Trading methodology comparison, and how to choose one

The stakes are not academic. SEBI's study of individual traders in the equity derivatives segment, published in September 2024, found that about 93 percent of individual traders in equity derivatives made net losses over FY22 to FY24, with aggregate net losses exceeding 1.8 lakh crore rupees. That study covers the leveraged, high-turnover end of the market rather than any particular school of chart reading, and it describes nobody's individual outcome. It is quoted here once, and for one reason: an environment that unforgiving is not one in which the difference between two schools of analysis is likely to be the deciding variable. Something more basic is going wrong, and it is usually that the method being traded was never written down clearly enough to be wrong about.

Start with the claims themselves, stated plainly and without the adjectives each school uses about itself. A methodology is a bet about where information hides. Price action bets it hides in the shape of the move. Indicator systems bet it hides in arithmetic summaries of the same data. Wyckoff and volume analysis bet it hides in the relationship between effort and result, that is, between volume and the price change that volume produced. Quantitative systematic work bets it hides in statistical regularities that are invisible bar by bar and only appear across thousands of instances. Fundamentals-driven positional work bets it hides outside the chart entirely, in the business, and that price will eventually be dragged towards it.

Those are genuinely different bets, and it matters that they are different rather than better or worse. Each one is coherent. Each one has been traded successfully by people whose results are not in doubt, and abandoned unsuccessfully by very many more. What almost never gets stated is the second half of each claim: the evidence it can offer. This is where the schools diverge sharply, and where an honest comparison has to be blunt. The evidence base for a fully mechanical rule is a computation anyone can repeat. The evidence base for a judgement-heavy reading is, in practice, the testimony of its practitioners. Those are not the same kind of thing, and treating them as though they are is the central intellectual mistake in this whole subject.

Accuracy is not one of those axes, and neither is profitability. Not because the question is impolite, but because no comparison of that kind survives scrutiny: the samples are self-selected, the practitioners differ far more than the methods do, the regimes differ across every study period, and the failures are systematically invisible because people who stop trading also stop writing about it. Any table that assigns a win rate to a school of thought is inventing that number, and you should treat it exactly as you would treat an invented number.

What can be compared are the demands. How many hours a week does the method need before it works at all? What capital does it implicitly assume, given its holding period and the way it sizes? What data, software and infrastructure does it require you to run and keep running? How much of the decision is left to your judgement in the moment? And how long is it before a decision you made is actually graded by the market, which is the loop through which any learning at all has to pass? Those five questions have honest answers, they differ sharply across schools, and none of them requires anyone to claim that one school is better.

Notice what the crossings do to the usual argument. If one line sat below all the others across all five axes, that school would simply be the correct answer and the debate would have ended decades ago. It has not ended, and the shape of this chart is why. Every school is cheap somewhere and expensive somewhere else, so the choice is genuinely determined by which costs you can absorb, which is a fact about you rather than about markets. A trader with three free hours a week and no interest in code has already eliminated two schools before considering their merits at all, and has done so correctly.

The feedback axis deserves particular attention because it is the one most people ignore. A method whose decisions are graded within days generates hundreds of gradeable decisions a year; a method whose decisions are graded across quarters generates a handful. The second trader is not lazier or less serious, but they are learning from a very much smaller sample, and their confidence should be calibrated accordingly. Slow feedback is not a flaw, it is simply a cost, and it should be priced into the choice rather than discovered three years in.

Here is the axis this page is really about. Take any method and put it through seven decisions, in the order a trade actually goes through them: what qualifies as a candidate, what triggers the entry down to the bar, where the idea is proved wrong, how large the position is, where you take the money, when you give up on a trade that has gone nowhere, and whether you are permitted to override the rules and on what grounds. Then ask, for each of the seven, whether you could write it as a sentence containing no adjectives, which a stranger could apply without asking you a question.

The count you end up with is the specifiability of your method. It is not a measure of quality. It is a measure of what kind of object you are holding, and it determines almost everything downstream. Every one of the seven you can write down is a decision that can be tested against history, taught to somebody else, checked after the fact against what you actually did, and repaired when it stops working. Every one you cannot write down is a decision that has to be made afresh by you, under pressure, on every trade, and which no test can ever reach.

A method you can specify is a method you can be wrong about. A method you cannot specify is a method you can only be disappointed by.

Two things in that matrix are worth staring at. The first is the position-size row, which is solid across every school. Not one of the five schools has any difficulty writing down how large the position should be, because sizing is arithmetic and arithmetic does not care what you believe about markets. The second is the entry row, which is the most variable of the seven. The entry trigger is what every online argument is about, it is what every course is sold on, and it is the decision on which the schools differ most in whether they can even state their own rule. There is something almost comic in that: the industry argues loudest about the line it writes down least well, and stays quiet about the line every school can write down perfectly.

The honest reading of this axis is that judgement dependence is a cost, not a virtue. That sentence is unpopular, because the schools that score low here tend to describe their judgement as depth, nuance or feel, and there is genuine substance to that: an experienced reader does perceive things that no rule captures. But the price is real and it is paid in a currency the enthusiast rarely counts. You cannot test what you cannot state. You cannot teach it, so you cannot check it against another person's reading. You cannot audit your own past decisions, because the record does not contain the rule you were applying. And when the method stops working you cannot tell whether the market changed or your reading drifted, which is the single most useful diagnostic in trading and it is unavailable to you.

The abstraction becomes concrete very quickly when you hand the same chart to two people. Take a single intention, one both a price-action reader and an indicator trader would recognise: buy strength after a pullback within an uptrend. Now write it twice. The first version is written the way a specification is written, with every term resolved to something a stranger can compute. The second is written the way almost everyone writes it in a journal, in a course, or in their own head.

The first version reads: close above the highest close of the prior ten bars, above the twenty-bar average, and nothing within five bars of the last entry. Every term in that sentence resolves to arithmetic. The second reads: enter when the pullback looks complete and buyers step back in. Every reader nods at that sentence, and no two readers mean the same thing by it. The figure below runs both sentences over the identical series and marks what two independent readers would mark.

Fifty percent agreement is a devastating number if you follow it through. It means that when you back-test that second sentence, you are not testing the method. You are testing your reading of the method on the day you ran the test, and a different day, a different mood or a different sequence of recent losses would have produced a different set of trades and therefore a different result. The test does not measure the edge, it measures you, and it measures you on one particular afternoon. That is the mechanism by which a genuinely honest person ends up with a genuinely misleading back-test, and it is worth understanding properly alongside the more familiar traps in back-testing integrity.

None of this means the second sentence is worthless. Experienced discretionary traders really do read pullbacks better than a ten-bar high rule does, and the information they are using is real even though they cannot fully state it. The point is narrower and unavoidable: the part of your method you cannot state is the part you cannot verify, and you should therefore hold your beliefs about it much more loosely than you hold your beliefs about the mechanical parts. Most blown accounts are not the result of a bad rule. They are the result of a confident belief about a rule that was never precise enough to be tested.

The matcher at the top of this page works one level below the school. It ranks twelve concrete, named setups by fit for a stated situation, and every one of those setups belongs to one of the five schools above. Seeing them sorted that way is instructive, because it shows how unevenly the schools populate a working shortlist. The indicator school supplies five of the twelve, and the most specifiable ones. The quantitative school supplies just one, because a systematic method is usually a whole process rather than a named setup, which is itself a fact worth noticing. And the codes in the last column are drilling addresses into the Master Encyclopedia, where each setup's rules, conditions and failure modes are written down rather than remembered.

Two patterns fall out of that table without anyone having to argue for them. First, the fully specified setups cluster in the indicator and quantitative schools, with the opening range breakout the one structural setup that is fully written down, and only because its window is fixed by the clock rather than by a reader. The partly specified ones cluster in structure and volume reading. That is not a coincidence, it is the same axis as before, seen from the level of individual setups. Second, and more usefully, the setups that are hardest to write down are also the ones with the longest apprenticeship. A Wyckoff phase read is difficult to specify precisely because so much of it lives in the reader, and that is exactly why it takes months rather than weeks to become reliable. If you are drawn to that end of the table, the honest version of the commitment is described in the guide to the Wyckoff method on Indian stocks, and the hours are not negotiable.

Now for the part that the entire methodology argument talks past. Whichever school you pick, it hands you a bar. That is all it hands you. It does not tell you where the idea is proved wrong, how many shares to buy, when to take the money, when to give up, or what to write down afterwards. Those four decisions are yours in all five schools, they are identical in structure across all five, and they determine the outcome far more directly than the choice of which bar you entered on.

The arithmetic makes this uncomfortably concrete. Take one entry, generated by the fully specified rule from the previous section, and write the risk line three different ways. All three are standard, all three appear in serious trading literature, and all three are defensible. Then hold the risk budget constant at one percent of a five lakh rupee account and solve for the position size, which is the only correct direction for that calculation to run.

This is why the claim at the top of this page, that the differentiator inside each school is process discipline rather than the school, is not a platitude. It is a structural observation about where the variance lives. Two traders using the identical signal from the identical school can run positions differing by a factor of three, and that difference compounds across every trade they take, while the difference between their schools applies only to the question of which bar they entered on. Whatever your school, the same four lines have to be written, which is the argument developed at length in the guide to a written swing trading strategy.

Every school fails in a way that is characteristic of it, and the failure is usually a straight consequence of its strength. This is worth studying before choosing, because you are not really choosing a set of advantages, you are choosing which failure mode you are prepared to live with and defend against. The person who tells you their school has no characteristic failure has simply not traded it long enough, or is not counting.

The final row is the one that matters most, and it is worth being precise about why. Every method, including good ones, produces losing runs that feel exactly like a broken method from the inside. There is no reliable way to tell the difference in the moment, which is why the decision has to be made in advance and in writing, when you are calm and have nothing at stake. A trader who switches schools after each bad month has, in effect, guaranteed that no method is ever held long enough to be evaluated, and that the switching itself becomes the strategy. The result is a portfolio of half-learned methods and a growing conviction that nothing works, which is a conclusion the evidence never actually supported.

There is also a quieter failure that spans all five and rarely gets named: adopting a school whose temperament requirement you do not meet. A method that fires often and fails often needs someone who can absorb many small losses without flinching, and handing it to someone who cannot is a guarantee that they will abandon it during the drawdown it was always going to have. This is not a character flaw, it is a fit problem, and it is entirely predictable in advance if anyone bothers to ask. If you are still early enough to be choosing, the ground-level orientation in technical analysis for beginners is a better first step than any comparison table, including this one.

Given all of that, the actual choice is much less dramatic than the debate around it. You do not select a methodology the way you select a favourite. You eliminate, using constraints that are facts about your circumstances rather than opinions about markets, and you accept whatever small set survives. The elimination does most of the work, and it does it without ever ranking anything.

The figure below runs exactly that subtraction over the twelve setups the matcher scores. Three constraints, applied in the order in which they are least negotiable: a person who works full time cannot trade an intraday method honestly, so the horizon constraint goes first; someone unwilling to spend months learning to read one signal cannot run the highest-complexity methods, so that goes second; and someone who cannot sit through many small losses in a row should not adopt a method that fires often and fails often, so that goes third.

What you do with the survivors matters more than which one you pick. Take one, or at most two, and go deep enough to know the failure signature: what it looks like just before it stops working, what conditions it hates, how it behaves in the regime it was not built for, and how far into a losing run you have historically been before it recovered. That knowledge is what converts a setup into an edge, it takes hundreds of observations rather than dozens, and it is unavailable to anyone spreading their attention across six methods at once.

This is also the honest answer to why people who have been doing this for a long time tend to trade very few things. It is not that they lack curiosity about the other schools, and most of them read widely across all five. It is that live attention is the binding constraint, and depth in one method beats acquaintance with six by a margin that is not close. The collecting instinct feels like diligence and functions as avoidance, because there is always another method to learn and learning one is far more comfortable than sitting with the one you have through a bad quarter.

This page answers one question and refuses several others, and it is worth being explicit about which is which. It compares schools on their demands and on how completely they can be written down, both of which are observable. It does not tell you which school is more accurate, which one makes more money, or which one has an edge in Indian markets today, because those questions cannot be answered honestly from any evidence available to anyone, including the people who answer them confidently.

Two further limits deserve naming. The first is that specifiability, the axis this page treats as central, is a measure of testability and not of value. A completely specified method can be completely useless, and the world contains a great many precisely written rules that measure nothing. The claim here is narrower: whatever value a method has, you can only verify the specified part of it, so the unspecified part should carry proportionally less of your confidence. The second is that regime recognition, which decides what any of these methods is being paid for at a given moment, is a skill this page assumes rather than teaches, and it is a substantial subject on its own.

Read plainly, then, the useful conclusion is unglamorous. The choice of school is a fit decision, made by subtraction, from constraints you already know. The discipline underneath it is the constant, it is identical across all five schools, and it is where essentially all of the difference between traders is actually generated. That is a less exciting answer than a ranking would be, and it has the compensating advantage of being true.

• SEBI study of individual traders in the equity derivatives segment (September 2024). The source of the loss figure quoted once above. The study covers the leveraged equity derivatives segment rather than any school of chart analysis, and it describes no individual's outcome. It is cited here only to indicate how unforgiving that end of the market has been in aggregate, and therefore how unlikely it is that the choice between two schools of analysis is the deciding variable. sebi.gov.in
• The two-reader agreement figure. Computed the way inter-rater agreement is normally computed: the number of bars both readings mark, divided by the number of bars either marks. The two readings compared are two defensible operationalisations of the same English sentence, one taking "buyers step back in" to mean a close in the upper half of the bar's range and the other taking it to mean a close back above the previous bar's close. Both are applied mechanically to the same drawn series, so the resulting figure describes the ambiguity of the sentence rather than the skill of any reader.
• The volatility stop distance. True range is the greatest of the session range, the distance from the prior close to the session high, and the distance from the prior close to the session low. The average true range used in the risk figure is that quantity averaged over the fourteen bars ending at the entry bar, computed from the same drawn series as the candles beside it. It is used only to draw one of three defensible stop distances, never to imply that one distance is better than another.
• The Master Encyclopedia codes. The codes shown against each setup, such as BL-021 and VP-022, are addresses into the Bharath Shiksha Master Encyclopedia, 1,308 documented methodologies across seven scanners, where each setup's conditions and failure modes are written down. bharathshiksha.com/encyclopedia
• The demand estimates and the specifiability scores. Both are considered editorial judgements, not measurements. The demand estimates describe typical practice in each school, and the specifiability scores record whether each of seven trade decisions is usually written as a rule, written with a judgement clause, or left to judgement entirely. A disciplined practitioner in any school can specify more than the row shown; the figures describe the common case, and they are presented as a comparison of kind rather than of quality.
• The price series and everything computed on it. The candles, the pullbacks, the signal counts, the reader agreement, the average true range, the stop distances and the position sizes are produced by applying the stated rules mechanically to a synthetic series generated for this guide, not to any real instrument. They measure how rules behave, not how any market performed. No outcome is scored and no success rate is implied for any methodology anywhere on this page.