I believe it's efficient, at least in the weak efficient market sense (not unique to betting, I believe the wem theorem.) Just because expectations don't correspond to reality doesn't, however, indicate opportunities for arbitrage. Arbitrage (at least from historical information) is...