A trade that won was not necessarily a good trade. A trade that lost was not necessarily a bad one. If your review process cannot tell the difference, you are systematically reinforcing the wrong behaviours and punishing the right ones.
This is the single most counter-intuitive piece of work in trading discipline. Your nervous system wants to grade trades the way it grades everything else: by result. Won = good. Lost = bad. Inside a probabilistic edge, that grading is not just wrong — it is actively destructive, because it reinforces lucky bad behaviour and discourages unlucky good behaviour.
This article explains the grading system that replaces outcome-based review, the four tiers and what each one actually means, how to apply it honestly when your defensive instincts are pushing against it, and how to use the data to improve. It builds on the journal article, which records the data this grading evaluates.
Two specific failure modes, both severe.
It reinforces lucky bad behaviour. You break your rules on a trade — entered without the full setup, sized larger than usual, ignored the stop. The trade happens to win. Outcome grading says it was a good trade. Your brain stores the win and the rule-breaking together, and next time the urge to break the rule arises, the precedent of the win supports it. You have just reinforced the exact behaviour your method exists to prevent. Over enough repetitions, the reinforcement accumulates until your discipline is so eroded that the inevitable rule-breaking losses cannot be controlled.
It punishes unlucky good behaviour. You execute a setup correctly. Stop where it should be. Size correct. Exit on plan. The trade loses — not because of anything you did, but because random distribution inside your edge put this one on the losing side. Outcome grading says it was a bad trade. Your brain stores the loss and the execution together, and next time the same setup appears, hesitation creeps in. You have just punished the behaviour your method depends on.
Across a year, outcome grading systematically degrades the quality of your trading by rewarding the wrong things and punishing the right things. The traders who escape this loop do so by grading on process — what they controlled — independent of outcome, which they did not.
Every trade gets one grade. The grade is recorded in your journal at exit. It is independent of P&L.
You followed every rule. The setup was a clean match for your defined criteria. Entry was at the planned level. Stop and target were preset before entry. Position size was calculated correctly. No mid-trade deviation. Exit was at the planned level or on a defined signal. The trade looks like the example diagram of your method.
An A-grade trade with a loss is a good trade. The edge was executed correctly; the random distribution produced this outcome; no learning required beyond logging the data point.
You did 90% of the plan. One small slip. Entered 1-2 ticks late because price moved fast. Exited 5 minutes before the time stop because the price action signalled exhaustion early. Sized slightly under what the rules indicated because you wanted some cushion. Nothing rule-breaking, but not the textbook version either.
B is fine. It is the realistic execution most disciplined traders achieve most of the time. A consistent stream of B-grade trades produces the expectancy your edge promises. The work is to keep the slips small and avoid the slow drift from B toward C.
You followed the letter of your rules but the spirit was off. Took a borderline setup that technically met criteria but you would normally have skipped. Sized lower because nervous (and the nervousness was a signal you ignored). Hesitated on the entry then chased it. Did not violate any specific rule, but the trade did not have the clean character of an A or B execution.
C trades are the warning signs. They are not failures, but a high frequency of C-grade trades in a week suggests state issues, regime issues, or rule-criteria drift. They are the lead indicator of approaching D-grade behaviour.
You broke at least one of your own rules. Moved a stop in real time. Took a trade without the setup confirming. Sized up beyond your defined risk parameters. Exited without a defined signal. Entered without a defined target.
D trades are failures regardless of outcome. A D-grade winner reinforces nothing good; if anything, it reinforces the rule violation. A D-grade loser is the expected outcome of operating off-plan. The work is to identify what specifically failed (state, criteria, sizing discipline) and address the cause — not the trade.
Three rules.
Grade at exit, not at end of day. While the trade is fresh, your memory of what you actually did is accurate. By evening, the rationalisations are constructed and the grade you give is partly a story. Grade in the moment, when honesty is cheapest.
Grade against what you wrote down, not against what you remember planning. If your journal records that you planned a 2R target and you exited at 1.5R because “the price action looked weak,” the deviation is on the page. The grade can be done objectively against the written plan. If the plan only lived in your head, the grade is whatever the post-trade rationalisation lets it be.
Grade strictly. Err toward lower grades. Most traders grade themselves too generously. If you are in doubt between A and B, give B. If in doubt between B and C, give C. The cost of grading strictly is mild self-criticism; the cost of grading generously is reinforced bad habits. Trade the second cost for the first.
The grades become decorative. The trader cannot bring themselves to record D against their own work. The journal looks pristine. The actual execution is unimproved. If your grades never drop below B, you are not grading honestly — you are recording a self-portrait.
The trader writes down the grade after seeing whether the trade won or lost. Wins receive A or B; losses receive C or D. This is the original failure the system exists to prevent. If you can predict the grade from the P&L column, the grading is not measuring process — it is measuring outcome with a letter on top.
On a quiet day, the trader fills in the grade carefully. On a day with five trades, they skip the grades because “there’s no time.” The busy days are the days the grades matter most — that is when execution quality is most variable. Skipping grades on busy days means the data is biased toward the quiet days when discipline is easy.
The trader fills in grades for the trades that stood out — the big winner, the bad loser, the one that triggered an emotional reaction. The unremarkable trades get no grade. Across a month, the graded sample is heavily skewed and tells you little about your average execution.
The trader maintains the discipline for two weeks. Then a few D grades arrive. They stop grading because “it’s not helping.” The grades were doing exactly what they were designed to do — surface the rule-breaking behaviour the trader needed to see. Stopping at the moment the system started working is the most common sabotage of all.
Three time horizons.
Weekly: distribution. Count the A, B, C, D grades for the week. A healthy distribution is roughly 30% A, 50% B, 15% C, 5% D. If the C’s and D’s climb above 30% combined, your execution is not yet at the level your method requires; the next week needs to be about that, not about taking more trades.
Monthly: cross-tab grade against outcome. Across a month, do your A and B grades produce better expected R than your C and D grades? They should. If they do not — if your highest-graded trades are losing more than your lowest-graded ones — either your edge is not real, or your grading is not honest, or both. Investigate.
Quarterly: identify the most common cause of B/C/D grades. Look at the journal entries for the lower grades. Is the pattern about entries, exits, sizing, state, criteria drift? Pick the single most frequent cause. That is the priority work area for the next quarter. Do not try to fix everything; fix the one most common leak, measure whether it improves, then move on.
Grading on process is the practical mechanism that makes random distribution a livable framework. The outcomes are noise inside the edge; the grades are signal about what you control. Without the grading, every win and loss feels like a verdict on you, and the verdicts produce the recalibrations that destroy edges. With the grading, you have a separate channel of information about your execution that is independent of the variance.
Grade at exit. Grade honestly. Use the data weekly, monthly, and quarterly. Done consistently, it is the single fastest way to improve the only thing you actually control: how you behave at the screen.
Next planned: How to Handle a Losing Streak Without Destroying Your Account — the math of streaks, the three damage paths, and the protocol that keeps you in the game when the variance is against you.