220 comments later: where I was wrong about "nothing simple works
retained text
a few days ago I posted that there's no edge left in what anyone can download for free, and asked people to prove me wrong. the title was bait, I said so in the comments. I didn't expect 220 replies. I went through them, and here is what I took away, including where I was wrong. **where I was wrong** "nothing is left" was too strong. one commenter put it better than I did: what my examples actually showed is that one-line rules on free candles don't survive costs. that's a narrower claim, and I'll take it. I also got my own math wrong. I wrote that the best of 1,000 random backtests lands around t = 3.7. that's the ceiling. the typical best is about 3.2 (picture, 10,000 runs, anyone can redo it in ten lines). that makes the point worse, not better: a t of 3 from a big parameter sweep is what pure noise gives you on an ordinary day. **what the people who said "it exists" had in common** nobody shared a setup and I didn't ask. but the descriptions had a pattern. they hold for days, not minutes. one person ran about 6,500 short-term FX configs and every two weeks picked the best one. in-sample it made +0.23R per trade, the week after −0.12R, same as a random pick. what survived for him was a plain trend system held for days, because one trade's cost is small next to the move. someone with NQ order book history said almost the same thing. there's a reason others don't take it. one commenter has run an ETF system since 2009 that has 50-70% drawdowns about once a decade. his point was that an edge survives when it has something that makes other people not want it: a drawdown a fund can't sit through, or capacity too small for a fund to bother with. breadth. one person trades intraday edges across many single stocks, found on plain minute bars, after "years and years of failure". not one chart, a whole cross-section. trend in futures over a century came up a lot. I'd still call that pay for sitting through long bad stretches more than something others missed, but it does make money, so it counts against my title. **the corrected version** a simple rule found by searching and traded short-term is almost certainly noise, or eaten by costs. what survives seems to hold longer, sits where big players can't or won't go, and comes with a reason you can say in one sentence. question for this round: if you have something that survived, what was the reason others weren't taking it? the drawdown, the size, the effort, or plain boredom? curious which one is most common.
document fetched 2026-11-18 00:48 · canonical https://reddit.com/r/algotrading/comments/1xyoe52/220_comments_later_where_i_was_wrong_about