Trade records and review: turning experience into comparable data
A trade that was never recorded cannot be improved — only remembered.
Why records are essential
The difficulty in trading is sample size. The outcome of a single trade contains a great deal of randomness, and judging by impression makes it easy to mistake luck for skill or normal variance for a broken strategy. Keeping records converts random outcomes into comparable data, so that judgements have a basis.
In their 2000 study in the Journal of Finance, Barber and Odean used data from tens of thousands of individual accounts to show that trading frequency is negatively correlated with net returns. The implication for an individual is that you need a mechanism to test whether the changes you make actually improve results — and a record is the only workable mechanism.
Fields to record for every trade
- Basic: date, instrument, direction (long or short), lot size
- Plan: entry, stop, target, amount at risk
- Execution: actual fill price, whether there was slippage and how many pips
- Outcome: exit price, profit or loss in money, and profit or loss expressed in R (R = result / amount at risk)
- Process: reason for entry, reason for exit, and whether the plan was followed (yes, no or partly)
The last field is the most often omitted and the most valuable: it separates a loss caused by the strategy from a loss caused by execution drift.
Four metrics to evaluate
Win rate: winning trades divided by total trades. A win rate on its own means nothing and must be read alongside the payoff ratio.
Payoff ratio: average win divided by average loss. Recorded in R, the ratio is directly comparable.
Expectancy: win rate x average win minus (1 − win rate) x average loss. In R terms: with a 40% win rate, an average win of 2.0R and an average loss of 1.0R, expectancy = 0.4 x 2.0 − 0.6 x 1.0 = +0.2R. That means each trade returns on average 0.2 times the amount risked. A positive expectancy is the precondition for a usable strategy.
Maximum drawdown: the largest peak-to-trough decline in account equity. It determines whether you can actually stick with the strategy in live conditions.
Sample size: why twenty trades is not enough
For a strategy with a 40% win rate, a run of five consecutive losses inside a twenty-trade sample is within normal bounds. With a small sample, any conclusion about whether the strategy has broken is highly unstable. A practical threshold is to accumulate at least thirty to fifty complete records before making substantive changes to parameters.
Telling a normal drawdown from a broken strategy
This is the most important step in any review. Compare three angles:
- Nature of the losses: did every loss occur at the planned stop? A loss beyond the planned stop points to execution, not to the strategy
- Market regime: has the volatility environment the strategy was designed for changed? A range strategy built on low volatility, for example, will perform worse once volatility expands materially
- Sample size: is the current losing run still inside the normal distribution for that win rate?
Having compared all three, the sequence is: first halve risk per trade (this step needs no conclusion), then use the comparison to decide whether to restore the original size, adjust parameters, or retire the strategy.
Review frequency and process
- After each close: complete the fields, record the result in R, mark whether the plan was followed
- Weekly: compute win rate, payoff ratio, expectancy and execution-drift rate, and check for behaviour that deviated from the plan
- Monthly: verify that expectancy is positive and that maximum drawdown is within tolerance, and check whether the strategy's assumptions still hold
Lesson summary and self-check
- The core purpose of records is to turn random outcomes into comparable data.
- Measuring in R (result / amount at risk) is what makes instruments and position sizes comparable.
- Expectancy = win rate x average win − (1 − win rate) x average loss, and must be positive to be usable.
- Make substantive parameter changes only after at least thirty to fifty trades.
Self-check: (1) With a 35% win rate, an average win of 2.5R and an average loss of 1.0R, what is the expectancy and is the strategy worth continuing? (2) During a losing streak, what is the first thing you should do?