The discipline score is the average of your adherence rates, rule by rule, computed from your real trades. Unlike plan adherence, which is a box you tick, this one is not declared: it is observed in the statement.
Out of 100 trades
Contents
The discipline score measures the gap between the rules you set and what your statement shows. Four rules are checked: was the trade opened inside one of your time windows, did its size increase right after a loss, did the day stop after the number of consecutive losses you set, did the day's loss stay under your limit.
Each gives an adherence rate, and the score is the average of the available rates. The first two rules count in trades, the last two in days, which is logical: trading outside your window is a per-trade decision, not stopping after three losses is a decision for the day.
The detail that changes everything sits in the word 'available'. A rule you disable leaves the calculation, it does not count as met at 100%. Without that, emptying your list of time windows would raise your discipline score, which is exactly the opposite of what the score must do.
The most rewarding reading is crossing it with daily P&L. When the score drops on the days the result collapses, you hold the cause and the remedy in the same line, and the remedy demands no new skill. When the two are uncorrelated over several months, it is the plan that needs revisiting: it governs something other than what costs you.
Then look at the rates rule by rule rather than at the mean. A score of 62% can describe four rules at 62%, which is a generally loose practice, or three rules at 95% and one at 5%, which is a single precise behaviour to fix. Both accounts have the same score and nothing else in common.
The classic mistake. Writing rules loose enough to always be met. Time windows covering the whole session, a daily loss limit at twice what you actually risk, a consecutive-loss threshold at five: the score climbs to 100% and measures nothing any more. The question to ask in front of the score is always the same, what exactly do these rules stop me doing.
It is the average of your adherence rates on four rules verifiable in your statement: trades opened inside your time windows, size not increased right after a loss, day stopped after the number of consecutive losses you set, daily loss under your limit.
Plan adherence is a box you tick trade by trade, so it covers your whole plan but it is declared. The discipline score covers only four rules, but it reads from the time, the size and the result of your trades, so it cannot lie. When the two contradict each other, the second one is right.
It leaves the calculation, it does not count as met. That is deliberate: if a disabled rule scored 100%, the simplest way to a great discipline score would be to have no rules at all. A trader managing risk by hand can therefore remove a rule without their score benefiting.
Not necessarily. Time windows covering the whole session and a loss limit at twice what you risk are met effortlessly and forbid nothing. A perfect score obtained with loose rules describes a plan that commits to nothing, not a practice held.
No, and they share no ingredient. The Oshi Score averages six performance indicators: win rate, profit factor, drawdown and the like. The discipline score looks at no result, only at adherence to rules you set. An account can be excellent on one and poor on the other, in either direction.
The score lives in the Psychology screen, which shows it rule by rule with the rate and the number of breaches behind each one, rather than the mean alone. The rules themselves are edited in Settings, under Trading and discipline: time windows, number of consecutive losses before stopping, maximum daily loss as a percentage of capital.
Two screens build on it. Progress tracking turns it into a series of days, so steadiness shows as something other than an average. And Live session checks the same rules while you trade, which is the only way to act on them before the day is written.
To go deeper on this: read the full guide. Every term is gathered in the glossary.