Trading strategies that didn't work, with the receipts.
Receipts for the strategies that failed
Failure
OVERFIT
Where Sharpe ratios go to die.
Most trading sites sell the thing that supposedly works. This one publishes the things that didn't: each idea's rules written down before the test, how many tries it took, how it did in more than one window and market, and a plain verdict on why it failed.
We ran about 1,500 recorded tests, most of them on 23 ETFs, then took the one survivor to markets it was never tuned on. So far there are 17 write-ups: 15 ideas that didn't work, 1 that passed on paper (and is not proven with money), and 1 that can't be tested.
Walk the graveyard How we test What survived
How things die
| Cause of death | What it means |
|---|---|
| Overfit | The winner doesn't survive an honest count of how many things we tried, or it's no better than scrambled data. |
| Not tradable | The effect may be real but we couldn't have captured it with realistic trading rules. |
| Edge decayed | It worked in one period and failed a later window we fixed in advance. |
| Works but not worth it | It does something real, but misses its bar or gives up the whole return. |
| Passed, unproven live | It cleared its pre-registered tests. It has not been proven with real money. |
Recently buried
- 144 signal tests, two winners, no edge
- 1,192 indicator combinations: the winner was no better than luck
- Weekly reversal: great on 2005–2016, nothing on 2017–2021
- Pre-Fed-announcement drift: real, but it happens overnight
- Wait for the stars to align: 24 setups, 0 survivors
Research and education only. Not investment advice, not a recommendation to buy or sell anything. Overfit is a product of Catalyst Forge, LLC.