The methods that professional traders and investors actually name and use. Each entry says who created it, what it asks you to do, and — just as importantly — where it breaks down.
Where machine learning and AI genuinely help in markets — and where they do not.
Walk-Forward Validation
Standard practice in quantitative research
Intermediate
Tests a model only on data from after the period it was fitted on, which is the minimum bar for any market model.
What it asks you to do
Fit on an in-sample window, then test on the period immediately following it.
Roll both windows forward and repeat across the full history.
Report only out-of-sample results.
Where it struggles
Prevents look-ahead bias but not survivorship bias or overfitting through repeated retesting. Every extra variant you try on the same data quietly weakens the result.
Holding for weeks, not minutes: spotting trends and riding them.
Stage Analysis
Stan Weinstein · 1988
Beginner
Classifies any stock into one of four phases so you buy in the phase where trends actually run.
What it asks you to do
Sorts price action into four stages: basing, advancing, topping, declining.
Uses a long-term moving average (Weinstein used the 30-week) and its slope to tell the stages apart.
Restricts buying to Stage 2 — the advancing phase — and selling into Stage 3 or 4.
Requires volume expansion to confirm a move out of a base.
Where it struggles
Stage labels are obvious in hindsight and ambiguous in real time. Ranges and false breakouts routinely produce stage changes that reverse within weeks.
Source: Secrets for Profiting in Bull and Bear Markets
CAN SLIM
William J. O'Neil · 1988
Intermediate
A seven-part checklist combining company fundamentals with price behaviour and overall market direction.
What it asks you to do
C and A: demand recent quarterly and sustained annual earnings growth.
N: look for something genuinely new — product, management or a new price high.
S: prefer a smaller share supply, and watch for buybacks.
L: buy leaders in a group, not laggards.
I: require evidence of institutional sponsorship.
M: only act when the general market direction supports it.
Where it struggles
Strict growth screens exclude most of the market and cluster into whichever sector is currently hot, so the approach concentrates risk exactly when a theme is most crowded.
Source: How to Make Money in Stocks
Volatility Contraction Pattern (VCP)
Mark Minervini · 2013
Advanced
Identifies bases where each pullback is shallower than the last, indicating supply drying up before a breakout.
What it asks you to do
Measures successive corrections within a base and requires each to be tighter than the previous.
Requires volume to contract alongside price, then expand sharply on the breakout.
Defines a pivot — the precise price where the trade is triggered.
Sets the stop below the final, tightest contraction, which keeps risk small.
Where it struggles
Highly discretionary. Two experienced traders will count contractions differently on the same chart, which makes it hard to test and hard to apply consistently.
Source: Trade Like a Stock Market Wizard
Relative Strength (RS)
Robert A. Levy; later popularised by O'Neil · 1967
Beginner
Ranks a stock by how it has performed against a benchmark, rather than by its own price move.
What it asks you to do
Compares a stock's return to an index over a chosen lookback.
Converts the comparison into a percentile rank across the universe.
Directs attention to the strongest names instead of the cheapest-looking ones.
Where it struggles
Backward-looking by construction. A high rank describes what already happened, and ranks decay fastest exactly at trend turns.
Source: Relative Strength as a Criterion for Investment Selection (Journal of Finance)
Market Breadth
No single originator; standard market-internals practice
Beginner
Measures how many stocks participate in a move, rather than how far a weighted index travels.
What it asks you to do
Counts the share of stocks above a chosen moving average.
Compares advancing to declining issues.
Flags divergence, where an index rises while participation falls.
Where it struggles
A description of the present, not a timing tool. Breadth confirms trends after they begin and deteriorates after tops have already formed.
Source: Advance/decline analysis, in use since the early 20th century
Working out what a business is worth, and paying less than that.
Margin of Safety
Benjamin Graham · 1949
Beginner
Buy only at a large enough discount to estimated value that being wrong still need not be ruinous.
What it asks you to do
Estimate what a business is worth independently of its share price.
Buy only at a meaningful discount to that estimate.
Treat the discount as protection against your own estimation error, not as the expected profit.
Where it struggles
The entire method rests on a value estimate that is itself uncertain. A large discount to a wrong number offers no protection at all.
Source: The Intelligent Investor
Piotroski F-Score
Joseph D. Piotroski · 2000
Intermediate
Scores financial health from 0 to 9 using nine accounting signals, to separate sound cheap companies from failing ones.
What it asks you to do
Checks profitability signals: positive net income, positive operating cash flow, improving return on assets, and cash flow exceeding net income.
Checks leverage and liquidity: falling long-term debt, rising current ratio, no new share issuance.
Checks efficiency: improving gross margin and asset turnover.
Sums the nine pass/fail tests into a single score.
Where it struggles
Built on reported accounting data, so it inherits every weakness of that data — restatements, aggressive recognition, and sector conventions that make cross-sector comparison misleading.
Source: Value Investing: The Use of Historical Financial Statement Information
Magic Formula
Joel Greenblatt · 2005
Beginner
Ranks companies on just two measures — earnings yield and return on capital — and buys the best combined ranks.
What it asks you to do
Rank the universe by earnings yield (cheapness).
Rank it again by return on capital (quality).
Add the two ranks and buy from the top of the combined list.
Hold for a fixed period, then rebalance.
Where it struggles
Deliberately ignores debt structure, cyclicality and accounting quality. It also requires sitting through long stretches of underperformance, which is where most people abandon it.
Source: The Little Book That Beats the Market
Fama–French Three-Factor Model
Eugene F. Fama and Kenneth R. French · 1992
Advanced
Explains stock returns using three factors — market, company size and value — rather than market exposure alone.
What it asks you to do
Extends the single-factor model with a size factor and a value factor.
Provides a way to check whether a strategy is genuinely skilful or merely loaded on known factors.
Underpins most modern factor-based index products.
Where it struggles
A model for explaining returns after the fact, not a trading system. Factor premia have long periods of underperformance, and the original findings are debated.
Source: The Cross-Section of Expected Stock Returns
How much to buy, when to sell, and how not to blow up.
Fixed-Fractional Position Sizing
Widely used; popularised by Van K. Tharp · 1998
Beginner
Risk the same small fraction of capital on every trade, so no single loss can do lasting damage.
What it asks you to do
Fix the percentage of capital you are willing to lose per trade before entering.
Derive position size from that amount and the distance to your stop — not from conviction.
Recalculate as capital changes, so size falls automatically during a losing run.
Where it struggles
Only works if the stop is honoured. It also assumes the stop can be filled at the intended price, which gaps and illiquid names routinely break.
Source: Trade Your Way to Financial Freedom
Kelly Criterion
John L. Kelly Jr. · 1956
Advanced
Calculates the bet size that maximises long-run growth given a known edge and payoff.
What it asks you to do
Takes the probability of winning and the win/loss payoff ratio as inputs.
Returns the fraction of capital that maximises compound growth.
Shows mathematically that both over-betting and under-betting reduce long-run outcomes.
Where it struggles
Assumes you know your edge precisely. In markets you never do, and overestimating it produces dangerously large positions — which is why practitioners use a fraction of the Kelly figure.
Source: A New Interpretation of Information Rate (Bell System Technical Journal)
How to actually use this page
1Pick one framework, not five. Combining several before you understand any of them produces a system you cannot debug when it stops working.
2Read its "where it struggles" section first. If that failure mode is one you cannot tolerate, the framework is wrong for you regardless of its reputation.
3Go to the original source. Every entry names one. Second-hand summaries — including this page — lose the conditions the author attached.
4Write down what would make you abandon it, before you start. A method with no exit condition is a belief, not a process.