How success breeds success

We investigate if, and why, an initial success can trigger a string of successes. Using random variations in real?effort laboratory experiment, we cleanly identify the causal effect early competition. confirm that indeed leads to increased chances later success. By alternatively eliminating strategic features competition, turn on off possible mechanisms driving Standard models dynamic contest predict due asymmetric incentives between winners losers. Surprisingly, find no evidence they explain positive winning. Instead, winning seems driven by information revelation effect, whereby players update their beliefs about relative strength after experiencing