The five-step review
Fix the sample before you look at anything
One account, one date range, closed trades only. Decide the range before you see the result, or you will unconsciously pick the window that flatters you. Write down the trade count; it governs how much weight anything that follows deserves.
Separate the result into profit and loss sides
Look at gross profit, gross loss, average win, average loss and win rate together. Profit factor and expectancy fall out of these. A single one of them in isolation cannot distinguish a good process from a lucky month.
Find the concentration
Group by instrument and by strategy. In most retail accounts a small number of trades or one instrument dominates both tails. If removing your three largest outcomes flips the sign of the period, your headline figure was describing those three trades, not your method.
Subtract the cost of doing business
Total fees, commissions and swap against net result. High-frequency approaches routinely lose their entire edge here, and overnight financing quietly reprices any strategy that holds through the rollover.
End with one testable change
A review that produces five changes produces none. Pick the single largest, measurable leak, write the rule change in your journal, and set the sample size at which you will judge it.
Reading a result honestly
The same numbers support very different conclusions depending on the sample behind them.
| What you see | The lazy reading | The honest reading |
|---|---|---|
| Win rate 72% | The strategy works | Check average loss against average win. High win rate with a large average loss is the classic shape of an unmanaged loser. |
| Profit factor 3.1 over 14 trades | Strong edge | Sample too small to separate from variance. Note it, keep trading the rules, and re-read at a few hundred closed trades. |
| Net P&L up, equity feels flat | Statement is wrong | Check fees, swap and whether the source reports net or gross, and whether open positions are excluded from the figure. |
| One strategy shows a large negative | Stop trading it | Group by instrument first. It is often one instrument inside the strategy, not the strategy, and often one session of it. |
| Drawdown smaller than it felt | Emotional overreaction | Closed-trade drawdown excludes open-position mark-to-market. What you felt was an unrealized excursion the metric does not include. |
Why the journal has to be in the loop
Trade outcomes alone cannot establish that you revenge traded, hesitated, or moved a stop. Those are claims about behaviour, and the only record of behaviour is what you wrote at the time.
This is the practical reason to keep the journal and the analytics in one workspace: the metric identifies where to look, and the entry supplies the cause. Either half alone produces confident conclusions with nothing underneath them.
Performance review questions
How many trades do I need before the numbers mean anything?
There is no single threshold, but treat anything under about 30 closed trades as directional at best, and expect metrics to keep moving until several hundred. The safer habit is to always read a metric with its trade count attached rather than looking for a magic sample size.
How often should I review?
Weekly for execution quality, with a larger review each quarter for strategy-level questions. Weekly reviews catch rule breaks while you still remember the trade; quarterly reviews have the sample size to say something about the edge itself.
Should I analyse open positions too?
Keep them separate. Mixing realized and unrealized results makes every metric ambiguous, because the unrealized part changes without you doing anything. Orderblock’s metrics are computed from closed trades for this reason.