I appreciate your comment and agree overall on the issue with Kelly on correlated bets. But the strategies enacted in this paper aren't affected by the issue you raised because we treat each market as having its own bankroll. The Kelly criterion is merely used as a guide for how to size bets between different markets, but it isn't being applied to the whole bankroll overall.

Note that this choice was forced upon us because the full joint distributions among states wasn't available to us from the 538 estimates. Otherwise, we would have incorporated dependence among the states and scaled the bets accordingly.

Also please note that we do not claim in our argument that it is impossible to beat the betting markets. Our analysis was only between the models posted at popular forecasting sites (such as 538 and Economist) and the betting market at PredictIt. We acknowledge there are profitable traders in all betting markets.

Kelly Criterion makes no sense in this context as it assumes every bet is completely independent. As states are highly correlated, kelly betting will lead to vast overbetting and negative returns (even if your predictions are much more accurate.) I made a >20% return betting on 538 and kelly correctly.