Previous empirical research on variable pay systems have suggested that possible gains can come from paying for performance, but highlight the difficulty firms face in measuring performance. Using contracts signed in Major League Baseball’s free agent market, I find that over the 2010-2014 period, teams utilized variable pay schemes with players that were more productive or signaled greater risk, either in their contract terms or via overspecialization. However, not all forms of risk signaling were correlated with greater use of performance incentives, including age and proxies for injury history. These findings have significant implications for labor practices more broadly, as the high-information environment of Major League Baseball can shed light on how principals behave when performance measurement costs are effectively eliminated.