How to Evaluate a Trading Strategy

Why No Single Metric Is Enough

Every metric used to judge a trading strategy answers one narrow question and stays silent on the others. That's why any single one, quoted on its own, can mislead:

• A win rate of 90% sounds excellent — but says nothing about how large the losses are, so it can hide a losing method.
• A tempting reward-to-risk ratio of 5:1 means nothing if the target is almost never reached.
• A positive expectancy can still belong to a strategy with a drawdown no one could survive.
• A strong profit factor can be produced by a single lucky trade.
• A shallow maximum drawdown in a backtest can hide the fact that the sample was tiny.
• Even a genuine edge carries a risk of ruin if it's staked too aggressively.

The through-line of the whole MPM approach to statistics is this: no number stands alone, and the more impressive a figure looks in isolation, the more the other questions need asking. Honest evaluation is a checklist, not a headline.

The Questions, In Order

A useful way to evaluate any strategy — MPM's or anyone else's — is to work through these questions in sequence. Each question builds on the previous one, and skipping the earlier ones can make later conclusions misleading — which is why the order matters. Each links to a full explanation.

1. Can I trust the evidence? → Sample Size. Before believing any figure, ask how many trades it rests on. A dazzling result over 20 trades is most likely luck; the same figure across thousands carries weight. Sample size comes first because it governs whether the rest of the numbers mean anything at all.

2. How often does it win? → Win Rate. The share of trades that end in profit. Useful, but incomplete on its own — it says nothing about the size of the wins and losses, so it can never be read alone.

3. What payoff is it aiming for? → Reward-to-Risk Ratio. How much each trade aims to make relative to what it risks. Win rate and reward-to-risk are two halves of the same question — a high ratio usually comes with a lower win rate, and neither means much without the other.

4. What's the average outcome per trade? → Expectancy. Win rate and reward-to-risk combined into a single figure: the average profit or loss per trade. If expectancy is positive, the strategy makes money over time; if negative, it loses — whatever the win rate looks like. This is the number that reveals whether there's an edge at all.

5. How efficient is it? → Profit Factor. Total winnings divided by total losses — a scale-free ratio above or below 1. Another view of profitability, most trustworthy when it agrees with expectancy and rests on a broad, well-distributed sample rather than one outlier.

6. How painful was the journey? → Maximum Drawdown. The worst peak-to-trough decline the strategy endured. Profitability metrics answer "does it make money?"; drawdown answers "could you actually survive trading it?" — because recovering from a deep loss takes a disproportionately larger gain.

7. Can it survive long enough for the edge to matter? → Risk of Ruin. The probability that a losing streak ends the account before the edge pays off. It sits beneath all the others: a measured edge is worthless if the strategy is staked in a way that doesn't survive its own inevitable losing runs. Position size, more than the edge itself, usually decides this.

How the Answers Work Together

The value of this framework comes from the relationships between the metrics rather than the individual numbers themselves. The point isn't to collect seven numbers — it's to see whether they tell a coherent story. A genuinely sound strategy tends to show an adequate sample, a positive expectancy, a profit factor above 1, a drawdown you could realistically sit through, and a negligible risk of ruin at sensible position sizes — all at once, and all consistent with each other.

Tension between the numbers is itself informative. A high win rate paired with a barely-positive profit factor points to small wins and large losses. A strong expectancy paired with a brutal drawdown points to a strategy that makes money on paper but is dangerous to trade. A spectacular figure paired with a tiny sample points to luck. Reading the metrics as a set — and being most suspicious of the one that looks best — is what separates evaluation from wishful thinking.

This is the discipline that runs through every page in this section: profitability, risk, and the quality of the evidence all matter together, and no single measure is allowed to stand in for the whole.

MPM Perspective

This evaluation discipline is not separate from how MPM works — it is how MPM works.

MPM's entire research posture is built on the idea that a good-looking number is a reason for more scrutiny, not less. Where trade-level results appear in MPM's work, they are presented as a complete set — sample size, win rate, expectancy, profit factor, drawdown, and the survivability considerations that follow — rather than as a single flattering headline. The same standard applies to MPM's measurements of historical price behaviour around zones: figures are reported with the number of observations behind them, and claims are withheld where the sample is too small.

The reason this section of the Learning Center exists is to give any reader the tools to apply that same scrutiny — to MPM's published work and to anyone else's. A platform confident in its research should want its readers to know exactly how to check it. Teaching how to evaluate evidence, rather than how to be impressed by it, is the clearest expression of "research, not signals."

Common Misconceptions

"There's one number that tells you if a strategy is good."

There isn't. Every metric answers one narrow question and is silent on the rest. Profitability, risk, and evidence quality each need their own measure, read together.

"If several numbers look good, the strategy is good."

Only if they rest on an adequate sample and tell a consistent story. Good-looking numbers from too few trades, or numbers that contradict each other, are warnings — not confirmation.

"The best strategy is the one with the highest returns."

Not necessarily. A slightly lower-returning strategy with a survivable drawdown and negligible risk of ruin can be far more useful than a higher-returning one that's likely to blow up. Return means little without the risk that produced it.

"Once a strategy passes evaluation, it's proven."

No. Every metric here is historical, and markets change. A thorough evaluation raises confidence and rules out obvious flaws; it never guarantees future results. Evaluation is ongoing, not a one-time certificate.

Limitations

This framework helps you ask the right questions, but it can't turn an uncertain future into a certain one. Every metric it draws on is historical: it describes what a strategy did in a tested period, under conditions that may not repeat. A strategy that passes every check can still fade as markets change, and no combination of statistics removes that fundamental uncertainty.

The framework is also only as good as the data fed into it. Metrics computed from unrepresentative samples, idealised backtests, or figures that ignore real trading costs will mislead no matter how carefully they're read together. The discipline of evaluation reduces the chance of fooling yourself; it does not eliminate it. Used honestly, it's a way of thinking clearly about evidence — not a formula that produces a verdict. Even when every metric points in the same direction, judgement remains necessary: statistics inform decisions, they do not replace them.

How This Fits Into MPM

This overview is the entry point to the MPM Learning Center's strategy-evaluation series — the set of pages that teach how to judge a trading strategy on complete evidence rather than a headline number.

You'll encounter this thinking in:

• Research Papers, where any trade-level result is presented as a complete set of metrics with its sample size and limitations, never as an isolated figure.
• Reaction Library, where MPM's published measurements of historical price behaviour (how often zones held versus broke) are always reported with the number of observations behind them.
• Intelligence Circle — a member-only research environment containing advanced market research, historical investigations, and trading strategies developed using the MPM research framework.

MPM's broader stance is that the ability to evaluate evidence is more valuable to a trader than any single strategy or statistic — and this section exists to build it.