Crypto · Exit test

Exit challengers

Two exits were written down before any test: a trailing stop from the highest price, and an exit as soon as price trades at or below the last final 20-day EMA. Each is replayed on the same trades as the control exit, on the same price basis and after the same costs.

Collecting

Forward sample: 2 of 40 closed trades.

Pre-registered as TREND-EXIT-1. From the test's own records with a 5-minute delay, as of 2026-11-11 15:34 UTC. Paper only, not investment advice.

Can a different exit keep more of a trend trade's gain than the trend program's control exit?

Trailing stop from the high: Rejected
Exit at the 20-day EMA: 2 of 40 trades
Deadline: 2027-04-16

Where it stands

The test, stage by stage

RegisteredDoneOct 16, 2026Both candidates and every bar written before any test.
Historical testDoneOct 16, 20261 rejected, 1 continued.
Forward recordNowOct 16, 20262 of 40 closed trades.
First readAheadAt 40 closed trades.
DecisionAheadApr 16, 2027Deadline: fewer than 20 trades by then means not demonstrated.

Results so far

Nothing to show yet, and why

No result yet. No result yet: the decision look needs 40 closed forward trades (an early stop for harm needs 15).

Candidate exit	Verdict	Forward trades
Trailing stop from the high	Rejected by the historical test: worse than the control in one window and not better in either.	-
Exit at the 20-day EMA	Passed the historical screen; forward test: 2 of 40 closed trades.	2 of 40

How it works

The rules, as registered

Historical test: replay both candidates on 2021 to 2023 and 2024 to 2026. A candidate that fails the bar is rejected and never retuned.

Forward: for every new trend trade after the registration, replay the surviving candidate on minute bars beside the exit that actually happened.

Decide at 40 closed trades: select only if the paired interval is above zero and the drawdown is not deeper. Stop early at 15 if it is clearly worse.

Limits

What this test cannot tell you

A selection would only be a proposal: nothing changes the live exit automatically.
Small samples decide nothing; the interval must clear zero.
Paper only: no orders, no real money, and not investment advice.

Related tests