Rules and grades that hold up
A strategy is only as good as the parts of it that were decided before the trade — the rules, and the conviction behind each call.
The entry is the part everyone obsesses over, but it is the exit that decides whether a strategy makes money. Rules that hold up name, in advance, the two levels that end a trade: the stop, where the idea is admitted wrong, and the target, where the expected move is judged complete. Written before the position opens, those levels protect you from the two classic mistakes — moving the stop to dodge a loss, and abandoning a target the moment a position turns green.
Conviction should be measured, not vibed
Most services attach mood words — “strong buy,” “high conviction” — that mean whatever the sender wants on the day. A measured conviction grade marks where a call sits in its own model's return distribution, with a number behind it. On the recommended method, every call carries a grade from A (highest) to D (lowest), and the threshold is set per model — which is why the same letter means a very different raw move on a fast clock than on a slow one:
| Model | Holding clock | Grade-A bar (per trade) |
|---|---|---|
| Day Trade | same session, a zero-to-sixty-minute window | 0.70% avg / trade |
| Multi Hour | half a session out to two sessions | 4.50% avg / trade |
| Swing Trade | the flagship, carried roughly seven to twenty-eight days | 6.00% avg / trade |
| Investing | carried over a long horizon | long-horizon |
An A is the top band of a model's own return distribution and a D is the lowest still published. The bar is set per clock, so an A on a same-session Day Trade call (around 0.70% a trade) and an A on a multi-week Swing call (around 6.00%) both read as “top-band for this horizon” rather than one absolute number stretched across very different holding times. There is no E grade — it was retired so the four-step scale keeps its meaning.
Why per-model calibration matters
Hold every model to one fixed percentage and the comparison collapses: a 0.70% same-session move and a 6.00% multi-week move are simply not the same size, so a single absolute cut-off would make the quick clock look feeble and the slow one look heroic while telling you nothing useful. Grading each call against its own model's spread means a B on a Day Trade call and a B on a Swing call each say the same thing — “above-typical for this horizon” — which is exactly the signal a trader who cannot take every call needs in order to know when to lean harder. There is no E grade; it was retired so the four-step scale keeps its meaning.
And because the grade is folded into the on-chain fingerprint (see proof you can check yourself), it is fixed before the outcome and cannot be revised once the trade closes — which is what stops a grade from being a marketing dial turned up after a winner prints.
What a bad version of this looks like
Rules fail this test the moment any of them can be decided after the trade is open. The fragile versions all share that flaw:
- Moving the stop to dodge a loss. The single most expensive habit there is. A stop that slides lower “just this once” the moment price approaches it has stopped being a stop — it is now an unlimited loss with a hopeful name. A bad strategy treats the stop as a suggestion; a real one treats it as the definition of being wrong.
- An entry vague enough to never be wrong. “Buy the dip” or “long around here” can be scored as a win against almost any later price, which is exactly why it proves nothing. If the entry is not a named level you could hand to a stranger, the record built on it is unfalsifiable.
- Catching the knife with no invalidation. Mean reversion fails ugly when a stretch is really the first leg of a genuine new trend. A fragile version keeps averaging down with no line that admits the thesis broke. Without a stop, “it has to bounce” is a prayer, not a plan.
- A win rate with no denominator. A bad record shows the winners and quietly drops the losers, then quotes a glittering percentage with no trade count beside it. A 90% win rate over an unstated number of cherry-picked trades is a billboard; it tells you nothing you can interrogate.
- Sizing by conviction instead of by cap. Betting big on the setups that “feel” strong and small on the rest is how a single bad week ends an account. Conviction belongs inside a fixed risk cap, not in place of one.
Every one of these is the same failure wearing a different hat: a decision made after the trade was open, where the rules could no longer be checked. The cure is to fix every level in advance — and, on a method worth trusting, to commit them in public before the outcome.