You close out a red month. Somewhere between the last trade and Sunday evening, you pick one of two stories.
Story one: the strategy is broken. Time to rebuild, add a filter, find something that works in this market.
Story two: the strategy is fine, you just kept doing stupid things. Time to get disciplined.
Most of us have flipped between those two stories in the same weekend. And here's the annoying part — they ask for opposite actions. If the strategy is broken, more discipline just means you follow bad rules more faithfully. If your execution is the problem, rebuilding the strategy destroys something that was already working.
So the first job is working out which one you're actually looking at, and a hunch on Sunday night doesn't count.
The good news is that this is one of the few questions in trading with a clean answer, because both stories leave different fingerprints in your journal. You just have to be recording the right thing.
"Discipline" is a number, not a personality trait
Trading psychology gets talked about as character. Are you patient enough, are you calm enough, do you have the right mindset.
I'd rather treat it as arithmetic. Discipline is the gap between the trade you planned and the trade you took. That gap has a size. You can measure it in R, and R is just your standard risk per trade. If you normally risk €200 and a trade loses €200, that's −1R. If it makes €400, that's +2R.
Four gaps are worth measuring:
- Entry. Did you enter where the plan said, or somewhere else?
- Stop. Was the stop where the plan put it?
- Size. Did you risk your normal unit, or more, or less?
- Exit. Did you exit on the plan's condition, or on a feeling?
That's it. Four comparisons per trade. Planned versus actual, and nothing else.
And a plan you can compare against has to exist before the trade. Written down, boring, specific. If your "plan" is reconstructed on Sunday from what you remember intending, you'll unconsciously write a plan that matches what you did. Everyone does it. Memory is helpful like that.
What the two diagnoses look like
Made-up numbers, but the shape is real. Imagine a month with 63 trades on one setup.
Month A
- 44 trades where you followed all four rules: +0.24R each, so +10.6R
- 19 trades where you broke at least one: −0.82R each, so −15.6R
- Net: −5.0R. Red month.
Month B
- 51 trades where you followed all four rules: −0.09R each, so −4.6R
- 12 trades where you broke something: −0.04R each, so −0.5R
- Net: −5.1R. Same red month.
Identical bottom line. Completely different problem.
In Month A the strategy did its job. The trades you took as designed had positive expectancy — the average result per trade across the whole batch, winners and losers together, which here was +0.24R. The damage came from somewhere else. In Month B you followed your rules almost all month and still bled. Nothing about your behaviour explains that. Either the edge was never as big as you thought, or the environment moved under it, which is a separate diagnosis worth running properly on its own.
You cannot tell A from B by looking at the equity curve. They're the same curve.
Not all four gaps are equally measurable
This is where most rule-adherence tracking goes soft, and it's worth being picky.
Some deviations have a clean counterfactual — you know exactly what the plan would have produced, because it's the same trade on the same price path. Others don't. Treat them differently or you'll fool yourself.
Size deviations are pure arithmetic. Say you doubled up on 5 trades after a losing streak. Four lost, one won. At double size that's −2R four times and +3.2R once, so −4.8R. Rescale those same five trades to your planned risk and it's −4R plus 1.6R, so −2.4R. The oversizing cost you 2.4R. No counterfactual price path required — same trades, same fills, one different multiplier. (Strictly, doubling your size can nudge your fill in a thin market. At retail size that's noise, but it's the one crack in the arithmetic.)
Exit deviations are nearly as clean. You were in the trade. The price path after your exit is a matter of record. Replay the plan's exit condition on that path and you get the number. Say 5 trades where you bailed early averaged +0.10R as taken, and the plan's exit on those same paths averaged +0.90R. That's 4.0R left behind. Exits are also where a surprising amount of expectancy lives, which is the whole argument for treating management as part of the edge rather than an afterthought.
Stop deviations are messier but usually answerable. Would the planned stop have been hit on the same path? Usually yes or no, occasionally ambiguous on a wick.
Entry deviations are the dirty one. If you chased an entry 8 ticks late, fine, you can roughly reprice it. But if you took a trade the plan didn't allow at all, there is no planned twin to compare against. The honest answer is "this trade should not exist," not "here's what it would have made."
Back to Month A. Two clean measurements — 2.4R of oversizing and 4.0R of early exits — account for 6.4R. The month lost 5.0R. Before touching a single entry, the arithmetic errors cost more than the month did.
That is not a promise you'd have been green. It's a much narrower claim, and a much more useful one: two things you can fix without changing your strategy at all were larger than the loss.
The trap: you can't just delete your bad trades
Here's the part that gets skipped, and it's the reason I don't fully trust anyone who says "without my three tilt trades I'd be up 4%."
Rule-breaking trades are not a random sample. You break rules under specific conditions — after a missed winner, after two losses, on Friday afternoon, when the move is already running. Those conditions have their own market character. Removing them from your data isn't like removing a random 19 trades. You're removing a particular slice of market states.
Which means the correct statement is "the trades I took as designed averaged +0.24R," not "my strategy makes +0.24R and my emotions cost me the rest." Only the first one is something you measured. The second quietly assumes those 19 broken trades would have behaved like the other 44 if you'd handled them properly, and you have no evidence for that.
Same caution on sample size. Five oversized trades is five trades. It's a record of one month, and one month of five trades predicts very little. Those five early exits cost 4.0R this time round; over the next thirty the same habit might cost half that, or nothing. Give the pattern a couple of months before you build a rule around it, and keep an eye on whether it survives. Roughly the same discipline you'd apply to any other segment of your journal — the numbers worth tracking are only worth tracking when there are enough of them to mean anything.
The answer nobody puts in the discipline course
Run this comparison honestly and sometimes you get a third result: the deviations made money.
The trades you weren't supposed to take outperformed the ones you were. The early exits saved you more than they cost. This happens more often than the mindset literature admits, and it means something specific.
It does not mean you should trade on instinct. It means your written plan is behind your actual judgement. Some condition you recognise on the chart never made it into the rules, so the rules keep telling you no while your hands keep saying yes and being right.
That's a plan problem, not a person problem. The fix is to find the condition, write it down, and test it as an actual hypothesis on trades you haven't seen yet. Not to declare yourself a discretionary genius. Most of the time this pattern evaporates on the next thirty trades, and you were just describing a lucky stretch. Sometimes it doesn't, and you've learned something real about your own process.
Either way you had to measure to find out.
Making this an actual weekly habit
You don't need a system. You need four extra columns and the discipline to fill them in before the outcome is known.
Write the plan before entry: entry trigger, stop level, position size, exit condition. After the trade, mark each of the four as followed or not, and tag which one broke. Binary is fine. Scoring adherence 0–100 sounds rigorous and mostly just gives you a number you can't act on.
Then once a week, split the trades in two — clean and not clean — and compare average R. That single comparison is the whole method. If clean trades are positive and the month is red, you have an execution problem and it's fixable this week. If clean trades are flat or negative, stop working on discipline and go audit the strategy.
One warning from experience: this only works if the plan is genuinely specific. "Enter on confirmation" cannot be graded. "Enter on the first close back above the level, market order, no more than 3 candles after the sweep" can. It doesn't matter whether your rules come from order blocks, moving averages or raw price action, since a well-built edge framework is strategy-agnostic by design. What matters is that the rules are written tightly enough to fail a check.
This is also, quietly, the difference between a journal that changes your behaviour and one that just accumulates screenshots. Plenty of traders log every trade for a year and improve at nothing, because the log records what happened without ever comparing it to what was supposed to happen. If that sounds familiar, the gap between journaling and improving is usually exactly this missing comparison.
Where software helps, and where it doesn't
Nothing above needs a tool. A spreadsheet with four extra columns does it.
What a tool buys you is structure. The plan gets captured as its own fields — entry trigger, stop, size, exit condition — and you fill them in while the outcome is still unknown. Adherence then lives in its own layer of the data, separate from the strategy. That's how we tag every trade in EdgeFlow: four layers, technical setup, execution, environment and management, so a red month can be pointed at a layer. Adherence ends up as a number you see each week next to your expectancy.
Be honest about the limit, though. No journal can stop you going back afterwards and quietly editing what you claimed the plan was. The fields are there; filling them in before you know the outcome is still on you.
It won't tell you why you sized up after two losses. Nothing will. That part is yours.
What it will tell you is whether the thing costing you money this month is the plan or the person following it. That's the fork in the road. Everything you do next depends on which way you turn, and almost nobody bothers to check.
Start with the four columns. If you want the comparison run for you every week, that's what our trading analytics are built to do — and the demo on the homepage is free to poke at before you decide anything.