occursio studies whether the behavior of top on-chain traders still holds up after fees, slippage, latency and out-of-sample testing, before any capital is put at risk.
A pipeline built to find out whether there is any real signal in elite-trader behavior, and to say so plainly if there isn't.
Scan public leaderboards, then vet traders on robustness: deflated Sharpe, consistency across periods, and filters for market makers and one-lucky-bet accounts.
Rebuild each trader's positions from their fills, model their behavior, and replay it through a simulator with realistic fees, funding, spread and delay.
Evaluate only on time the model has never seen, using purged walk-forward splits. Anything promising is paper traded before real money is involved.
Most trading backtests fool their authors. These are the rules we hold ourselves to.
Join the waitlist for findings, tooling releases and early access as the project progresses.