It is pretty well agreed upon that the Kelly criterion strongly overestimates bet sizes for continuous bets where accurate probabilities are hard to determine, like the stock market. It has given rise to many variations such as Half-Kelly, Kelly minus constant, etc, which is a pretty good tell that it's an inaccurate approximation. Using Kelly to invest in the stock market is roughly like using half the Black-Scholes formula to invest in options. You'll get good results when the market agrees with your direction, but spectacular failures when it doesn't.