Reading your own numbers

How to grade your own trades

Most trade grading is circular: the winners get an A. Grade the decision instead of the result, and the grade starts answering a question you cannot answer any other way.

Ask a trader to grade a trade and they will usually grade the outcome. It is almost impossible not to — you know how it ended before you write the letter, and a trade that made $600 does not feel like a C. But a grading system that knows the result is not measuring anything. It is restating the P&L in letters.

The tautology

Grade on the result and you can prove your A trades make money. Of course they do; that is how they became A trades. The table looks impressive, means nothing, and cannot answer the only question grading exists for:

The question worth asking

Do the trades where I broke my own rules actually cost me money? That question only has an answer if the grade and the money are measured independently. The moment the outcome touches the grade, the two are the same number wearing different clothes.

So the rule is uncomfortable but simple: a grade must be computable before the trade resolves. If you could not have assigned it at the moment you clicked, it is not a grade, it is a verdict.

Ten things worth grading on

Everything below is knowable at the moment of the decision, and every one of them is a behaviour rather than an opinion.

CheckFlags the trade when
Size disciplineit is several times your own median size on that instrument
Re-entry pauseyou opened it soon after a loss closed
Session depthit is the Nth or later trade of that session
Losing runlosses were already stacking up behind it
Added to a losera later entry filled worse than the first, in the trade’s own direction
Overtradingit is past the number of trades a normal day holds for you
Hold-time outlieryou held it far longer, or cut it far shorter, than you usually do
Daily stop respectedthe account was already down more than your daily stop
Near a firm ruleit was opened with most of the daily loss limit already used
Stop recordedno stop was written down on it

The thresholds have to be your own

This is where most rubrics fail. “Flag anything over four contracts” is a rule about a stranger. Four is a normal clip for one trader and a catastrophe for another, and a fixed number quietly grades the size of your account rather than the quality of your decisions.

Make every threshold a multiple of your own median — twice what you usually trade on that instrument, three times your normal session length — and the same rule means the same thing to a one-lot trader and a twenty-lot one. It also moves with you: as your size grows, the rule grows, and it keeps flagging the trades that were unusual for you.

Per instrument, too. Ten MNQ and ten MCL are not the same bet — one is $2 a point and the other $100.

Not every check can answer

A check needs data. “Did you add to a loser” needs the individual fills; “hold-time outlier” needs an exit timestamp; a size rule needs enough trades on that instrument for a median to exist. When the data is missing, the honest answer is “cannot tell” — not “clean”. A trade that nothing could be judged on has not earned an A, and counting it as one is the fastest way to build a grade that flatters.

Your grade and a computed score are different tools

Keep both. They are answering different questions and the disagreement between them is the most useful thing on the screen.

A computed score is consistent. It applies the same rule to trade one and trade four thousand, it never gets tired, and it never grades generously because the week was hard. What it cannot see is context: that the setup was the cleanest one this month, or that you knew you were tilted and took it anyway.

Your own grade carries exactly that. So when you write B and the checks say a clean A, one of you is missing something. Either you are being hard on yourself about a trade that broke no rule, or you know something about that trade the rules do not capture — and that second case is where the eleventh check comes from.

What a grade cannot tell you

It cannot tell you the trade was a good idea. Grading measures compliance, not quality — a disciplined trade on a bad thesis is still a losing trade, and it will grade beautifully.

That is a feature. If your A trades break no rules and still lose money, you have learned something precise and valuable: the problem is not your discipline. The conversation moves from “I need to stop tilting” to “this method does not work in this market”, which is a much more productive place for it to be, and one you cannot reach with a grade built on outcomes.

Making it stick

Two practical things. Grade at the close, the same day — memory does not merely fade, it edits, and it edits flatteringly. And grade every trade, not the memorable ones; a grading habit applied only to disasters produces a record of disasters.

What Choptick does with this

All ten checks are built in, and you choose which ones apply to you — a scalper and a swing trader do not share a rulebook. Every threshold is a multiple of your own median, per instrument, recomputed as your trading changes.

  • A letter on every trade, in the Trade Log, with the reason spelled out: “40 contracts against your usual 2 on MNQ”, “back in 12 seconds after the last loss closed”.
  • What each check costs you — every rule you score on, worst first, with the money beside it. This is the panel to act on.
  • Whether you are improving — broken rules per trade, month by month.
  • Your own grade beside it, which Choptick never writes to, so the two can disagree.
  • The day’s flags on your daily reflection, so you rate yourself against something rather than from memory.
  • Nothing stored — the score is recomputed from your current settings, so changing a threshold restates your whole history instead of leaving old trades judged by an old rule.
See Choptick

Figures in the examples are invented and illustrative. Nothing here is financial advice.