You already have a favorite setup. You could name it right now. The uncomfortable question is whether the setup you love is the same one your journal says actually pays — because more often than not, it isn’t.
Why gut feeling picks the wrong favorite setup
Ask a trader which setup is their best and they’ll answer instantly and confidently. Ask for the data and the confidence usually evaporates — because the answer was built from memory, and memory is a rigged jury.
Your sense of your “best” setup is distorted by predictable biases:
- Recency — the setup that won yesterday feels hot, regardless of its long-run record.
- Vividness — one unforgettable trade (the huge win, the perfect entry) colors your whole read of a setup that’s otherwise mediocre.
- Effort justification — you like the setups you worked hardest to master, whether or not they earn their keep.
- Survivorship — you remember the times a setup worked and quietly forget the times it didn’t, especially the boring small losses.
The net effect is that traders routinely pour size into a setup that merely feels like their edge while a quieter, less glamorous setup does the actual earning in the background. The journal exists precisely to overrule the gut. If you’ve tagged your trades consistently, the answer is already sitting in the data — you just have to slice it.
Slicing trades by setup tag and sample size
Finding your real edge starts with grouping every trade by its setup tag and, before anything else, checking how many trades sit in each bucket. Sample size is the gatekeeper, and skipping this step is how traders convince themselves of edges that don’t exist.
A setup with six trades tells you almost nothing, no matter how good its numbers look. Random chance produces gaudy short-run records all the time. As a rough discipline:
- Under ~20 trades in a bucket: treat any conclusion as a hypothesis, not a finding.
- 30-plus trades: the numbers start to mean something, though the confidence band is still wide.
- The more, the better — and the more evenly your trades are spread across setups, the more comparable the rankings.
This is exactly where consistent tagging pays off and where a spreadsheet starts to creak: hand-maintained tags drift into synonyms that split your samples, and you end up comparing half-populated buckets. Purpose-built journals aggregate by tag automatically and flag thin samples for you — a distinction we dig into in Shibiki vs Tradezella. The point of the slice isn’t just to rank; it’s to rank only the setups you have the evidence to rank.
Ranking by expectancy, not win rate
Here’s the trap that sends traders after the wrong setup: they rank by win rate. Win rate feels like skill — how often you’re right — but it’s a near-useless way to compare setups, because it ignores the size of the wins and losses entirely.
Rank by expectancy instead: the average amount a setup earns per trade, combining how often it wins with how much it wins and loses. Consider two setups:
| Setup | Win rate | Avg win | Avg loss | Expectancy |
|---|---|---|---|---|
| “Comfortable” | 70% | +0.5R | −1R | +0.05R |
| “Uncomfortable” | 40% | +3R | −1R | +0.80R |
The setup that wins most often is barely profitable; the one that loses more often than it wins is the real workhorse — by a wide margin. Rank by win rate and you’d promote the wrong setup and starve your actual edge. The trading expectancy explainer unpacks why the per-trade average is the number that compounds, and the expectancy calculator lets you drop each setup’s stats in and compare them directly. Expressing everything in R-multiples keeps the comparison honest across setups you trade at different sizes.
Killing the setups that quietly bleed
Ranking by expectancy does something uncomfortable: it exposes the setups running a negative expectancy. These are the quiet bleeders — setups that never blow up spectacularly, so they escape scrutiny, but drip small net losses trade after trade and drag on your whole account.
They survive because they’re camouflaged. A setup with a decent win rate can still lose money if its losses outweigh its wins, and the frequent small victories feel like success while the arithmetic runs the other way. Only the expectancy slice unmasks them.
Once a setup is clearly negative across a real sample, the highest-value action in your entire trading is boring: stop trading it. Not tweak it, not trade it smaller for now — cut it. Every trade you don’t take on a losing setup is money saved with zero effort and zero risk. Traders spend months chasing new edges while a bleeder they already trade drains the gains; killing it is faster, cheaper, and more certain than anything new you could learn.
Reallocating size toward your proven edge
Finding your best setup is only half the payoff. The other half is acting on it — moving size toward the edge the data proved and away from the ones it didn’t. For most traders this reallocation is the single biggest improvement available, and it requires no new skill, only the discipline to follow their own journal.
Do it deliberately, and do it with a safety net:
- Concentrate, don’t over-concentrate. Weight your book toward the proven setups, but keep every trade sized so that a normal losing streak on your favorite can’t threaten your account. Confirm each position against your risk limits with the expectancy calculator feeding your sizing math.
- Scale in with the sample, not ahead of it. Add size as a setup’s edge holds up over more trades, not on the strength of a hot week. Shibiki’s live edge health with a Wilson confidence interval is built for exactly this: it tells you when a setup’s edge is statistically real versus still inside the noise, so you press size on confirmed edges and hold off on tempting-but-thin ones.
- Let the data keep voting. Because Shibiki auto-journals every fill and tracks each strategy’s edge continuously — across every prop account you copy it to — your ranking stays current instead of frozen at last quarter’s read. If your best setup starts to decay, the edge health flags it before it eats a payout, and you reallocate again.
Your journal already knows where your money comes from. Slice it by setup, rank by expectancy, cut the bleeders, and put your size where the evidence is — then let the data keep the ranking honest as you trade.
Related: Expectancy calculator · Trading expectancy · Shibiki vs Tradezella