A coded exit
Most trades are lost not at the entry but at the exit, where a plan meets a feeling and the feeling usually wins. The fix is to code the exit before the feeling arrives.
The entry is the part every builder fusses over, but the exit is the part that decides whether the system makes money. A coded exit names, 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 guard against the two classic system-killers — sliding the stop to avoid booking a loss, and tearing up the target the instant a position turns green.
The discipline is easy to describe and brutal to keep, which is exactly the advantage of a coded system: its rules do not lose their nerve. The levels are set when the call is made and they do not drift because the afternoon turned ugly.
What a bad version looks like
An exit goes bad in predictable ways, and all of them are the same move: a level that was supposed to be fixed gets edited the moment it costs something to honour it.
- The sliding stop. Moving the stop further away as the price approaches it, to avoid booking the loss. It turns a planned small loss into an unplanned large one — the single most expensive habit in trading.
- The torn-up target. Closing early the instant a position turns green, out of nerves rather than rule. It quietly caps the winners that are supposed to pay for the losers, so the whole edge leaks away.
- No exit at all. Opening with only an entry in mind and deciding the exit later. “Later” arrives mid-trade, under pressure, which is the one moment judgement is worst.
- An unprovable exit. Levels that exist only in a chat the operator can edit, so a stop “hit” or a target “reached” can be re-written after the fact. An exit no stranger can re-check is a story, not a rule.
The honest version is the inverse: the stop and the target are written with the entry, in the same breath, and then frozen somewhere they cannot be edited once the trade is live.
How fixing the exit on-chain removes the temptation
The strongest form of “decided in advance” is a level a stranger can confirm was set before the trade resolved. On the codified system here, the entry, target, stop and grade are written into a hash anchored to Bitcoin at publication — in the shape a SHA-256 of the call's entry, target, stop, grade and signal time. Because the target and stop live inside that hash, they cannot be quietly nudged after the fact: any change would break the receipt. The exit stops being a tale told afterward and becomes a fact fixed before the outcome is known.
What that looks like on one call
The mechanism is easier to trust once you have walked a single call through it. The flow below is the whole proof — published, fingerprinted, anchored, re-checked — and it is what turns a coded exit from a promise into a receipt.
Picture a made-up call published with five fields: entry 247.00, target 254.50, stop 243.40, grade B, signal time 09:41:18 UTC. At publication those exact fields are combined into one fingerprint and anchored to a Bitcoin block. The trade resolves later that week. Anyone can take the published call, recompute the fingerprint from those same five fields, and confirm it matches the receipt recorded against a block mined before the trade closed.
Now try to cheat it: imagine the stop were quietly shifted from 243.40 to 244.90 after the price dipped, to dodge a loss. The fingerprint recomputed from the edited call would no longer match the on-chain receipt, and the tamper would be exposed. That is what a coded exit anchored before the outcome buys you — not a promise that the levels were honoured, but proof that they could not have been moved.
The same receipt protects the conviction grade, which is one of the hashed fields — so a call cannot be re-graded from a C to an A after it wins. The reader-side version of this check is on how to test a trading system is honest.