Lately I've been wondering if I'm paying attention to the wrong trades. Not the winners. The ones that never fully worked.
Most systems treat failed execution as waste. A route misses, timing is off, liquidity disappears, the trade gets abandoned and everyone moves on. But the more I watch how traders behave across chains, the less convinced I am that failure is actually being lost.
What if a failed trade is just execution data that hasn't been priced yet?
I keep noticing that successful trades get recognized because they settle on-chain, meaning the transaction is visible and completed. The failed attempts usually stay off-chain, hidden inside searches, route comparisons, rejected paths, and decisions that never reached final execution. Yet those moments contain friction. They show what conditions were avoided, where liquidity broke down, and which opportunities looked attractive but weren't worth the risk.
"Sometimes the miss contains more information than the fill."
That feels strange, but I keep coming back to it.
If thousands of participants repeatedly avoid the same routes, fail at the same moments, or abandon similar setups, a pattern starts forming. Not participation data. Selection data.
Maybe the advantage isn't learning from what traders choose. Maybe it's learning from what they consistently refuse to do. I'm not sure most systems know how to recognize that difference yet.