Good article, I've used the Kelly formula in the past but abandoned it because my systems fell outside the criteria of my maximum risk amount a lot of the time.
Example, a system is 40% winners with a 5:1 reward, Kelly suggests using 28% capital risk on each trade... a little too scary lol.
One day I'll get round to plotting it, but there seems to be a narrow range of win probability and risk reward ratios that end with realistic risk levels for trading, like 1 or 2%.
I think the Kelly system is optimised for probability of 50:50 like a coin toss, when you start changing the R:R it wildy skews the values.
Resteeming, because I love all things probability !