Exit Lab
The fleet already spent this SOL. This replays every real closed position through the exact exit code the live worker runs, with the rules changed to whatever you choose, and shows what those same trades would have returned. Nothing here is simulated: every entry, peak and final price is a recorded quote from a position that cost real money and settled on-chain.
Exit policy
Which rule did the selling
Every closed trade is attributed to the one rule that fired. A policy that books its profit through a single rule is a policy resting on a single assumption.
Search for a better policy
Replays the whole corpus against every combination on a grid and ranks the results. This is an exact search over trades that already happened, not a prediction: it tells you what the fleet's past would have paid, which is the only thing this data can honestly say.
Trade by trade
The positions where this policy diverges most from what the fleet actually did. Every row links to the buy and the sell on Solscan.