SBF@SBF_FTX·Aug 29, 2020Replying to @tbr90 and @brttbYeah kelly implies really large bet sizes, also often it's too conservative (!!!)115
SBF@SBF_FTX·Aug 29, 2020Replying to @tbr90 and @brttbSo I agree that kelly is best in its class at *median* outcome but it's obviously not at *mean*. Why is median what you're maximizing for?
SBF@SBF_FTXReplying to @tbr90 @brttb and @nntalebya you have to include those still though kelly doesn't maximize mean4:18 AM · Aug 29, 2020·Twitter Web App1 Like
SBF@SBF_FTX·Aug 29, 2020Replying to @tbr90 @brttb and @nntalebnope, math be weird if you have a 50/50 coin that pays out 1.01 : 1, then lim_n-->inf [ (1/2)^n * 2.01^n ] = inf, and in fact is mean maximizing even though lim_n-->inf [(1/2)^n] = 0, e.g. its median outcome is 0.110