do those four exit variants translate to crypto overnight?

so i read the exit rules comparison in the library, the one with the four blocks you bolt onto the same entry. really liked the framing. Keep the entry fixed, swap the exit, see what actually changes. i've been guilty of endlessly fiddling with entries and never touching the exit, so this hit home a bit. my issue is it's written for US stocks and i basically only trade crypto between like midnight and 5am my time. couple of things i'm stuck on: 1. the time exit. for stocks a 20 or 30 min hold cap makes sense because there's a session, there's a close, dead trades get killed off before the day ends. crypto just... keeps going. is a time exit even doing the same job here, or is it just an arbitrary chop? my gut says it still cuts the trades that went nowhere, but the reasoning behind the number feels different. 2. the -1.0 stop. on the stuff i trade at night that's often inside normal wiggle. do i scale all four numbers by some volatility factor, or does that break the comparison because then i'm changing two things at once? i'd rather keep the exact numbers so my results are comparable to what other people got, but i suspect i'd just get stopped out constantly and learn nothing. 3. variant 3 vs 4 — give-back protection vs breakeven stop. on paper 4 looks safer but i have a feeling in thin overnight books it'd get triggered a lot by a single wick and then i'm out flat right before the actual move. anyone tested those two head to head on anything volatile? not asking anyone to do my homework, just wondering if people ran these outside of stocks and what you had to change. happy to post my numbers once i've done all four properly on one entry.
N
NightOwl_Yuki
27 posts · 0 followers
+ Follow

Related reading

6 replies

H
HalfKelly· Jun 2026 ago· edited Aug 2026 ago
The clean fix for #2 is to stop thinking in points and start thinking in R. Define R as one unit of whatever your normal noise measure is, ATR over your session, or the stdev of 5min returns times sqrt(bars) — then the stop is 1.0R, the target is whatever multiple the original used, and size = risk_per_trade / (1.0R in price). Your numbers stay comparable because the ratios are identical, only the yardstick changed. What breaks comparability is keeping the raw number on an asset that moves four times as much.
DataDrivenDee· Jul 2026 ago· edited Aug 2026 ago
Genuinely curious how you'd tell the two apart though, if you scale the stop by vol and the results differ from the stock version, you won't know if that's the asset or the scaling. Could you run both arms? Same entry, same logged trades, one set with the literal numbers and one with vol-scaled, and just note how many of the literal-stop trades died in the first bar. That first-bar stopout rate on its own would answer your question about whether it's noise or signal.
Q
QuietVol· Jul 2026 ago
On the time exit: in the stock version the cap is doing two jobs at once, capping the hold and respecting the close. Overnight only the first job survives. I'd derive the number rather than import it — take your winners, look at the distribution of bars-to-target, and set the cap somewhere past the bulk of it, maybe the 80th percentile. If most of your winners resolve in 12 bars, a 30 minute cap is just paying for a lot of dead time. Do you have enough logged trades yet to look at that distribution?
CryptoKarl· Jul 2026 ago
3am books do exactly what you think they do to a breakeven stop. I've been wicked out flat and then watched the thing run without me more times than I want to admit. If you test 4, at least try it with the stop parked slightly below entry rather than exactly at it, the difference is not small.
S
SpreadWatcher· Jul 2026 ago
Two things nobody's said yet. A 20 minute cap means a lot of round trips, and at 3am your spread is wider than your average edge might be — the time exit is the variant that gets taxed hardest, so measure it net of fees or the comparison is fiction. Also log your actual exit fills vs the trigger price on all four. Variant 4 will look great on paper and mediocre in the account, and that gap is the whole point.
R
RiskFirstRita· Jul 2026 ago
Checklist before you run any of this: fixed risk per trade, written down, same for all four variants. The trap with widening the stop for crypto vol is quietly keeping the same position size, and then variant 1 isn't a wider stop, it's a bigger bet. Size down proportionally or you're testing your account balance, not the exit rules. Also decide now how many trades per variant before you look at results — deciding after is how everyone talks themselves into the one that got lucky.
Sign in to reply →
← All strategies & backtesting