Chart Patterns
Getting Started in Chart Patterns
If you only buy one Bulkowski, buy this one. It carries the same measured numbers as his Encyclopedia plus the part the Encyclopedia leaves out, which is how he actually enters, stops out, and exits, and it has the busted-pattern chapter, the best idea in either book.
The one idea
The Encyclopedia of Chart Patterns counts. This book chooses. Bulkowski spent 1035 pages there putting a measured base rate on 63 patterns in alphabetical order and refusing to tell you which ones to trade. Here he ranks them and hands you a short list: the ten best bullish patterns, in order, with the numbers attached. High and tight flag first (69% average rise, 0% break-even failure across 253 patterns), then pipe bottoms on the weekly chart (45%, 5%), inverted and ascending scallops (43%, 4%), three rising valleys (41%, 5%), rounding bottoms and descending triangles tied for fifth, ascending broadening wedges and Eve and Eve double bottoms tied for seventh, triple bottoms and head-and-shoulders bottoms tied for ninth. The descending triangle entry is the one that stops people: 64% of them break out downward, and the whole trade is patiently waiting for the 36% that do not, because those rise 47% on average, the second best number in the book.
The second idea is the one with no equivalent in the 2005 Encyclopedia at all, and it is worth the price on its own. Bulkowski calls a pattern busted when price travels less than 10% after the breakout and then reverses and comes out the other side. Chapter 9 covers 18 of them with sample counts. A descending triangle that breaks down, goes nowhere, and closes back above its highest peak rose 60% on average across 180 samples. Busted double tops rose 74% across 62. Busted head-and-shoulders tops rose 53% across 68. Ordinary upward breakouts in the same bull market averaged 36%. The failed pattern beat the working pattern, and it comes with a clean entry trigger, which is a close past the pattern's own high or low.
The third thing is plumbing, and it is the reason a beginner can trade off this book and cannot trade off the Encyclopedia. Trendlines get measured instead of asserted: long ones beat short ones (43% versus 33% average rise around a 139-day median), widely spaced touches beat clustered ones, five touches beat three (57% versus 33%), shallow angles of 30 to 45 degrees beat steep ones. Support and resistance gets measured: round numbers, old peaks, congestion zones, all with stopping rates. Throwbacks get measured across 12,256 patterns. And stops get a formula you can run in a spreadsheet. The Encyclopedia tells you which patterns are good. This one tells you what to do on Monday at 9:31.
What it actually teaches
The book is a fictional dialogue, but a real six-step trade sequence runs underneath it, spelled out in the Chapter 11 checklists.
CHECK THE MARKET AND THE INDUSTRY FIRST
Before he looks at a single pattern he reviews the S&P 500, the Dow industrials, transports, utilities and the NASDAQ composite, then flips to the weekly chart to strip out daily noise. Buy upward breakouts only when the market is rising, and sit in cash otherwise. His own numbers make the case: ascending triangles rose 35% on average after an upward breakout in a bull market but only 30% in a bear market, and fell 24% in a bear market against 19% in a bull. He then ranks roughly 35 industries daily by six-month price change and trades only from the top 10 and from industries climbing the list fast.
PICK FROM THE SHORT LIST, NOT THE WHOLE ZOO
Chapter 6 gives ten bullish patterns in performance order and Chapter 12 gives a bull-market table for about 50. Each pattern comes with identification rules tight enough to use. A high and tight flag needs price up at least 90% in two months or less, then a pause, then a close above the pattern high, and its measure rule (half the run-up added to the flag's low) hits 90% of the time. A pipe bottom is two adjacent downward weekly spikes with heavy volume, a large overlap, and a median gap of just 24 cents between the two lows. Selection inside a pattern matters as much as picking the pattern: high and tight flags with a throwback rose 49% while those without rose 100%, and loose ones rose 50% against 85% for tight ones.
MAP THE CEILING BEFORE YOU COUNT THE PROFIT
Chapter 4 puts stopping rates on every kind of resistance, which no other beginner book does. Old peaks and valleys halt 19% to 27% of breakouts. Horizontal congestion zones halt 30% of upward breakouts and 35% of downward ones. Round numbers such as 10, 15 and 20 halt 22% of upward breakouts within fifty cents and 42% within a dollar. In a 38-stock study covering 1999 to 2004, 56% of valleys, 63% of peaks and 73% of congestion zones sat on a round number. The practical rule: resistance closer than about 5% usually gets punched through, and the danger band is 10% to 20% out, where 52% of stalls happen and most cluster near 15%.
BUY A PENNY ABOVE THE BREAKOUT
Wait for the close outside the pattern, then use a buy stop one cent past the breakout price rather than a market order the next morning. He says price moves most in the first week after a breakout, so a late fill eats the trade, and premature breakouts (price pokes out then comes straight back) run only 3% to 22% in triangles, which is a small price for getting in on time. Buy near the yearly high, not the yearly low. Of the 29 upward-breakout patterns he charts, 13 perform best when the breakout sits in the top third of the yearly range and only 9 in the bottom third. Downward breakouts flip: 14 of 25 do best in the bottom third.
SET A VOLATILITY STOP, THEN RATCHET IT
Take the daily high minus low for each of the last 30 days, average them, multiply by 1.5, and subtract that from the current low. His worked example: the 30-day average range was 21 cents, so 1.5 times 0.21 is 32 cents, and against a 7.25 breakout the stop goes no closer than 6.94. For stocks under $20 he uses 2.0 instead of 1.5. Then raise the stop below each new valley as price makes new highs, and never lower it. Expect the retrace before it arrives: across 12,256 patterns price threw back to the breakout 53% of the time, turned a median of three days out, and took about ten days for the round trip. Pullbacks after downward breakouts ran 56% across 10,878 patterns, and a high-volume breakout throws back 70% of the time.
WHEN THE PATTERN FAILS, TAKE THE OTHER SIDE
A pattern is busted when price moves less than 10% after the breakout and then reverses. Do not act on the retrace itself, because throwbacks and pullbacks look identical early. Wait for a close above the pattern's highest peak to go long, or below its lowest valley to go short. The payoffs across a bull market: busted double tops rose 74% on 62 samples, busted descending triangles 60% on 180, busted head-and-shoulders tops 53% on 68, busted symmetrical triangles 43% on 97, busted rising wedges 43% on 84. He is blunt that these are perfect trades and that some of the busted broadening patterns rest on five or twenty samples.
What it looks like Monday morning
You stop hunting every shape on the chart and hunt ten. Then you add the two checks almost nobody does before entry. First, you find the ceiling: the nearest old peak, congestion block or round number, and if the only real room is under 5% you skip the trade instead of taking it and hoping. Second, you compute the stop instead of picking a number that feels safe, which for most people means the volatility formula puts the stop further away than they wanted and tells them to trade smaller. You also change how you enter, replacing the next-morning market order with a buy stop a penny above the breakout price, and you write down in advance that price probably comes back to your entry inside ten days so you do not sell into it.
The larger change is that failure becomes a setup. A pattern that breaks the wrong way used to be a loss you forgot about. Now the pattern that breaks down, travels less than 10%, and then closes back above its own highest peak is a trade you take in the opposite direction, with the entry trigger and the stop already defined before you look. One more thing changes on Monday: you start disqualifying stocks. Anything that had a dead-cat bounce in the last six to twelve months comes off your list, because in his sample of roughly 700 of them, 26% got hit with another drop of at least 15% within three months and 38% within six.
| Pattern that broke down, then reversed | Average rise after the bust | Samples |
|---|---|---|
| Double top (Eve and Eve) | +74% | 62 |
| Descending triangle | +60% | 180 |
| Broadening top | +56% | 54 |
| Head-and-shoulders top | +53% | 68 |
| Rectangle top | +52% | 22 |
| Ascending triangle | +45% | 46 |
| Symmetrical triangle | +43% | 97 |
| Rising wedge | +43% | 84 |
Ordinary upward breakouts in the same bull market averaged 36%, so the failed pattern outperformed the working one, on perfect trades with no costs.
Unwillingness to use a stop is what separates an amateur trader from a professional.Thomas N. Bulkowski, Getting Started in Chart Patterns
Where it fails
- It sells you two more books, and the checklist does not work without them. The Before Buying checklist in Chapter 11 has a step that reads 'Score the chart pattern,' and the scoring system lives in Trading Classic Chart Patterns, not here. The General Trading Tips say tall patterns are 'the single best predictor of chart pattern performance,' then send you to the Encyclopedia for the tall and short thresholds, because this book does not print them. Chapter 5 and Chapter 12 both tell you to buy the Encyclopedia for the full bull and bear numbers. Three of his own most-used inputs are deliberately withheld.
- There is no bear-market column anywhere in the summary tables. Chapter 12 opens by stating the tables are from bull markets only. That is the one axis where the Encyclopedia genuinely wins, and it matters if you ever plan to short. Even in the Encyclopedia the bear market is a single event, March 24 2000 to October 10 2002, with everything else in the 1991 to 2004 window classed as bull. Either way, 2008, 2020 and 2022 are not in this data, and no measurement here has been re-run against a market with algorithmic execution or sub-penny spreads.
- The research that makes this book different runs on small samples. The throwback study is 12,256 patterns and is solid. The material you cannot get elsewhere is nowhere near that. The trendline chapter, the single best chapter in the book, rests on roughly 200 trendlines: 85 downsloping lines with three touches against 40 with five touches. The 1-2-3 trend change method was tested on 67 upsloping and 101 downsloping lines. The claim that the most powerful triangle breakouts come 80% to 85% of the way to the apex rests on 31 samples, which he says. Busted ascending broadening wedges with upward breakouts: five samples.
- The 1-2-3 trend change method fails its own test and he keeps it anyway. For upsloping trendlines, only 43% of the 67 qualifying lines actually delivered the 20% decline the method is supposed to predict. For downsloping lines the hit rate is a respectable 73%, but the average rise that follows is 35%, which is lower than the 38% posted by all the trendlines that ignored the 1-2-3 filter entirely. So the screen made you more certain and less profitable. He prints both numbers and then writes that he considers the method to add value.
- The headline numbers were already drifting on the day it printed. The high and tight flag leads the whole ranking on a 0% break-even failure rate across 253 patterns. In the same paragraph he adds, in parentheses, that since writing it he has found high and tight flags that moved less than 5%. Every figure in the book is a perfect trade too: bought at the exact breakout, sold at the ultimate high, no commissions, no slippage. He says so repeatedly. Use the numbers to rank setups against each other, never as expected return.
- Some of the measurement is not normalized for price. The round-number research reports that 42% of upward breakouts stop within a dollar of a round number. A dollar is 10% of a $10 stock and 1% of a $100 stock, and the study does not separate them. He acknowledges elsewhere that stocks under $20 are more volatile, which is exactly the population that would inflate that hit rate. The 22% within fifty cents is the more usable figure, and it is a lot less impressive.
- Two hundred of the pages are a novel. The book is framed as Bulkowski teaching a fictional student named Jake Murphy, who is bleeding money on health insurance premiums and who, in the epilogue, arrives by helicopter with Dom Perignon after buying the insurance company and firing its executives. Jake breaks the fourth wall to tell jokes about peanuts. If that reads as charming you will get through it fast. If it reads as filler, the actual content compresses to about eighty pages and you will resent the rest.
- It assumes daily bars, US stocks, and a lot of screen time. Every number comes from individual US stocks on daily or weekly charts with real exchange volume. Futures, forex and crypto traders lose the volume analysis, the gap analysis, the event patterns and most of the support and resistance work. His own routine is reviewing 350 securities across roughly 35 industries every day, holding under 10 positions, and timing entries on one and five minute charts. There is no position sizing, no risk of ruin math, and no portfolio construction anywhere in the book. His stated lifetime win rate is 49%, which tells you the whole edge lives in the exit rules he does give you.
- This edition has not been revised, and the buy button may not point at it. Everything above comes from the 2006 first edition, whose data stops in 2004. A later edition exists and Bulkowski re-measured his statistics for the 2021 third edition of the Encyclopedia, so specific thresholds quoted here (the 11.64% tall pipe cutoff, the 15-day flag width, the 47% descending triangle rise) may read differently in whatever copy arrives. Check any number you intend to trade against the edition in your hands. He also wrote a separate later book on what happens after the buy, which means the busted-pattern chapter here, the main reason to own this one, is the 2006 version of that idea.
Who it is for
Buy it if
You trade or want to trade US stocks on daily bars, and you want one book that both teaches chart patterns and puts measured odds on them. This is the Bulkowski you read cover to cover, and it is the only place he lays out the full sequence: market check, short list, resistance map, entry a penny above the breakout, volatility stop, trail, and the busted-pattern trade when it all goes wrong. Buy it especially if you already own the Encyclopedia and have never opened it, because this is the book that would have told you what to do with it.
Skip it if
Futures, forex and crypto traders get close to nothing, since every measurement comes from individual US stocks and leans on stock volume, gaps and earnings events. Skip it if you need per-pattern bear-market numbers or the tall and short thresholds, because those were deliberately left out to sell you the Encyclopedia. And skip it if you want position sizing, risk management math or a complete system, because there is none of that here, only entries, stops and exits.
As an Amazon Associate I earn from qualifying purchases. It costs you nothing extra, and it does not change what gets recommended or what the verdict says.
Reviewed here is the first edition (2006), data ending 2004. The link goes to the second edition. Specific thresholds quoted above, including the tall-pipe cutoff and the flag width, should be checked against the copy that arrives.
Back to all titles