the exit block in that volume-confirmed continuation setup bugs me

Read the volume-confirmed continuation entry writeup in the strategy library this morning and I keep coming back to it, mostly because I like that the author flagged their own arbitrary numbers instead of pretending 1.5 and 125 came down from a mountain. Refreshing. But my brain went straight to the sell block, because that's where I always look first. Sorry, it's a condition. 2.5% target, -1.2% stop, 900 second time cap. That's roughly a 2:1 payoff, which means you need something north of a third of trades to hit the target before the stop or the clock. And the clock is the part nobody budgets for. A trade that times out at -0.4% is not a winner and it's not a loser, it's a small fee-paying nothing, and if a big chunk of your fills end up there the 2:1 on paper isn't the 2:1 you actually live with. Has anyone here who's run continuation entries like this actually logged what fraction exit on time vs target vs stop? That distribution seems way more important than the entry filter and I never see people talk about it. Second thing, and this is the bit I want to get right before I touch anything. The note about ChangeAngle scaling with the window is genuinely the most useful paragraph in the piece. If the angle is degrees, then the same threshold means a totally different demand at 60s vs 600s, and you can silently write a condition that can basically never fire. My instinct is to widen to 300s because 60s of volume feels thin to me for a ratio test. But then I have to drop the angle, and I have no principled way to pick the new number. I'd just be eyeballing it. Is there a sane way to translate a threshold across windows, or is it purely 're-test it at the new window and forget the old number ever existed'? And my checklist question: if I'm wiggle-testing the 1.5 ratio, the 125, AND the angle, that's three knobs. At what point am I just fitting noise? My rule has always been one variable at a time, hold the rest still, and if the result only works at one exact value it wasn't a result. Curious whether people think that's too conservative.