiF4tkGVIM",{"id":696,"title":697,"author":8,"body":698,"coverImg":1151,"description":1152,"extension":683,"head":8,"meta":1153,"navigation":685,"path":1154,"publishedAt":1155,"schemaOrg":8,"seo":1156,"sitemap":1157,"stem":1161,"updatedAt":687,"__hash__":1162},"blog\u002Fblog\u002F2026\u002F10\u002Fwhat-is-a-stop-loss.md","What Is a Stop-Loss? Stop, Stop-Limit and Trailing Stops",{"type":10,"value":699,"toc":1136},[700,703,706,713,716,720,729,732,741,744,748,757,760,786,789,793,796,803,806,809,812,816,823,826,831,906,910,913,922,931,934,937,940,946,950,957,960,963,966,970,973,982,985,988,992,995,1021,1025,1028,1035,1037,1048,1052,1055,1064,1067,1070,1077,1081,1126,1130,1133],[13,701,702],{},"You buy a stock at $50 and set a stop at $47. You go to bed feeling responsible. Overnight the company warns on its earnings, and in the morning your broker tells you the shares sold at $41.",[13,704,705],{},"Nothing broke. The order did exactly what it was built to do. It just was not built to do what you thought.",[13,707,708,709,712],{},"A stop-loss is an order to sell a stock automatically if its price falls to a level you pick in advance. ",[356,710,711],{},"The level is a trigger, not a price you are promised."," That one distinction explains most stop-loss surprises, and it is where the stop-limit and the trailing stop come in.",[13,714,715],{},"Meta's open on February 3, 2022 shows how far apart the trigger and the price can get.",[23,717,719],{"id":718},"how-a-stop-loss-order-works","How a Stop-Loss Order Works",[13,721,722,723,728],{},"The SEC's investor bulletin calls it a stop order, also known as a ",[282,724,727],{"href":725,"target":285,"rel":726},"https:\u002F\u002Fwww.investor.gov\u002Fintroduction-investing\u002Fgeneral-resources\u002Fnews-alerts\u002Falerts-bulletins\u002Finvestor-bulletins-15",[287],"stop-loss order",": an order to buy or sell once the stock reaches a price you set, the stop price. When the stop price is reached, the order becomes a market order.",[13,730,731],{},"A sell stop sits below today's price on a stock you own, to cap a loss or to protect a position that has already risen. A buy stop sits above today's price and does the same job for a short seller, who loses money when the price rises.",[13,733,734,735,740],{},"The words that matter are \"market order\". A market order is built to fill, not to fill at a particular price. As ",[282,736,739],{"href":737,"target":285,"rel":738},"https:\u002F\u002Fwww.investor.gov\u002Fintroduction-investing\u002Finvesting-basics\u002Fhow-stock-markets-work\u002Ftypes-orders",[287],"Investor.gov puts it",", the last traded price is not necessarily the price at which a market order will execute.",[13,742,743],{},"Brokers also differ on what counts as reaching the stop. The SEC notes that some use the last traded price and others use quotes. FINRA's rule for stop orders says a stop order is triggered by a trade at the stop price, and an order a broker triggers some other way has to carry a different name. So check what your broker means by \"stop\" before you rely on one.",[23,745,747],{"id":746},"the-stop-price-is-a-trigger-not-a-promise","The Stop Price Is a Trigger, Not a Promise",[13,749,750,751,756],{},"FINRA, the self-regulatory body for US brokers, put out guidance on this in May 2016. ",[282,752,755],{"href":753,"target":285,"rel":754},"https:\u002F\u002Fwww.finra.org\u002Frules-guidance\u002Fnotices\u002F16-19",[287],"Regulatory Notice 16-19"," leads its list of warnings with exactly this point: the stop price is not the execution price. A triggered stop is a market order, firms have to fill it promptly at the current market price, and in a fast market that price \"may be very different\" from the stop.",[13,758,759],{},"Three situations make the difference bigger:",[550,761,762,774,780],{},[553,763,764,767,768,773],{},[356,765,766],{},"Fast markets."," The ",[282,769,772],{"href":770,"target":285,"rel":771},"https:\u002F\u002Fwww.investor.gov\u002Fintroduction-investing\u002Finvesting-basics\u002Fglossary\u002Fstop-order",[287],"Investor.gov glossary"," warns that the fill can differ from the stop price, especially when prices change rapidly.",[553,775,776,779],{},[356,777,778],{},"Gaps."," If bad news lands overnight and a stock opens well below your stop, the stop is triggered at the open and sells at what buyers are paying then. The SEC's point is that a triggered stop fills at the prices of the liquidity available, not at the number you typed.",[553,781,782,785],{},[356,783,784],{},"Brief dips."," A short, sharp drop during the day can trigger a stop, and the SEC notes the sale can come out substantially worse than that day's closing price. FINRA adds that the stock may go back to its earlier level the same day, after you have sold.",[13,787,788],{},"FINRA's notice has one more uncomfortable line: in extreme volatility, sell stops being triggered can add downward pressure on the price, and a stop triggered in a steep decline is more likely to fill well below its price.",[23,790,792],{"id":791},"stop-vs-stop-limit-orders-on-the-same-gap","Stop vs Stop-Limit Orders on the Same Gap",[13,794,795],{},"Picture a stock that closes at $50. Two people own it and protect it differently.",[13,797,798,799,802],{},"Ana places a sell stop at $47. Ben places a sell ",[356,800,801],{},"stop-limit order"," with a stop price of $47 and a limit of $46: once $47 trades, his order becomes a limit order to sell at $46 or better, never worse.",[13,804,805],{},"Overnight, bad news. The stock opens at $41 and drifts lower all day.",[13,807,808],{},"Ana's stop triggers at the open, becomes a market order and sells close to $41. She is out, $6 a share below plan, but she is out.",[13,810,811],{},"Ben's order triggers too and becomes a limit order at $46. Nobody pays $46 for a stock trading at $41, so the order just sits there. Ben still owns every share, and they keep falling.",[291,813],{"src":814,"alt":815,"width":295,"height":296},"https:\u002F\u002Fmedia.primotrade.app\u002Fblog\u002Fwhat-is-a-stop-loss\u002Fstop-vs-stop-limit-gap-panel.jpg","Diagram of a price line that gaps down overnight past a stop price and a limit price, where the stop sells at the lower open and the stop-limit is left unfilled.",[13,817,818,819,822],{},"This is the trade-off FINRA describes. ",[356,820,821],{},"A stop buys you an exit at a price you did not choose; a stop-limit buys you a price, with the risk of no exit at all."," FINRA's words for the stop-limit are \"additional certainty\" on price and \"the possibility that the order will not be executed at all\". It even suggested brokers consider making the stop-limit the default, so a plain stop becomes a deliberate choice.",[13,824,825],{},"The SEC's own example uses small numbers: a sell stop-limit with a stop price of $3.00 and a limit of $2.50 becomes an active limit order when $3.00 trades, and can only sell at $2.50 or better. If the price falls straight through $2.50, nothing sells.",[827,828,830],"h3",{"id":829},"stop-stop-limit-and-trailing-stop-side-by-side","Stop, Stop-Limit and Trailing Stop Side by Side",[31,832,833,848],{},[34,834,835],{},[37,836,837,839,842,845],{},[40,838],{},[40,840,841],{},"Stop (stop-loss)",[40,843,844],{},"Stop-limit",[40,846,847],{},"Trailing stop",[50,849,850,864,878,892],{},[37,851,852,855,858,861],{},[55,853,854],{},"When triggered, it becomes",[55,856,857],{},"A market order",[55,859,860],{},"A limit order",[55,862,863],{},"A market order, or a limit order if set up as a trailing stop-limit",[37,865,866,869,872,875],{},[55,867,868],{},"What you can count on",[55,870,871],{},"A sale, at whatever price is there",[55,873,874],{},"A price at your limit or better",[55,876,877],{},"A stop that rises as the price rises",[37,879,880,883,886,889],{},[55,881,882],{},"What you give up",[55,884,885],{},"Control of the price",[55,887,888],{},"Certainty of a sale",[55,890,891],{},"Room: a normal pullback can set it off",[37,893,894,897,900,903],{},[55,895,896],{},"When the stock gaps past it",[55,898,899],{},"Sells at the lower price",[55,901,902],{},"May not sell at all",[55,904,905],{},"Behaves like the stop or stop-limit it is",[23,907,909],{"id":908},"a-real-gap-meta-on-february-3-2022","A Real Gap: Meta on February 3, 2022",[13,911,912],{},"Here is the same thing on a real stock, read after the fact: a past event that shows how orders behave, not a view on the stock.",[13,914,915,916,921],{},"Meta Platforms, then still trading as FB, reported fourth-quarter results after the close on Wednesday, February 2, 2022. It had closed that day at $323, according to ",[282,917,920],{"href":918,"target":285,"rel":919},"https:\u002F\u002Fmarkets.businessinsider.com\u002Fnews\u002Fstocks\u002Ffacebook-stock-price-meta-q4-earnings-mark-zuckerberg-tiktok-revenue-2022-2",[287],"Business Insider",". At 9:34 a.m. Eastern the next morning it was at $240.31.",[13,923,924,925,930],{},"Yahoo Finance's daily price history for February 3 shows an open of $244.65 and a high of $248.00. By the close it had fallen $85.24, or 26%, to $237.76, as ",[282,926,929],{"href":927,"target":285,"rel":928},"https:\u002F\u002Fwww.cbsnews.com\u002Fnews\u002Fstock-market-down-meta-facebook-earnings",[287],"CBS News reported",", the biggest one-day loss of market value for a public company on record at the time.",[13,932,933],{},"Now imagine two holders going into that night.",[13,935,936],{},"The first had a sell stop at $300, about 7% under Wednesday's close. The stock opened far below $300, so the stop would have triggered at the open and sold at or near $244.65, about $55 a share below the stop.",[13,938,939],{},"The second had a stop-limit, stop $300 and limit $295. It triggered at the open too and became a limit order to sell at $295 or better. The stock never traded above $248 that day, so the order never filled.",[13,941,942,943],{},"A chart read on Wednesday afternoon could not have known this: the price that mattered came from a release published while the market was shut. ",[356,944,945],{},"No stop placement on a chart protects against a price that never trades through it.",[23,947,949],{"id":948},"trailing-stop-loss-a-stop-that-follows-the-price-up","Trailing Stop Loss: A Stop That Follows the Price Up",[13,951,952,953,956],{},"A ",[356,954,955],{},"trailing stop"," is a stop or stop-limit order whose stop price is not a fixed number. It is a dollar amount or a percentage below the market price, and it moves up as the price rises. When the price falls, it stays put.",[13,958,959],{},"The SEC's own example in the same bulletin runs like this. You buy a stock at $20. It rises to $22, and you set a sell trailing stop $1 below the market price. The stock climbs to $24, so the stop climbs to $23. The price then turns down, the stop stays at $23, and when $23 trades the shares are sold, though the SEC notes the sale price may differ from $23.",[13,961,962],{},"That is the appeal: the floor rises with the stock and you never have to move it by hand. The cost is choosing the distance. Set the trail tight and an ordinary wobble sells you out; the SEC warns that short-term fluctuations can trigger a trailing stop. Set it wide and you give back more of the rise before it fires.",[13,964,965],{},"And on a gap, a trailing stop is just a stop. Had one been trailing Meta at $300 that February night, it would have met the same open as everyone else.",[23,967,969],{"id":968},"when-a-blip-sets-it-off-august-24-2015","When a Blip Sets It Off: August 24, 2015",[13,971,972],{},"Gaps are one problem. The other is the drop that comes and goes before you have finished your coffee.",[13,974,975,976,981],{},"On Monday, August 24, 2015, the most actively traded US equity product, the SPDR S&P 500 ETF Trust, opened sharply lower and by 9:35 a.m. was 7.8% below Friday's close. By 9:40 it had recovered past its opening price, and it closed the day down 4.2%. That is from the SEC staff's ",[282,977,980],{"href":978,"target":285,"rel":979},"https:\u002F\u002Fwww.sec.gov\u002Fmarketstructure\u002Fresearch\u002Fequity_market_volatility.pdf",[287],"research note"," on the day.",[13,983,984],{},"Run a sell stop 6% under Friday's close through that morning. It triggers within the first five minutes, sells somewhere in the drop, and the fund closes the day above the price you got. That is the SEC's \"substantially worse than the closing price\" warning, played out in real time.",[13,986,987],{},"FINRA's notice suggested that brokers consider limiting stop triggers around the open and the close, when markets may be more volatile, and setting an expiry on good-til-cancelled stops, giving 90 days as an example. Your broker may do one, both or neither. It is worth finding out which.",[23,989,991],{"id":990},"where-traders-set-a-stop","Where Traders Set a Stop",[13,993,994],{},"There is no official rule. The common approaches:",[550,996,997,1009,1015],{},[553,998,999,1002,1003,1005,1006,1008],{},[356,1000,1001],{},"Below a level on the chart."," Many traders put the stop just under a recent low or support line, so it is hit only if the chart's story breaks. Our guide to ",[282,1004,445],{"href":444}," explains the levels traders watch, and the ",[282,1007,532],{"href":531}," draws them.",[553,1010,1011,1014],{},[356,1012,1013],{},"A fixed percentage."," Some simply use a set distance below the entry.",[553,1016,1017,1020],{},[356,1018,1019],{},"By account risk."," A widely repeated trading convention is to size each position so hitting the stop costs only a small, fixed slice of the account. It is a practitioner habit, not a regulatory or academic standard.",[291,1022],{"src":1023,"alt":1024,"width":295,"height":296},"https:\u002F\u002Fmedia.primotrade.app\u002Fblog\u002Fwhat-is-a-stop-loss\u002Fpencil-line-under-low-2.jpg","A hand drawing a pencil line just under the most recent candles on a printed candlestick chart on a desk in evening light.",[13,1026,1027],{},"Whatever the method, the failure modes are the same. Too close, and normal movement takes you out. Too far, and the stop does not limit much.",[13,1029,1030,1031,1034],{},"If you want a second read of a chart's trend and pattern, ",[282,1032,104],{"href":1033},"https:\u002F\u002Fprimotrade.app\u002Fdownload?c=post_inline"," can analyze a ticker or a chart photo, name the chart pattern its AI sees and suggest a strategy idea with a risk level. A risk level in Primo Trade's Trading Strategy section is the AI's read of the data, not an order. Nothing sells because an analysis said \"high risk\"; only an order at your broker does that.",[13,1036,270],{},[13,1038,1039,1040,1043,1044,1047],{},"Our article on ",[282,1041,1042],{"href":543},"how AI reads a stock chart"," explains how to check its read, and our guide to ",[282,1045,1046],{"href":516},"reading candlestick charts"," covers the basics.",[23,1049,1051],{"id":1050},"do-stop-loss-rules-actually-help","Do Stop-Loss Rules Actually Help?",[13,1053,1054],{},"The honest answer is \"it depends on how prices behave\", and there is real research behind that.",[13,1056,1057,1058,1063],{},"Kathryn Kaminski and Andrew Lo studied ",[282,1059,1062],{"href":1060,"target":285,"rel":1061},"https:\u002F\u002Fpapers.ssrn.com\u002Fsol3\u002Fpapers.cfm?abstract_id=968338",[287],"when stop-loss rules"," actually stop losses. They showed that if prices move like a random walk, simple stop-loss rules always lower a strategy's expected performance. When prices have momentum, meaning recent moves tend to continue, stop-loss rules can add value.",[13,1065,1066],{},"The 2007 working paper tested this on US stocks with monthly data from 1950 to 2004, with a strategy that moved into long-term government bonds whenever it was stopped out. Certain rules added 50 to 100 basis points per month during stop-out periods. That covers only the months spent in bonds, and it describes a whole basket of US stocks checked monthly, not a single-stock stop placed during the day.",[13,1068,1069],{},"The 2014 version in the Journal of Financial Markets used index futures with daily prices. There, certain stop-loss policies improved expected performance and cut volatility over longer intervals, which again says little about a stop that fires on one morning's dip.",[13,1071,1072,1073,1076],{},"So here is the limit, said plainly. ",[356,1074,1075],{},"A stop-loss is a risk control, not a source of performance."," It caps what one decision can cost you when the price moves through it in an orderly way. It does nothing against a price that opens below it, and it can cost you when a stock dips, sells you out, and recovers by lunch.",[23,1078,1080],{"id":1079},"how-to-use-a-stop-loss","How to Use a Stop-Loss",[1082,1083,1084,1090,1096,1102,1108,1114,1120],"ol",{},[553,1085,1086,1089],{},[356,1087,1088],{},"Decide the stop before you buy."," It is much harder to choose a sensible level once a trade is already losing.",[553,1091,1092,1095],{},[356,1093,1094],{},"Pick the order type on purpose."," A stop gets you out at an uncertain price; a stop-limit holds your price but may not get you out.",[553,1097,1098,1101],{},[356,1099,1100],{},"Set the limit gap on purpose, too."," On a stop-limit, the further the limit sits below the stop, the more room it has to fill and the worse the price it may accept.",[553,1103,1104,1107],{},[356,1105,1106],{},"Leave room for normal movement."," Look at how much the stock usually moves in a day before choosing the distance.",[553,1109,1110,1113],{},[356,1111,1112],{},"Ask your broker four things."," Which of the three order types it offers at all, since the SEC notes they are not available at every firm; what triggers a stop (trades or quotes); whether stops work outside regular hours; and when a good-til-cancelled stop expires.",[553,1115,1116,1119],{},[356,1117,1118],{},"Know when gaps are likely."," Results releases and other news outside market hours are when a price can jump straight past your stop.",[553,1121,1122,1125],{},[356,1123,1124],{},"Do not move it down."," Lowering a stop as the price approaches it turns a plan into a hope.",[23,1127,1129],{"id":1128},"conclusion","Conclusion",[13,1131,1132],{},"A stop-loss order sells a stock automatically once its price reaches a level you set in advance. A plain stop turns into a market order and can fill well below the stop price in a fast market or a gap; a stop-limit holds the price but may not sell at all; a trailing stop follows a rising price up and stays put when it falls.",[13,1134,1135],{},"Meta's open on February 3, 2022 and the first minutes of August 24, 2015 show the two ways a stop surprises people: a price that jumps past it, and a dip that takes it out before recovering. Research by Kaminski and Lo found stop-loss rules lower expected performance when prices move randomly and can help when prices have momentum. Treat a stop as a limit on what one decision can cost, set it before you enter, and know how your broker triggers and fills it.",{"title":663,"searchDepth":664,"depth":664,"links":1137},[1138,1139,1140,1144,1145,1146,1147,1148,1149,1150],{"id":718,"depth":664,"text":719},{"id":746,"depth":664,"text":747},{"id":791,"depth":664,"text":792,"children":1141},[1142],{"id":829,"depth":1143,"text":830},3,{"id":908,"depth":664,"text":909},{"id":948,"depth":664,"text":949},{"id":968,"depth":664,"text":969},{"id":990,"depth":664,"text":991},{"id":1050,"depth":664,"text":1051},{"id":1079,"depth":664,"text":1080},{"id":1128,"depth":664,"text":1129},"https:\u002F\u002Fmedia.primotrade.app\u002Fblog\u002Fwhat-is-a-stop-loss\u002Fcover-graphite.jpg","What a stop-loss order does, how stop, stop-limit and trailing stops differ, and why a gap can sell yours far below the stop price or leave it unfilled.",{},"\u002Fblog\u002F2026\u002F10\u002Fwhat-is-a-stop-loss","2026-10-07",{"title":697,"description":1152},{"loc":1154,"images":1158,"lastmod":687},[1159,1160],{"loc":814},{"loc":1023},"blog\u002F2026\u002F10\u002Fwhat-is-a-stop-loss","IwOBbRWFaV03uEnuQHrPjB1itFDVPC8eYFgCuN9--pk",{"id":1164,"title":1165,"author":8,"body":1166,"coverImg":1672,"description":1673,"extension":683,"head":8,"meta":1674,"navigation":685,"path":1675,"publishedAt":1676,"schemaOrg":8,"seo":1677,"sitemap":1678,"stem":1682,"updatedAt":687,"__hash__":1683},"blog\u002Fblog\u002F2026\u002F10\u002Fhow-ai-reads-a-stock-chart.md","AI Chart Analysis: What a Screenshot Can and Cannot Show",{"type":10,"value":1167,"toc":1654},[1168,1171,1174,1181,1185,1188,1191,1194,1198,1201,1205,1314,1317,1321,1325,1334,1337,1340,1343,1346,1355,1359,1362,1366,1369,1378,1382,1391,1394,1397,1401,1404,1419,1423,1426,1500,1509,1512,1516,1525,1534,1537,1541,1555,1563,1566,1570,1619,1623,1631,1639,1642,1644,1648,1651],[13,1169,1170],{},"It is late, a chart on your monitor looks interesting, and you grab a screenshot on your phone. You crop it so it looks tidy, drop it into an AI and ask what it sees. Back comes a confident paragraph: a steady uptrend, a short pause that looks like a bull flag, volume picking up into the right edge.",[13,1172,1173],{},"Here is the question nobody asks at that moment. You cropped off the price axis, the ticker and the dates. So what exactly was the AI reading?",[13,1175,1176,1177,1180],{},"The shape. ",[356,1178,1179],{},"AI chart analysis works from the pixels in front of it, and a screenshot keeps the shape of the price while throwing away most of what gives that shape a meaning."," This post walks through what an AI can read from a chart image, what it cannot know no matter how clever it is, where published research says these models slip, and a short routine for getting a read you can check.",[23,1182,1184],{"id":1183},"what-an-ai-sees-when-it-reads-a-stock-chart","What an AI Sees When It Reads a Stock Chart",[13,1186,1187],{},"What the AI got from your crop was pixels, not prices. A multimodal model, the kind that takes in an image and text together, never sees the numbers behind a chart as a table. It sees lines, colours, candle bodies and maybe a few labels along the edges. From those it infers what the chart says, much as you would, and then puts it into words.",[13,1189,1190],{},"That has two consequences. The model can take in the overall shape of a chart fast, which is often exactly what you want from a second read. And everything it says depends on how well it read the picture, so a blurred photo, a cropped axis or an unusual chart style can change its answer.",[13,1192,1193],{},"Think of it as a very quick reader handed a page torn out of a book. It can tell you the story on the page. It cannot tell you which book it came from, or how the next chapter goes.",[23,1195,1197],{"id":1196},"ai-chart-analysis-readable-versus-missing","AI Chart Analysis: Readable Versus Missing",[13,1199,1200],{},"Some things live in the pixels. Others live in labels that a crop removes, and a few were never on the screen at all.",[291,1202],{"src":1203,"alt":1204,"width":295,"height":296},"https:\u002F\u002Fmedia.primotrade.app\u002Fblog\u002Fhow-ai-reads-a-stock-chart\u002Fscreenshot-readable-vs-missing-panel.jpg","Diagram of a cropped candlestick chart screenshot, with the rising trend, a candle, the sideways flag pattern and the volume marked as readable, and the ticker, timeframe, price axis and dates marked as missing.",[31,1206,1207,1220],{},[34,1208,1209],{},[37,1210,1211,1214,1217],{},[40,1212,1213],{},"What you want to know",[40,1215,1216],{},"In the pixels?",[40,1218,1219],{},"Where it has to come from",[50,1221,1222,1233,1242,1252,1263,1274,1284,1294,1304],{},[37,1223,1224,1227,1230],{},[55,1225,1226],{},"Direction of the trend",[55,1228,1229],{},"Yes",[55,1231,1232],{},"The candles themselves",[37,1234,1235,1238,1240],{},[55,1236,1237],{},"Shape of a pattern, such as a flag or a double top",[55,1239,1229],{},[55,1241,1232],{},[37,1243,1244,1247,1250],{},[55,1245,1246],{},"Long wicks, small bodies, gaps between candles",[55,1248,1249],{},"Yes, if the image is sharp",[55,1251,1232],{},[37,1253,1254,1257,1260],{},[55,1255,1256],{},"Whether volume rose or fell",[55,1258,1259],{},"Yes, if the volume bars are in frame",[55,1261,1262],{},"The volume panel",[37,1264,1265,1268,1271],{},[55,1266,1267],{},"Which stock it is",[55,1269,1270],{},"No",[55,1272,1273],{},"The ticker label, or you",[37,1275,1276,1279,1281],{},[55,1277,1278],{},"The price level, and how big a move was in percent",[55,1280,1270],{},[55,1282,1283],{},"The price axis",[37,1285,1286,1289,1291],{},[55,1287,1288],{},"Whether each candle is five minutes or one week",[55,1290,1270],{},[55,1292,1293],{},"The timeframe label or the dates",[37,1295,1296,1299,1301],{},[55,1297,1298],{},"When it happened, and what news was out",[55,1300,1270],{},[55,1302,1303],{},"The dates, then a news check",[37,1305,1306,1309,1311],{},[55,1307,1308],{},"What happened after the right edge",[55,1310,1270],{},[55,1312,1313],{},"Nothing: no screenshot has it",[13,1315,1316],{},"The top half of that table is where an AI chart read pulls its weight. The bottom half is where it can only guess, and a guess written in fluent sentences looks a lot like knowledge.",[23,1318,1320],{"id":1319},"three-things-a-cropped-screenshot-throws-away","Three Things a Cropped Screenshot Throws Away",[827,1322,1324],{"id":1323},"the-timeframe","The timeframe",[13,1326,1327,1328,1333],{},"Traders have long said that charts look alike at every zoom level, and a famous mathematician took the folklore seriously. Benoit Mandelbrot, the man behind fractals, wrote that once a stock chart is enlarged or shrunk to fit the same frame, an observer ",[282,1329,1332],{"href":1330,"target":285,"rel":1331},"https:\u002F\u002Fwww.scientificamerican.com\u002Farticle\u002Fmultifractals-explain-wall-street\u002F",[287],"cannot tell which"," data change week to week, day to day or hour to hour. He published that in Scientific American in 1999.",[13,1335,1336],{},"For a screenshot, that means a bull flag on a five-minute chart and a bull flag on a weekly chart can be the same picture. One describes an afternoon, the other most of a year. Without a timeframe label or readable dates, the AI cannot tell them apart, and neither can you.",[827,1338,1283],{"id":1339},"the-price-axis",[13,1341,1342],{},"Without the axis, the AI cannot know whether the stock trades at 3 dollars or 300, or whether a tall candle was a 1% wobble or a 10% drop. Those are not small details.",[13,1344,1345],{},"The scale matters too. On a linear chart, a move from 10 to 20 takes the same height as a move from 70 to 80; on a log chart, equal percentage moves take equal height. Over a long stretch, the same prices drawn both ways can give different trendlines, and a screenshot rarely says which scale it used.",[13,1347,1348,1349,1354],{},"Reading values off an axis is also one of the weakest skills these models have. FinChart-Bench, the first benchmark built only from real financial charts (1,200 charts from 2015 to 2024 and 7,016 questions, posted in 2025 and presented at ACL 2026), found that models did relatively well when numbers were printed next to the points and poorly when a value had to be inferred ",[282,1350,1353],{"href":1351,"target":285,"rel":1352},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2507.14823",[287],"from the axes",". Its authors call spatial reasoning a significant bottleneck. A price chart is exactly that kind of chart: almost nothing is labelled, and everything is read against an axis.",[827,1356,1358],{"id":1357},"the-ticker","The ticker",[13,1360,1361],{},"No ticker means no company, no earnings date and no news, so no way to know whether the gap on the right was a company announcement or a market-wide sell-off. An AI working from pixels alone reads the chart the way a technical analyst would with a sticky note over the name. That is an interesting exercise, and an incomplete one.",[23,1363,1365],{"id":1364},"when-the-picture-lies-nvidias-split-in-june-2024","When the Picture Lies: Nvidia's Split in June 2024",[13,1367,1368],{},"Here is the failure case as one real, closed example from June 2024, read after the fact.",[13,1370,1371,1372,1377],{},"Nvidia split its stock ten for one that month. As ",[282,1373,1376],{"href":1374,"target":285,"rel":1375},"https:\u002F\u002Ffinance.yahoo.com\u002Fnews\u002Fnvidia-stock-rises-after-10-for-1-stock-split-204412528.html",[287],"Yahoo Finance reported"," on Monday, June 10, 2024, the shares began trading that day on the new basis, Friday's closing price of $1,208.88 was revised to $120.88, and the stock closed up nearly 1% on its first day after the split. Nothing bad happened to the company that weekend. Each shareholder simply held ten shares for every one.",[291,1379],{"src":1380,"alt":1381,"width":295,"height":296},"https:\u002F\u002Fmedia.primotrade.app\u002Fblog\u002Fhow-ai-reads-a-stock-chart\u002Fsplit-cliff-chart-panel.jpg","Stylised candlestick chart that climbs steadily, then drops in one sheer vertical cliff and carries on calmly at a much lower level.",[13,1383,1384,1385,1390],{},"Now picture a chart that does not adjust its history for the split. It shows a stock that climbed for months and then lost about 90% over a weekend. StockCharts, which adjusts its own data for exactly this reason, warns that an unadjusted split ",[282,1386,1389],{"href":1387,"target":285,"rel":1388},"https:\u002F\u002Fhelp.stockcharts.com\u002Fdata-and-ticker-symbols\u002Fdata-availability\u002Fprice-data-adjustments",[287],"gives the impression"," that something bearish happened, and that most technical indicators on such a chart would give sell signals.",[13,1392,1393],{},"Show a screenshot of that cliff to an AI with no ticker and no dates, and \"a collapse\" is a perfectly reasonable description of the pixels. It is also completely wrong. The fix is not a smarter model. It is context: the ticker, the date and a quick news check.",[13,1395,1396],{},"The general rule holds well beyond splits: the same picture of a price can come from two very different stories, and the picture alone cannot say which.",[23,1398,1400],{"id":1399},"how-well-ai-reads-charts-what-the-benchmarks-say","How Well AI Reads Charts: What the Benchmarks Say",[13,1402,1403],{},"So how often do these models get a chart right? Researchers measure it with benchmarks: sets of real charts with questions that have checkable answers.",[13,1405,1406,1407,1412,1413,1418],{},"Early ones such as the ",[282,1408,1411],{"href":1409,"target":285,"rel":1410},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2203.10244",[287],"ChartQA benchmark",", published in 2022, asked thousands of questions about charts, including questions that need arithmetic over the values shown, and models improved on them quickly. The authors of CharXiv, a 2024 benchmark, argued that such datasets used simple, uniform charts and template questions, which gave an ",[282,1414,1417],{"href":1415,"target":285,"rel":1416},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2406.18521",[287],"over-optimistic measure"," of progress. When they changed the charts or questions only slightly, scores fell by up to 34.5%.",[827,1420,1422],{"id":1421},"harder-tests-wider-gap","Harder tests, wider gap",[13,1424,1425],{},"Harder tests show a clear gap between models and people:",[31,1427,1428,1447],{},[34,1429,1430],{},[37,1431,1432,1435,1438,1441,1444],{},[40,1433,1434],{},"Benchmark",[40,1436,1437],{},"What it tests",[40,1439,1440],{},"People",[40,1442,1443],{},"Best model at release",[40,1445,1446],{},"Best open-source model",[50,1448,1449,1466,1483],{},[37,1450,1451,1454,1457,1460,1463],{},[55,1452,1453],{},"CharXiv (2024), reasoning questions",[55,1455,1456],{},"2,323 real charts from scientific papers",[55,1458,1459],{},"80.5%",[55,1461,1462],{},"47.1% (GPT-4o)",[55,1464,1465],{},"29.2%",[37,1467,1468,1471,1474,1477,1480],{},[55,1469,1470],{},"ChartMuseum (NeurIPS 2025)",[55,1472,1473],{},"1,162 questions on real charts from 184 sources",[55,1475,1476],{},"93%",[55,1478,1479],{},"63.0% (Gemini 2.5 Pro)",[55,1481,1482],{},"38.5%",[37,1484,1485,1488,1491,1494,1497],{},[55,1486,1487],{},"FinChart-Bench (2025), open questions",[55,1489,1490],{},"1,200 real financial charts",[55,1492,1493],{},"Not reported",[55,1495,1496],{},"63.59%",[55,1498,1499],{},"59.78%",[13,1501,1502,1503,1508],{},"The ",[282,1504,1507],{"href":1505,"target":285,"rel":1506},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2505.13444",[287],"ChartMuseum authors"," also split questions that need mainly visual reasoning from questions that lean on text. On the visual ones, every model they tested dropped 35% to 55% from its score on the text-heavy ones. In a synthetic test that could only be solved by looking, models got worse as the picture got busier, while people held steady. A price chart is almost all visual, so the parts that need careful looking, such as exact levels, small candles and crowded stretches, are where a model is most likely to slip.",[13,1510,1511],{},"CharXiv and ChartMuseum use scientific and general charts, FinChart-Bench's 14 chart types are bars, lines, pies and the like rather than candlesticks, and newer models keep raising the scores. The direction is still clear: on charts that need real looking, the best models trail people by a wide margin.",[23,1513,1515],{"id":1514},"can-ai-predict-a-stock-from-its-chart","Can AI Predict a Stock From Its Chart?",[13,1517,1518,1519,1524],{},"The most cited research here is about text, not charts. Alejandro Lopez-Lira and Yuehua Tang had GPT-4 score news headlines published after its knowledge cutoff. The scores captured the market's first reaction to the news, which they describe as non-tradable, and also predicted some of the drift afterwards, especially for small stocks and bad news. The edge also shrank: in their data it ",[282,1520,1523],{"href":1521,"target":285,"rel":1522},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2304.07619",[287],"faded as adoption"," of these models rose.",[13,1526,1527,1528,1533],{},"Testing such claims is tricky. Paul Glasserman and Caden Lin showed that backtests of AI news strategies can be ",[282,1529,1532],{"href":1530,"target":285,"rel":1531},"https:\u002F\u002Farxiv.org\u002Fabs\u002F2309.17322",[287],"quietly biased"," when the model already knows what happened to the stock after the news, or when its general knowledge of a company distracts it from the text in front of it.",[13,1535,1536],{},"None of this research shows that an AI reading a chart can tell you where a price will go. A chart is a record of past prices, and an AI read of it is one more reading of that same record: quick, often useful, and never a view of the future.",[23,1538,1540],{"id":1539},"confident-fluent-and-sometimes-wrong","Confident, Fluent and Sometimes Wrong",[13,1542,1543,1544,1548,1549,1554],{},"The SEC, NASAA and FINRA, in a joint alert on AI and investment fraud, warn that even when the input is correct, AI output ",[282,1545,1547],{"href":589,"target":285,"rel":1546},[287],"can be faulty",", or even completely made up. The same alert quotes pitches such as \"Our proprietary AI trading system can't lose!\" as the kind to run from, and it urges caution about ",[282,1550,1553],{"href":1551,"target":285,"rel":1552},"https:\u002F\u002Fwww.investor.gov\u002Fintroduction-investing\u002Fgeneral-resources\u002Fnews-alerts\u002Falerts-bulletins\u002Finvestor-alerts\u002Fartificial-intelligence-fraud",[287],"using AI-generated information"," to predict the market's direction or a stock's price, because it may rest on data that is incomplete, outdated or misleading.",[13,1556,1557,1558,1562],{},"The SEC has acted on overstated AI claims too. In 2024 it settled charges against two firms, Delphia and Global Predictions, for false and misleading statements about their use of AI, with ",[282,1559,1561],{"href":580,"target":285,"rel":1560},[287],"$400,000 in penalties"," between them.",[13,1564,1565],{},"A chart description reads just as smoothly whether it is right or wrong. So the useful habit is neither trusting nor distrusting the AI. It is checking it.",[23,1567,1569],{"id":1568},"how-to-get-an-ai-chart-read-you-can-check","How to Get an AI Chart Read You Can Check",[1082,1571,1572,1578,1584,1590,1602,1608],{},[553,1573,1574,1577],{},[356,1575,1576],{},"Keep the labels in."," Crop to the chart, but leave the ticker, the price axis and the dates in frame. They are the whole bottom half of the table above.",[553,1579,1580,1583],{},[356,1581,1582],{},"Name the timeframe if it is not shown."," \"Daily candles, about six months\" changes what a pattern means.",[553,1585,1586,1589],{},[356,1587,1588],{},"Shoot straight and sharp."," A photo of a screen at an angle, with glare, blurs the small candles a model already finds hardest.",[553,1591,1592,1595,1596,1598,1599,1601],{},[356,1593,1594],{},"Check every claim against the chart."," If it names a level or a pattern, find it yourself. You can only do that if you can read the chart, which our guides to ",[282,1597,1046],{"href":516}," and ",[282,1600,445],{"href":444}," cover.",[553,1603,1604,1607],{},[356,1605,1606],{},"Ask what it cannot see."," A good follow-up is \"what would change this read?\" Earnings next week, a split or a market-wide drop are all invisible in a cropped chart.",[553,1609,1610,1613,1614,1618],{},[356,1611,1612],{},"Plan for being wrong."," Decide in advance where you would get out. Our guide to ",[282,1615,1617],{"href":1616},"https:\u002F\u002Fprimotrade.app\u002Fblog\u002Fwhat-is-a-stop-loss","what a stop-loss is"," explains how.",[23,1620,1622],{"id":1621},"how-primo-trade-reads-a-chart-photo","How Primo Trade Reads a Chart Photo",[13,1624,1625,1627,1628,1630],{},[282,1626,104],{"href":1033}," lets you take a photo of a chart or choose a screenshot from your gallery. Its AI reads the trend, volatility, volume and sentiment it sees, the chart pattern, and what that pattern usually signals. Our ",[282,1629,532],{"href":531}," shows what confirms each one. If the image is not a chart, the app says so instead of guessing.",[13,1632,1633,1634,1638],{},"The limits in this post apply to it as well, and our ",[282,1635,1637],{"href":1636},"https:\u002F\u002Fprimotrade.app\u002Fblog\u002Fai-stock-analysis-tools-compared-what-each-one-reads","comparison of AI stock analysis tools"," sets it beside six others. When the AI can read a US ticker on the image, the app runs the full ticker analysis instead, pulling prices, financial metrics and recent news before the AI writes, and the result adds a price chart and a Forecast Value, the model's 6 to 12 month estimate. When it cannot read a ticker, it works from the image alone, so the result has no price chart, no Forecast Value and no news. One more reason to keep the ticker in frame.",[13,1640,1641],{},"Either way, the result shows a signal from Strong Buy to Strong Sell as the AI's read, a Primo Score of five ratings from 0 to 100, a Trading Strategy idea with a risk level, and Key Insights. A Risk score of 80 is a rating on a scale, not an 80% chance of anything.",[13,1643,270],{},[23,1645,1647],{"id":1646},"what-an-ai-chart-read-is-good-for","What an AI Chart Read Is Good For",[13,1649,1650],{},"An AI reads a stock chart the way it reads any image: from pixels into words. From a screenshot it can describe the trend, the candles, the pattern and the volume in seconds. It cannot know the ticker, the price level, the timeframe or the date unless they are in the picture, and the best models still trail people on charts that need careful looking, most of all when a value has to be read off an axis.",[13,1652,1653],{},"So keep the labels in, check what it says against the chart, and treat it as a quick second opinion on past prices. Any tool that claims an AI can see the future in a chart is the warning sign regulators describe.",{"title":663,"searchDepth":664,"depth":664,"links":1655},[1656,1657,1658,1663,1664,1667,1668,1669,1670,1671],{"id":1183,"depth":664,"text":1184},{"id":1196,"depth":664,"text":1197},{"id":1319,"depth":664,"text":1320,"children":1659},[1660,1661,1662],{"id":1323,"depth":1143,"text":1324},{"id":1339,"depth":1143,"text":1283},{"id":1357,"depth":1143,"text":1358},{"id":1364,"depth":664,"text":1365},{"id":1399,"depth":664,"text":1400,"children":1665},[1666],{"id":1421,"depth":1143,"text":1422},{"id":1514,"depth":664,"text":1515},{"id":1539,"depth":664,"text":1540},{"id":1568,"depth":664,"text":1569},{"id":1621,"depth":664,"text":1622},{"id":1646,"depth":664,"text":1647},"https:\u002F\u002Fmedia.primotrade.app\u002Fblog\u002Fhow-ai-reads-a-stock-chart\u002Fcover-v3.jpg","AI chart analysis reads trend, candles and patterns from a screenshot, but not the ticker, the timeframe or the price scale. What it sees, where it slips.",{},"\u002Fblog\u002F2026\u002F10\u002Fhow-ai-reads-a-stock-chart","2026-10-06",{"title":1165,"description":1673},{"loc":1675,"images":1679,"lastmod":687},[1680,1681],{"loc":1203},{"loc":1380},"blog\u002F2026\u002F10\u002Fhow-ai-reads-a-stock-chart","oH2xwoJGjea3JgXemVUfFbDP4tXnpO_lFqgvPsJTCXY",{"id":1685,"title":1686,"author":8,"body":1687,"coverImg":2233,"description":2234,"extension":683,"head":8,"meta":2235,"navigation":685,"path":2236,"publishedAt":2237,"schemaOrg":8,"seo":2238,"sitemap":2239,"stem":2243,"updatedAt":687,"__hash__":2244},"blog\u002Fblog\u002F2026\u002F10\u002Fbullish-and-bearish-chart-patterns.md","Bullish and Bearish Chart Patterns and Their Lookalikes",{"type":10,"value":1688,"toc":2216},[1689,1692,1695,1698,1702,1705,1714,1723,1727,1730,1817,1820,1824,1840,1843,1852,1855,1859,1862,1871,1874,1877,1886,1893,1897,1903,1912,1916,1919,1923,1934,1943,1946,1949,1953,1959,1968,1971,1975,1988,2061,2064,2090,2094,2097,2101,2109,2112,2116,2119,2122,2126,2135,2138,2142,2145,2148,2152,2155,2190,2197,2206,2208,2210,2213],[13,1690,1691],{},"In December 2024, a chart blog called Nvidia's daily chart a textbook head and shoulders. It is the most famous bearish shape in the book. Six weeks later, was it right?",[13,1693,1694],{},"The honest answer is that it depends on which day you asked. That is the whole story of bullish and bearish chart patterns in one stock: a shape the price has already drawn, a name for it, and a lot of room between the name and what happens next.",[13,1696,1697],{},"This guide walks through the patterns you will meet most often, each one beside the lookalike that fools people, with the numbers one practitioner has published on them and what academic studies found when they tested them.",[23,1699,1701],{"id":1700},"bullish-vs-bearish-what-the-labels-actually-claim","Bullish vs Bearish: What the Labels Actually Claim",[13,1703,1704],{},"A chart pattern is a shape drawn by the price over weeks or months: three peaks with the middle one highest, two dips at the same level, two lines closing in on each other. A bullish pattern is one traders say tends to come before a rise. A bearish one is said to come before a fall.",[13,1706,1707,1708,1713],{},"Notice the \"said to\". The label describes a tradition, not a law. Andrew Lo, Harry Mamaysky and Jiang Wang, writing in the Journal of Finance in 2000, put the core problem plainly: the shapes in a price chart are often ",[282,1709,1712],{"href":1710,"target":285,"rel":1711},"https:\u002F\u002Fwww.nber.org\u002Fpapers\u002Fw7613",[287],"in the eyes"," of the beholder. Two people can look at the same chart and see different patterns, or none.",[13,1715,1716,1719,1720,1722],{},[356,1717,1718],{},"So every useful pattern comes with a rule for when it counts."," Until the price crosses a specific line, you are looking at squiggles with a name attached. If you want every shape on one page, our ",[282,1721,532],{"href":531}," draws them side by side.",[23,1724,1726],{"id":1725},"five-patterns-and-the-lookalikes-that-fool-people","Five Patterns and the Lookalikes That Fool People",[13,1728,1729],{},"Here are the shapes this guide covers, what has to happen before each one counts, and the impostor that looks the same until it does not.",[31,1731,1732,1748],{},[34,1733,1734],{},[37,1735,1736,1739,1742,1745],{},[40,1737,1738],{},"Pattern",[40,1740,1741],{},"Taught as",[40,1743,1744],{},"Counts once the price",[40,1746,1747],{},"Its lookalike",[50,1749,1750,1764,1778,1791,1804],{},[37,1751,1752,1755,1758,1761],{},[55,1753,1754],{},"Head and shoulders top",[55,1756,1757],{},"Bearish",[55,1759,1760],{},"Closes below the neckline",[55,1762,1763],{},"Three bumps that bounce off the neckline and climb",[37,1765,1766,1769,1772,1775],{},[55,1767,1768],{},"Head and shoulders bottom",[55,1770,1771],{},"Bullish",[55,1773,1774],{},"Closes above the neckline",[55,1776,1777],{},"Three dips that roll over before the line",[37,1779,1780,1783,1785,1788],{},[55,1781,1782],{},"Double bottom",[55,1784,1771],{},[55,1786,1787],{},"Closes above the peak between the two lows",[55,1789,1790],{},"Two dips that never clear the middle peak",[37,1792,1793,1796,1798,1801],{},[55,1794,1795],{},"Descending triangle",[55,1797,1757],{},[55,1799,1800],{},"Closes outside either trendline",[55,1802,1803],{},"The same triangle breaking upward",[37,1805,1806,1809,1811,1814],{},[55,1807,1808],{},"Cup with handle",[55,1810,1771],{},[55,1812,1813],{},"Rises above the right lip of the cup",[55,1815,1816],{},"A breakout that hands the whole rise back",[13,1818,1819],{},"The right column is where most of the pain lives. Let's take them one at a time.",[23,1821,1823],{"id":1822},"head-and-shoulders-the-real-one-and-the-fake","Head and Shoulders: The Real One and the Fake",[13,1825,952,1826,1829,1830,1833,1834,1839],{},[356,1827,1828],{},"head and shoulders top"," is three peaks in a row, the middle one (the head) higher than the two beside it (the shoulders). A line under the two troughs between the peaks is the ",[356,1831,1832],{},"neckline",". The pattern is read as a rise turning into a fall, and practitioners have long called it one of the most reliable patterns, as a ",[282,1835,1838],{"href":1836,"target":285,"rel":1837},"https:\u002F\u002Fwww.newyorkfed.org\u002Fmedialibrary\u002Fmedia\u002Fresearch\u002Fstaff_reports\u002Fsr4.pdf",[287],"New York Fed"," staff report by Carol Osler and Kevin Chang notes.",[13,1841,1842],{},"The catch is the word \"complete\". The pattern only counts once the price closes below the neckline. Before that, it can just as easily bounce off the line and climb past the head. How often does the fake version happen? More often than the textbooks let on, as the computer test further down shows.",[13,1844,1845,1846,1851],{},"Thomas Bulkowski, who has cataloged chart patterns for decades on ",[282,1847,1850],{"href":1848,"target":285,"rel":1849},"https:\u002F\u002Fthepatternsite.com\u002Fhst.html",[287],"his pattern site",", adds a second check: the price must have something to reverse. Three bumps after a long climb mean more than three bumps in a market that has been going sideways for a year.",[13,1853,1854],{},"The mirror image, a head and shoulders bottom, has three troughs with the deepest in the middle and is read as bullish. The same rule applies upside down: it counts once the price closes above the neckline.",[23,1856,1858],{"id":1857},"a-real-head-and-shoulders-that-changed-its-mind-nvidia-late-2024","A Real Head and Shoulders That Changed Its Mind: Nvidia, Late 2024",[13,1860,1861],{},"This is a closed window from December 2024 to January 2025, read after the fact. It is not a view on Nvidia today.",[13,1863,1864,1865,1870],{},"On December 15, 2024, the Keystone Speculator blog called Nvidia's daily chart a textbook head and shoulders and ",[282,1866,1869],{"href":1867,"target":285,"rel":1868},"https:\u002F\u002Fthekeystonespeculator.blogspot.com\u002F2024\u002F12\u002Fnvda-nvidia-daily-chart-h-head-and.html",[287],"put numbers"," on it: a neckline near $132, a head near $150, and a measured target of $114 if the neckline gave way. On Christmas Eve the same blog noted the stock had dipped to $126.86 and bounced back to about $140.",[13,1872,1873],{},"Then the stock did two opposite things in three weeks.",[13,1875,1876],{},"On January 6, 2025, Nvidia closed at a record $149.43, Yahoo Finance's daily price history shows. That beat the head's best close of $148.88 on November 7. The right shoulder was supposed to stay below the head, and on a closing basis it had grown into a new high. Anyone grading the pattern that evening would have filed it under failed.",[13,1878,1879,1880,1885],{},"On January 27, after China's DeepSeek shook AI chip stocks, Nvidia fell 17% in a single day to $118.42, The Wall Street Journal reported in its ",[282,1881,1884],{"href":1882,"target":285,"rel":1883},"https:\u002F\u002Fwww.wsj.com\u002Flivecoverage\u002Fstock-market-today-dow-sp500-nasdaq-live-01-27-2025\u002Fcard\u002Fnvidia-stock-is-down-more-than-10-here-s-why--sZmsM8tvQFTS3iUBASHa",[287],"market coverage",". That was far below the $132 neckline and not far above the $114 target.",[13,1887,1888,1889,1892],{},"Read on January 6, the head and shoulders looked failed. Read on January 27, the price had fallen past the neckline toward the drawn target. A strict reader would say the December pattern died when the price made a new high, and the January drop was a new story driven by news. A loose reader would say the top just took its time. ",[356,1890,1891],{},"Both readings exist because the shape could not contain the thing that actually moved the price",": news that nobody drawing lines in mid-December could have known.",[23,1894,1896],{"id":1895},"double-bottom-or-just-squiggles","Double Bottom or Just Squiggles",[13,1898,952,1899,1902],{},[356,1900,1901],{},"double bottom"," is two lows at roughly the same price with a peak between them, like a W. It is read as bullish: sellers pushed the price down twice and could not break through.",[13,1904,1905,1906,1911],{},"Its lookalike is so common that Bulkowski gave it a blunt name. On his ",[282,1907,1910],{"href":1908,"target":285,"rel":1909},"https:\u002F\u002Fthepatternsite.com\u002Faadb.html",[287],"Adam and Adam"," double bottom page, he writes that the twin dips become a true double bottom only when the price closes above the peak between them. Until then, in his words, the squiggles on the chart are just squiggles.",[291,1913],{"src":1914,"alt":1915,"width":295,"height":296},"https:\u002F\u002Fmedia.primotrade.app\u002Fblog\u002Fbullish-and-bearish-chart-patterns\u002Fdouble-bottom-vs-squiggles-panel-2.jpg","Two W shapes side by side: on the left the price closes above the line at the middle peak and keeps rising, on the right it turns back below that line and falls under both lows.",[13,1917,1918],{},"His sample shows why waiting matters. If you buy before that confirmation, he puts the chance of the price continuing lower at 48%. That is close to a coin flip, on a pattern that is taught as bullish.",[23,1920,1922],{"id":1921},"triangles-same-shape-either-direction","Triangles: Same Shape, Either Direction",[13,1924,1925,1926,1929,1930,1933],{},"A triangle forms when the highs and lows start to converge between two trendlines. In a ",[356,1927,1928],{},"descending triangle",", the lower line is flat and the upper line slopes down. An ",[356,1931,1932],{},"ascending triangle"," is the mirror: a flat upper line and a rising lower one.",[13,1935,1936,1937,1942],{},"Descending triangles are usually taught as bearish. In Bulkowski's own sample, the price actually broke out upward 53% of the time, his ",[282,1938,1941],{"href":1939,"target":285,"rel":1940},"https:\u002F\u002Fthepatternsite.com\u002Fdt.html",[287],"triangle page"," shows. The shape on its own did not decide the direction.",[13,1944,1945],{},"He also lists what makes one count: at least three touches of one line and two of the other, and a close outside a line to confirm it. His page adds a sobering note: the pattern's performance has dropped almost in half since the 1990s. A shape that everyone knows tends to stop surprising anyone.",[13,1947,1948],{},"The lesson for a triangle is simple. Do not decide the direction from the slope of the lines. Let the close outside the triangle decide, and know in advance what you will do if it goes the other way.",[23,1950,1952],{"id":1951},"cup-with-handle-when-a-success-is-also-a-drop","Cup with Handle: When a Success Is Also a Drop",[13,1954,952,1955,1958],{},[356,1956,1957],{},"cup with handle"," is a rounded dip, like the side view of a teacup, followed by a smaller dip on the right (the handle). The breakout comes when the price rises above the right lip of the cup, and the pattern is read as bullish.",[13,1960,1961,1962,1967],{},"On paper it is one of the best performers Bulkowski tracks: a break-even failure rate of 5% across 913 trades. Then you read further down the ",[282,1963,1966],{"href":1964,"target":285,"rel":1965},"https:\u002F\u002Fthepatternsite.com\u002Fcup.html",[287],"same page",". In a separate study of 300 cups with handles from 1990 to March 2024, he found that 47% dropped substantially within two months of the upward breakout, and almost a quarter rose no more than 15% before dropping.",[13,1969,1970],{},"Both numbers can be true at once. The failure rate only asks whether the price ever cleared 5% after the breakout. A cup that rose 6% and then fell back below the breakout counts as a success in the first number and as a drop in the second. The target drawn from the cup's depth was reached by only 61% of the cups he measured.",[23,1972,1974],{"id":1973},"bulkowskis-numbers-with-the-fine-print-attached","Bulkowski's Numbers, With the Fine Print Attached",[13,1976,1977,1978,1981,1982,1987],{},"Bulkowski publishes a ",[356,1979,1980],{},"break-even failure rate"," for each pattern: the share of patterns where the price failed to move at least 5% in the expected direction after the breakout, as his ",[282,1983,1986],{"href":1984,"target":285,"rel":1985},"https:\u002F\u002Fthepatternsite.com\u002Fglossary.html",[287],"glossary defines"," it. Here are his bull-market figures for the patterns above.",[31,1989,1990,2005],{},[34,1991,1992],{},[37,1993,1994,1996,1999,2002],{},[40,1995,1738],{},[40,1997,1998],{},"Break-even failure rate",[40,2000,2001],{},"Reached the measured target",[40,2003,2004],{},"Sample, bull markets",[50,2006,2007,2020,2034,2048],{},[37,2008,2009,2011,2014,2017],{},[55,2010,1754],{},[55,2012,2013],{},"19%",[55,2015,2016],{},"51%",[55,2018,2019],{},"More than 2,800 trades",[37,2021,2022,2025,2028,2031],{},[55,2023,2024],{},"Descending triangle, up \u002F down breakout",[55,2026,2027],{},"22% \u002F 23%",[55,2029,2030],{},"64% \u002F 50%",[55,2032,2033],{},"More than 1,300 trades",[37,2035,2036,2039,2042,2045],{},[55,2037,2038],{},"Double bottom, Adam and Adam",[55,2040,2041],{},"16%",[55,2043,2044],{},"73%",[55,2046,2047],{},"1,154 trades",[37,2049,2050,2052,2055,2058],{},[55,2051,1808],{},[55,2053,2054],{},"5%",[55,2056,2057],{},"61%",[55,2059,2060],{},"913 trades",[13,2062,2063],{},"Read this table as a description of his past samples, measured on his own rules. It is not a forecast for the next chart you see, and it is not peer reviewed. Four details change how the numbers feel:",[550,2065,2066,2072,2078,2084],{},[553,2067,2068,2071],{},[356,2069,2070],{},"Bull markets only."," His glossary defines a bull market as everything outside the 2000 to 2002 and 2007 to 2009 bear markets, so these figures leave out the worst stretches.",[553,2073,2074,2077],{},[356,2075,2076],{},"Perfect trades."," He measures from the breakout to the ultimate high or low, which assumes buying at exactly the breakout and selling at exactly the top. No real trader does that.",[553,2079,2080,2083],{},[356,2081,2082],{},"Pullbacks hurt."," When the price comes back to the breakout level soon after breaking out, he notes, the move afterwards tends to be weaker.",[553,2085,2086,2089],{},[356,2087,2088],{},"Targets are rough."," Even on his own rules, the head and shoulders target was met about half the time.",[23,2091,2093],{"id":2092},"what-academic-studies-found","What Academic Studies Found",[13,2095,2096],{},"So did anyone test this properly? Three teams of researchers did, with computers doing the looking instead of eyes.",[827,2098,2100],{"id":2099},"some-information-not-a-money-machine","Some information, not a money machine",[13,2102,2103,2104,2108],{},"Lo, Mamaysky and Wang built a computer method to detect patterns objectively and ran it on a large number of US stocks from 1962 to 1996. They compared what happened after patterns like head and shoulders and double bottoms with what happened otherwise, and concluded that several of them provide ",[282,2105,2107],{"href":1710,"target":285,"rel":2106},[287],"incremental information"," and may have some practical value.",[13,2110,2111],{},"That is a careful phrase. It means the spread of daily price moves after a pattern looked different from the usual spread. The paper does not claim that trading the patterns paid off.",[827,2113,2115],{"id":2114},"head-and-shoulders-tested-by-computer","Head and shoulders, tested by computer",[13,2117,2118],{},"Osler and Chang tested a computer-identified head and shoulders rule on daily exchange rates of six currencies against the dollar from 1973 to 1994. It is a currency study, not a stock study, but the way it counts patterns applies to any chart.",[13,2120,2121],{},"Their program found plenty of head and shoulders shapes. Only 25 to 40 percent of them, depending on the currency, were confirmed by a break of the neckline within the time window their rules allowed. Most of the shapes did not finish in time to count. The rule had predictive power for the German mark and the Japanese yen but not for the other four. The same report found that another common claim, that the pattern's price target has predictive power, was not supported by the data.",[827,2123,2125],{"id":2124},"the-data-snooping-problem","The data snooping problem",[13,2127,2128,2129,2134],{},"Try enough rules on the same price history and some will look good by chance. Ryan Sullivan, Allan Timmermann and Halbert White tackled this in a Journal of Finance ",[282,2130,2133],{"href":2131,"target":285,"rel":2132},"https:\u002F\u002Fonlinelibrary.wiley.com\u002Fdoi\u002Fabs\u002F10.1111\u002F0022-1082.00163",[287],"paper from 1999"," that tested a large universe of trading rules on 100 years of daily Dow Jones data, adjusting for every rule tried, not only the one that looked best.",[13,2136,2137],{},"The same trap waits for anyone scanning charts by eye. The more shapes you look for, the more you will find, and some of them will have worked in the past by luck alone.",[23,2139,2141],{"id":2140},"how-to-read-a-pattern-without-fooling-yourself","How to Read a Pattern Without Fooling Yourself",[13,2143,2144],{},"None of this means patterns are useless. A head and shoulders tells you something true about the past: buyers tried to push past an earlier high and failed. That is context. The mistake is treating it as a forecast.",[13,2146,2147],{},"Most of the discipline is patience. A daily candle is not finished until the market closes, so the honest time to grade a pattern is the evening, not the middle of the session.",[291,2149],{"src":2150,"alt":2151,"width":295,"height":296},"https:\u002F\u002Fmedia.primotrade.app\u002Fblog\u002Fbullish-and-bearish-chart-patterns\u002Flaptop-after-the-close.jpg","An open laptop on a dark table in the evening, its screen showing a blurred line chart with a flat line across it.",[13,2153,2154],{},"Before you call a shape bullish or bearish, run through this:",[1082,2156,2157,2163,2169,2175,2181],{},[553,2158,2159,2162],{},[356,2160,2161],{},"Name the lookalike."," For every pattern you see, ask what the impostor would look like, and what would tell the two apart.",[553,2164,2165,2168],{},[356,2166,2167],{},"Wait for the close."," A pattern counts when the price closes beyond the neckline or trendline, not when it touches it. A screenshot taken during the trading day shows a last candle that is still forming, so any read of it, a photo read in Primo Trade included, can name the shape but cannot tell you whether it counts yet.",[553,2170,2171,2174],{},[356,2172,2173],{},"Check what came before."," A reversal pattern needs a trend to reverse.",[553,2176,2177,2180],{},[356,2178,2179],{},"Treat the target as a sketch."," Even on Bulkowski's rules, targets were met somewhere between half and three quarters of the time.",[553,2182,2183,2186,2187,2189],{},[356,2184,2185],{},"Decide where you are wrong before you start."," Our guide to ",[282,2188,1617],{"href":1616}," explains one way traders set that point in advance.",[13,2191,2192,2193,2196],{},"Patterns are built from candles, so if you are new to reading them, start with ",[282,2194,2195],{"href":516},"how to read candlestick charts",".",[13,2198,2199,2200,2202,2203,2205],{},"If you want a second read of a chart in front of you, ",[282,2201,104],{"href":1033}," can look at a photo or screenshot of it and name the chart pattern its AI sees, along with what that pattern usually signals. The same rules apply to that read: a name is an interpretation, not a confirmation, and our article on ",[282,2204,1042],{"href":543}," explains where it can go wrong.",[13,2207,270],{},[23,2209,1129],{"id":1128},[13,2211,2212],{},"Bullish and bearish chart patterns are names for shapes the price has already drawn. Each one has a twin that looks the same until the price closes past a line: the head and shoulders that bounces off its neckline, the double bottom that never clears its middle peak, the bearish triangle that breaks upward about half the time.",[13,2214,2215],{},"Practitioner counts show these patterns fail often enough to matter and reach their targets roughly half to three quarters of the time. Academic studies found a little information in some of them, in some markets, and warned that testing many shapes on the same history will always turn up a few that worked by luck. The Nvidia chart from the winter of 2024 shows the rest: the same pattern can look failed one week and confirmed three weeks later, depending on the day you stop reading. Use patterns to describe what the price has done, and weigh them as one piece of evidence among several.",{"title":663,"searchDepth":664,"depth":664,"links":2217},[2218,2219,2220,2221,2222,2223,2224,2225,2226,2231,2232],{"id":1700,"depth":664,"text":1701},{"id":1725,"depth":664,"text":1726},{"id":1822,"depth":664,"text":1823},{"id":1857,"depth":664,"text":1858},{"id":1895,"depth":664,"text":1896},{"id":1921,"depth":664,"text":1922},{"id":1951,"depth":664,"text":1952},{"id":1973,"depth":664,"text":1974},{"id":2092,"depth":664,"text":2093,"children":2227},[2228,2229,2230],{"id":2099,"depth":1143,"text":2100},{"id":2114,"depth":1143,"text":2115},{"id":2124,"depth":1143,"text":2125},{"id":2140,"depth":664,"text":2141},{"id":1128,"depth":664,"text":1129},"https:\u002F\u002Fmedia.primotrade.app\u002Fblog\u002Fbullish-and-bearish-chart-patterns\u002Fcover-panel.jpg","Bullish and bearish chart patterns beside the lookalikes that fool traders, with Bulkowski's failure rates, their samples, and a dated Nvidia example.",{},"\u002Fblog\u002F2026\u002F10\u002Fbullish-and-bearish-chart-patterns","2026-10-05",{"title":1686,"description":2234},{"loc":2236,"images":2240,"lastmod":687},[2241,2242],{"loc":1914},{"loc":2150},"blog\u002F2026\u002F10\u002Fbullish-and-bearish-chart-patterns","bfNCSDjtuWWbFQCMupPJ_2AXpK0fQ_pMUkCETk_ZC_0",{"id":2246,"title":2247,"author":8,"body":2248,"coverImg":2901,"description":2902,"extension":683,"head":8,"meta":2903,"navigation":685,"path":2904,"publishedAt":2905,"schemaOrg":8,"seo":2906,"sitemap":2907,"stem":2912,"updatedAt":687,"__hash__":2913},"blog\u002Fblog\u002F2026\u002F10\u002Fhow-to-read-candlestick-charts.md","How to Read Candlestick Charts Without Fooling Yourself",{"type":10,"value":2249,"toc":2870},[2250,2259,2262,2265,2269,2278,2281,2285,2289,2296,2299,2308,2312,2319,2325,2329,2336,2340,2343,2346,2349,2353,2362,2467,2470,2474,2477,2481,2490,2493,2496,2499,2503,2506,2512,2516,2519,2523,2528,2539,2542,2546,2552,2556,2559,2619,2622,2625,2629,2633,2640,2649,2652,2656,2660,2667,2676,2680,2689,2698,2701,2705,2708,2712,2761,2765,2774,2778,2790,2793,2797,2800,2847,2851,2860,2862,2864,2867],[13,2251,2252,2253,2258],{},"At 2:47 p.m. on May 6, 2010, shares of Procter & Gamble, about as sleepy a blue chip as Wall Street has, traded at $39.37, as ",[282,2254,2257],{"href":2255,"target":285,"rel":2256},"https:\u002F\u002Fwww.cnbc.com\u002F2010\u002F05\u002F06\u002Fstock-selloff-may-have-been-triggered-by-a-trader-error.html",[287],"CNBC reported"," that evening. A few minutes earlier they had been above $60. Soon after, they were back above $60 as if nothing had happened.",[13,2260,2261],{},"By the close, the whole episode had shrunk into one candle on P&G's daily chart: a small body sitting on a tail so long it looks like a printing error. If you know how to read candlestick charts, you can read that day in two seconds. If you know a little of the textbook, you may also be tempted to call it a hammer and wait for a bounce.",[13,2263,2264],{},"This guide is about the first skill, and about why the second deserves suspicion: a candle is an excellent record of what happened and a much weaker clue about what happens next.",[23,2266,2268],{"id":2267},"how-to-read-candlestick-charts-the-four-prices-in-every-candle","How to Read Candlestick Charts: The Four Prices in Every Candle",[13,2270,2271,2272,2277],{},"Each candle covers one period: a day, an hour, five minutes, whatever the chart is set to. To draw it, the chart needs four numbers for that period, the open, the high, the low and the close, as the ",[282,2273,2276],{"href":2274,"target":285,"rel":2275},"https:\u002F\u002Fchartschool.stockcharts.com\u002Ftable-of-contents\u002Fchart-analysis\u002Fcandlestick-charts\u002Fintroduction-to-candlesticks",[287],"StockCharts ChartSchool"," explains.",[13,2279,2280],{},"That is the whole alphabet. Everything else on a candlestick chart is those four numbers, drawn again and again.",[291,2282],{"src":2283,"alt":2284,"width":295,"height":296},"https:\u002F\u002Fmedia.primotrade.app\u002Fblog\u002Fhow-to-read-candlestick-charts\u002Fcandle-anatomy-panel.jpg","Diagram of a green up candle and a red down candle with the high, low, open, close, body and both wicks labelled",[827,2286,2288],{"id":2287},"the-body","The body",[13,2290,2291,2292,2295],{},"The thick part is the ",[356,2293,2294],{},"body",", sometimes called the real body. It runs from the open to the close.",[13,2297,2298],{},"If the close is above the open, the price rose during the period. Most charts draw that candle green or hollow, with the open at the bottom of the body and the close at the top. If the close is below the open, the price fell, and the candle is red or filled with the open at the top. Colors are a software setting, not a law, so check yours before you trust them.",[13,2300,2301,2302,2307],{},"Body length is distance. A long body means the price traveled a long way between open and close. A short body means the two ended up close together, which ",[282,2303,2306],{"href":2304,"target":285,"rel":2305},"https:\u002F\u002Fwww.cmegroup.com\u002Feducation\u002Fcourses\u002Ftechnical-analysis\u002Fchart-types-candlestick-line-bar",[287],"CME Group's course"," reads as not much strength in the move either way.",[827,2309,2311],{"id":2310},"the-wicks","The wicks",[13,2313,2314,2315,2318],{},"The thin lines above and below the body are the ",[356,2316,2317],{},"wicks",", also called shadows or tails. The tip of the upper wick is the highest price traded in the period, and the tip of the lower wick is the lowest.",[13,2320,2321,2324],{},[356,2322,2323],{},"Wicks show the part of the move that did not hold."," A long upper wick means buyers pushed the price up during the period and sellers dragged it back before the close. A long lower wick is the mirror case: sellers pushed it down and buyers brought it back. CME adds a useful reading of where the body sits: a small body near the top of the range means buyers had taken control by the end of the period, and a small body near the bottom means sellers had.",[827,2326,2328],{"id":2327},"when-there-are-no-wicks","When there are no wicks",[13,2330,2331,2332,2335],{},"A candle with no wicks is a ",[356,2333,2334],{},"marubozu",", a Japanese word the StockCharts dictionary translates as close-cropped. In a green marubozu the open was the low and the close was the high, so buyers held control from the first trade to the last. A red marubozu is the reverse.",[827,2337,2339],{"id":2338},"what-a-candle-hides","What a candle hides",[13,2341,2342],{},"ChartSchool makes a point most guides skip: two very different days can draw the same candle. A small dip, a big rally and a small fade can leave exactly the same open, high, low and close as a big rally, a crash and a full recovery. The candle keeps the four prices and throws away the order they came in, and it never tells you why any of it happened.",[13,2344,2345],{},"The same limit follows the candle into a photo. Snap a chart for Primo Trade's AI read, or for any other reader of chart images, and the picture carries the shapes, not the order of the trades inside each candle.",[13,2347,2348],{},"Which brings us back to P&G.",[23,2350,2352],{"id":2351},"read-a-real-candle-procter-gamble-on-may-6-2010","Read a Real Candle: Procter & Gamble on May 6, 2010",[13,2354,2355,2356,2361],{},"Here is P&G's daily price history around that afternoon, from ",[282,2357,2360],{"href":2358,"target":285,"rel":2359},"https:\u002F\u002Ffinance.yahoo.com\u002Fquote\u002FPG\u002Fhistory\u002F?period1=1272844800&period2=1273795200",[287],"Yahoo Finance",", with prices as traded rather than adjusted for dividends.",[31,2363,2364,2385],{},[34,2365,2366],{},[37,2367,2368,2371,2374,2377,2380,2382],{},[40,2369,2370],{},"Day",[40,2372,2373],{},"Open",[40,2375,2376],{},"High",[40,2378,2379],{},"Low",[40,2381,311],{},[40,2383,2384],{},"Volume",[50,2386,2387,2407,2427,2447],{},[37,2388,2389,2392,2395,2398,2401,2404],{},[55,2390,2391],{},"Wed, May 5",[55,2393,2394],{},"$61.62",[55,2396,2397],{},"$62.57",[55,2399,2400],{},"$61.50",[55,2402,2403],{},"$62.16",[55,2405,2406],{},"11.6 million",[37,2408,2409,2412,2415,2418,2421,2424],{},[55,2410,2411],{},"Thu, May 6",[55,2413,2414],{},"$61.91",[55,2416,2417],{},"$62.67",[55,2419,2420],{},"$39.37",[55,2422,2423],{},"$60.75",[55,2425,2426],{},"28.6 million",[37,2428,2429,2432,2435,2438,2441,2444],{},[55,2430,2431],{},"Fri, May 7",[55,2433,2434],{},"$60.61",[55,2436,2437],{},"$61.20",[55,2439,2440],{},"$60.00",[55,2442,2443],{},"$60.31",[55,2445,2446],{},"24.0 million",[37,2448,2449,2452,2455,2458,2461,2464],{},[55,2450,2451],{},"Mon, May 10",[55,2453,2454],{},"$61.57",[55,2456,2457],{},"$62.70",[55,2459,2460],{},"$61.16",[55,2462,2463],{},"$62.42",[55,2465,2466],{},"16.0 million",[13,2468,2469],{},"Read the May 6 candle out loud: opened at $61.91, went up to $62.67, down to $39.37, closed at $60.75. The close is below the open, so the body is red, and it is short, just $1.16. The upper wick is 76 cents. The lower wick runs $21.38, more than 18 times the length of the body, on about two and a half times the previous day's volume.",[827,2471,2473],{"id":2472},"the-shape-says-hammer","The shape says hammer",[13,2475,2476],{},"By shape alone, that is a textbook hammer or hanging man: a small body near the top, a long lower wick, almost nothing above. Now look left. P&G had closed at $62.68, $61.71 and $62.16 from Monday to Wednesday that week. There was no decline for a hammer to end and no rally for a hanging man to top, so by the book it is neither.",[827,2478,2480],{"id":2479},"the-cause-says-something-else","The cause says something else",[13,2482,2483,2484,2489],{},"That evening, CNBC's sources had a culprit: a trader who typed \"b\" for billion instead of \"m\" for million, in a trade possibly involving P&G. It was a neat story about one stock. The SEC and CFTC staffs' ",[282,2485,2488],{"href":2486,"target":285,"rel":2487},"https:\u002F\u002Fwww.sec.gov\u002Fnews\u002Fstudies\u002F2010\u002Fmarketevents-report.pdf",[287],"joint report",", published that September, told a bigger one.",[13,2491,2492],{},"The selling started in the futures market, where a large trader began a program at 2:32 p.m. to sell 75,000 E-Mini S&P 500 contracts, worth about $4.1 billion, into an already stressed market. P&G's own slide did not begin until 2:44 p.m., after the broad indices had started their drop at about 2:40. The stock was a casualty, not the cause.",[13,2494,2495],{},"Once it started, P&G fell from more than $60 to $39.37 in about three and a half minutes, a 36.14% drop from its 2:40 p.m. price, then recovered above $60 in about a minute.",[13,2497,2498],{},"It had plenty of company. More than 20,000 trades across over 300 securities were executed more than 60% away from their prices moments before, and those trades were later cancelled. P&G's drop stopped short of that line, so its trades stood, and the $39.37 low is still printed on its chart.",[827,2500,2502],{"id":2501},"what-came-next","What came next",[13,2504,2505],{},"What followed was not much. P&G closed at $60.31 the next day and $62.42 the Monday after, right back in the range it had traded in all week. Anyone who read the long wick as buyers defending the stock, or as the start of a turn, was reading a story into a market malfunction.",[13,2507,2508,2511],{},[356,2509,2510],{},"The candle recorded what happened perfectly. It said nothing about why."," That is the whole method in one example: read the four prices precisely, then ask what else you would need to know before they mean anything.",[23,2513,2515],{"id":2514},"candlestick-patterns-built-from-one-candle","Candlestick Patterns Built From One Candle",[13,2517,2518],{},"Traders give names to candles with a recognizable shape. Each name describes the period that just ended, and most of them depend on what came before.",[827,2520,2522],{"id":2521},"doji","Doji",[13,2524,952,2525,2527],{},[356,2526,2521],{}," forms when the open and the close are virtually equal, so the body shrinks to a line and the candle looks like a cross or a plus sign. StockCharts calls a doji neutral on its own: any reading of it depends on the price action before it and on what happens after.",[13,2529,2530,2531,2534,2535,2538],{},"Two variants have their own names. A ",[356,2532,2533],{},"dragonfly doji"," has a long lower wick and no upper one, like a letter T. A ",[356,2536,2537],{},"gravestone doji"," is the opposite, a long upper wick with nothing below.",[13,2540,2541],{},"Where you see one matters more than its name. A doji on a five-minute chart can be nothing more than a quiet stretch before lunch. A doji on a weekly chart means a whole week of buying and selling ended almost exactly where it began. Same cross, very different amount of argument behind it.",[827,2543,2545],{"id":2544},"spinning-top","Spinning top",[13,2547,952,2548,2551],{},[356,2549,2550],{},"spinning top"," has a small body with wicks on both sides that are longer than the body. Buyers and sellers both moved the price during the period and neither held the move, so it is read as indecision.",[827,2553,2555],{"id":2554},"hammer-hanging-man-and-their-upside-down-twins","Hammer, hanging man and their upside-down twins",[13,2557,2558],{},"These are the lookalikes that catch everyone out. The shape is identical; only the trend before it changes the name.",[31,2560,2561,2577],{},[34,2562,2563],{},[37,2564,2565,2568,2571,2574],{},[40,2566,2567],{},"Shape",[40,2569,2570],{},"After a decline",[40,2572,2573],{},"After a rise",[40,2575,2576],{},"What the period recorded",[50,2578,2579,2593,2607],{},[37,2580,2581,2584,2587,2590],{},[55,2582,2583],{},"Small body, long lower wick, little or no upper wick",[55,2585,2586],{},"Hammer",[55,2588,2589],{},"Hanging man",[55,2591,2592],{},"The price was pushed well down and came back by the close",[37,2594,2595,2598,2601,2604],{},[55,2596,2597],{},"Small body, long upper wick, little or no lower wick",[55,2599,2600],{},"Inverted hammer",[55,2602,2603],{},"Shooting star",[55,2605,2606],{},"The price was pushed well up and fell back by the close",[37,2608,2609,2612,2614,2616],{},[55,2610,2611],{},"Open and close almost equal",[55,2613,2522],{},[55,2615,2522],{},[55,2617,2618],{},"Both sides moved the price and the close ended near the open",[13,2620,2621],{},"StockCharts treats all four of the named reversal candles as needing confirmation from the candles that follow. For the shooting star it adds that the upper wick should be at least twice the length of the body before the shape means much. P&G's lower wick was 18 times its body, which shows that a dramatic wick on its own proves nothing.",[13,2623,2624],{},"A long wick in the middle of a flat stretch, like the one below, is exactly the case the book leaves without a name.",[291,2626],{"src":2627,"alt":2628,"width":295,"height":296},"https:\u002F\u002Fmedia.primotrade.app\u002Fblog\u002Fhow-to-read-candlestick-charts\u002Flong-lower-wick-panel.jpg","Illustration of a flat row of small candlesticks in which one candle has a lower wick far longer than any other",[23,2630,2632],{"id":2631},"candlestick-patterns-built-from-two-candles","Candlestick Patterns Built From Two Candles",[13,2634,2635,2636,2639],{},"An ",[356,2637,2638],{},"engulfing pattern"," is a small candle followed by a larger one whose body completely covers the first body and closes in the opposite direction. At the end of a decline, a green candle engulfing a red one is called bullish engulfing. At the end of a rise, a red candle engulfing a green one is called bearish engulfing.",[13,2641,2642,2643,2648],{},"Only the bodies count here. Per the ",[282,2644,2647],{"href":2645,"target":285,"rel":2646},"https:\u002F\u002Fchartschool.stockcharts.com\u002Ftable-of-contents\u002Fchart-analysis\u002Fcandlestick-charts\u002Fcandlestick-pattern-dictionary",[287],"StockCharts pattern dictionary",", the second body has to cover the first open and close, not its whole range from high to low.",[13,2650,2651],{},"Picture it at the end of a decline. Day one is a small red candle whose upper wick spikes well above its body. Day two opens below that red close and finishes above the red open, while day one's wick still pokes out above the green body. That still counts as bullish engulfing. The wick is the part of day one's move that did not hold, and the pattern compares bodies, not extremes.",[291,2653],{"src":2654,"alt":2655,"width":295,"height":296},"https:\u002F\u002Fmedia.primotrade.app\u002Fblog\u002Fhow-to-read-candlestick-charts\u002Fengulfing-vs-harami-panel.jpg","Diagram of a bullish engulfing pair, where a green body covers the red body before it, beside a harami, where a small green body sits inside a tall red one",[827,2657,2659],{"id":2658},"the-harami-lookalike","The harami lookalike",[13,2661,2662,2663,2666],{},"The lookalike runs the other way. In a ",[356,2664,2665],{},"harami",", the same dictionary says, the small body comes second and sits completely inside the previous body. One candle swallows, the other gets swallowed, so mixing up the order turns one name into the other.",[13,2668,2669,2670,2672,2673,2675],{},"Shapes built across many candles, such as head and shoulders or triangles, are a different family. Our guide to ",[282,2671,445],{"href":444}," covers them, and the ",[282,2674,532],{"href":531}," puts the common shapes on one page.",[23,2677,2679],{"id":2678},"where-candlesticks-came-from","Where Candlesticks Came From",[13,2681,2682,2683,2688],{},"Candlesticks are usually traced to Munehisa Homma, a rice merchant from Sakata who traded on the Dojima rice exchange in Osaka in the eighteenth century and died in 1803. His birth year is given as 1716 in a ",[282,2684,2687],{"href":2685,"target":285,"rel":2686},"https:\u002F\u002Fcatalogimages.wiley.com\u002Fimages\u002Fdb\u002Fpdf\u002F9781118732922.excerpt.pdf",[287],"Wiley book excerpt"," and as 1724 elsewhere.",[13,2690,2691,2692,2697],{},"The attribution is shakier than most guides admit. Steve Nison, who brought the charts to Western traders, wrote in his 1994 book Beyond Candlesticks that, based on his research, it is unlikely Homma used candle charts, as the ",[282,2693,2696],{"href":2694,"target":285,"rel":2695},"https:\u002F\u002Fen.wikipedia.org\u002Fwiki\u002FCandlestick_chart",[287],"Wikipedia history of the chart"," quotes him. He thought they more likely appeared in the early Meiji period, in the late 1800s. In Japan the chart is called ashi, meaning leg or foot.",[13,2699,2700],{},"Nison's 1991 book Japanese Candlestick Charting Techniques changed how the West drew charts. Until the late 1980s, real-time chart services in the US and Europe offered only bar charts. Within about two years of the book, almost every charting package offered candlesticks, according to the same Wiley excerpt.",[23,2702,2704],{"id":2703},"what-candlesticks-do-and-do-not-predict","What Candlesticks Do and Do Not Predict",[13,2706,2707],{},"Candlestick patterns are easy to name and hard to test, because a test needs a rule for when to get in and when to get out. Researchers have tried on several markets, and the results do not agree.",[827,2709,2711],{"id":2710},"tests-that-came-up-empty","Tests that came up empty",[550,2713,2714,2732,2749],{},[553,2715,2716,2719,2720,2725,2726,2731],{},[356,2717,2718],{},"US large caps:"," Ben Marshall tested candlestick strategies on the stocks of the Dow Jones Industrial Average from 1992 to 2002 in his ",[282,2721,2724],{"href":2722,"target":285,"rel":2723},"https:\u002F\u002Fmro.massey.ac.nz\u002Fhandle\u002F10179\u002F1604",[287],"2005 doctoral thesis",", which shares its title with a 2006 Journal of Banking and Finance paper he wrote with Martin Young and Lawrence Rose. He found no evidence that a trader following them would do better than the market. In a ",[282,2727,2730],{"href":2728,"target":285,"rel":2729},"https:\u002F\u002Fideas.repec.org\u002Fa\u002Fris\u002Fjofitr\u002F0834.html",[287],"2007 paper",", the three concluded that basing trading decisions solely on these techniques does not seem sensible, while leaving open that candlesticks could complement other methods.",[553,2733,2734,2737,2738,2743,2744,2196],{},[356,2735,2736],{},"Tokyo:"," a ",[282,2739,2742],{"href":2740,"target":285,"rel":2741},"https:\u002F\u002Fideas.repec.org\u002Fa\u002Fkap\u002Frqfnac\u002Fv31y2008i2p191-207.html",[287],"2008 study"," by Marshall, Young and Rochester Cahan in the Review of Quantitative Finance and Accounting found the same in the Japanese stock market from 1975 to 2004: no added value over the full 30 years, in any ten-year stretch, or in bull and bear markets, according to ",[282,2745,2748],{"href":2746,"target":285,"rel":2747},"https:\u002F\u002Fcas.xmu.edu.cn\u002Finfo\u002F1020\u002F2092.htm",[287],"Young's seminar abstract",[553,2750,2751,2754,2755,2760],{},[356,2752,2753],{},"Intraday:"," Matthieu Duvinage, Paolo Mazza and Mikael Petitjean tested dozens of candlestick rules on five-minute data for the 30 Dow stocks, published in ",[282,2756,2759],{"href":2757,"target":285,"rel":2758},"https:\u002F\u002Fideas.repec.org\u002Fa\u002Ftaf\u002Fquantf\u002Fv13y2013i7p1059-1070.html",[287],"Quantitative Finance"," in 2013. About a third of the rules beat simply holding the stocks before trading costs, but once costs were counted, the authors found no evidence that any did better.",[827,2762,2764],{"id":2763},"a-test-that-found-something","A test that found something",[13,2766,2767,2768,2773],{},"In Taiwan, Tsung-Hsun Lu, Yung-Ming Shiu and Tsung-Chi Liu found that three bullish two-day reversal patterns held up on the stocks of the Taiwan Top 50 Tracker Fund from 2002 to 2008, including in an out-of-sample test, in a ",[282,2769,2772],{"href":2770,"target":285,"rel":2771},"https:\u002F\u002Fideas.repec.org\u002Fa\u002Feee\u002Frevfin\u002Fv21y2012i2p63-68.html",[287],"2012 paper"," in the Review of Financial Economics. Same kind of pattern, different market, different answer.",[827,2775,2777],{"id":2776},"why-the-studies-disagree","Why the studies disagree",[13,2779,2780,2781,2786,2787],{},"One detail explains much of the disagreement. A 2015 study by Tsung-Hsun Lu, Yi-Chi Chen and Yu-Chin Hsu ran three-day reversal patterns on Dow stocks and found that eight of them ",[282,2782,2785],{"href":2783,"target":285,"rel":2784},"https:\u002F\u002Fresearchoutput.ncku.edu.tw\u002Fen\u002Fpublications\u002Ftrend-definition-or-holding-strategy-what-determines-the-profitab",[287],"passed or failed"," depending on the holding rule. With one exit rule they held up after a 0.5% trading cost and a correction for data snooping; with the exit rule Marshall, Young and Rose had used, they did not. ",[356,2788,2789],{},"The pattern alone did not decide the result; the exit did.",[13,2791,2792],{},"So here is the honest limit, plainly. A candlestick tells you what happened in a period, clearly and quickly. Whether a named pattern says anything reliable about the next period is still argued over, and the answer seems to depend on the market, the timeframe and the rules wrapped around it.",[23,2794,2796],{"id":2795},"how-to-practice-reading-candles","How to Practice Reading Candles",[13,2798,2799],{},"Reading candles is a habit more than a body of knowledge. These six steps are the habit.",[1082,2801,2802,2808,2814,2820,2832,2838],{},[553,2803,2804,2807],{},[356,2805,2806],{},"Say the four prices first."," Before naming anything, read the candle out loud: opened here, went up to here, down to here, closed here.",[553,2809,2810,2813],{},[356,2811,2812],{},"Look left."," A hammer and a hanging man are the same shape. The trend before the candle gives it a name, and a flat stretch, like P&G's week, gives it none.",[553,2815,2816,2819],{},[356,2817,2818],{},"Wait for the next candle."," Most single-candle patterns are only treated as signals once later candles confirm them.",[553,2821,2822,2825,2826,2831],{},[356,2823,2824],{},"Check the timeframe."," As with the doji above, the same shape covers very different amounts of trading on different charts. Gregory Morris argued that intraday candles lack the overnight break that gives the open and close their weight, a point a ",[282,2827,2830],{"href":2828,"target":285,"rel":2829},"https:\u002F\u002Fwww.econ.sinica.edu.tw\u002F~econ\u002FpdfPaper\u002F14-A010.pdf",[287],"2015 working paper"," from Academia Sinica quotes.",[553,2833,2834,2837],{},[356,2835,2836],{},"Ask why before asking what next."," If the whole market lurched that day, as it did on May 6, 2010, the candle may be telling you about the market rather than the stock.",[553,2839,2840,2843,2844,2846],{},[356,2841,2842],{},"Plan the exit before the entry."," The research above suggests the exit rule shapes the result as much as the pattern does. Our guide to ",[282,2845,1617],{"href":1616}," explains one common way traders set it.",[827,2848,2850],{"id":2849},"a-second-opinion-on-a-chart","A second opinion on a chart",[13,2852,2853,2854,2856,2857,2859],{},"If you want a second opinion on a chart you are looking at, ",[282,2855,104],{"href":1033}," can read it from a photo or a screenshot and give the AI's view of the trend and the chart pattern it sees. Like any reader of a chart image, it sees the shape, not the reason behind it. Our article on ",[282,2858,1042],{"href":543}," explains what that read can and cannot do.",[13,2861,270],{},[23,2863,1129],{"id":1128},[13,2865,2866],{},"Every candle is four prices: open, high, low and close. The body shows how far the price moved between open and close, and the wicks show how far it went and came back. Named patterns like the doji, the hammer and the engulfing pattern are shorthand for those shapes, and each one depends on the trend before it and the candles after it.",[13,2868,2869],{},"P&G's candle on May 6, 2010 is the reminder to keep close. The shape was perfect, the story behind it was a market malfunction, and nothing about the candle could tell the two apart. Most tests on US and Japanese stocks found that trading on candlestick patterns alone did not hold up. One study of Dow stocks found that a set of three-day patterns passed under one exit rule and failed under another, so treat a pattern as one piece of evidence rather than an answer.",{"title":663,"searchDepth":664,"depth":664,"links":2871},[2872,2878,2883,2888,2891,2892,2897,2900],{"id":2267,"depth":664,"text":2268,"children":2873},[2874,2875,2876,2877],{"id":2287,"depth":1143,"text":2288},{"id":2310,"depth":1143,"text":2311},{"id":2327,"depth":1143,"text":2328},{"id":2338,"depth":1143,"text":2339},{"id":2351,"depth":664,"text":2352,"children":2879},[2880,2881,2882],{"id":2472,"depth":1143,"text":2473},{"id":2479,"depth":1143,"text":2480},{"id":2501,"depth":1143,"text":2502},{"id":2514,"depth":664,"text":2515,"children":2884},[2885,2886,2887],{"id":2521,"depth":1143,"text":2522},{"id":2544,"depth":1143,"text":2545},{"id":2554,"depth":1143,"text":2555},{"id":2631,"depth":664,"text":2632,"children":2889},[2890],{"id":2658,"depth":1143,"text":2659},{"id":2678,"depth":664,"text":2679},{"id":2703,"depth":664,"text":2704,"children":2893},[2894,2895,2896],{"id":2710,"depth":1143,"text":2711},{"id":2763,"depth":1143,"text":2764},{"id":2776,"depth":1143,"text":2777},{"id":2795,"depth":664,"text":2796,"children":2898},[2899],{"id":2849,"depth":1143,"text":2850},{"id":1128,"depth":664,"text":1129},"https:\u002F\u002Fmedia.primotrade.app\u002Fblog\u002Fhow-to-read-candlestick-charts\u002Fcover-v2.jpg","How to read candlestick charts: the four prices in every candle, the doji, hammer and engulfing patterns, a real 2010 candle, and what the research found.",{},"\u002Fblog\u002F2026\u002F10\u002Fhow-to-read-candlestick-charts","2026-10-04",{"title":2247,"description":2902},{"loc":2904,"images":2908,"lastmod":687},[2909,2910,2911],{"loc":2283},{"loc":2627},{"loc":2654},"blog\u002F2026\u002F10\u002Fhow-to-read-candlestick-charts","h_ARc3WAneYN3CrEWRx5cBa5Z11Y6UDtLE0SG3_5gTk",41,1794742768361]