Definition
What is fractional Kelly?
Last updated
Definition
Fractional Kelly is sizing each bet at a constant multiple (less than 1) of the full Kelly recommendation — typically 0.25× or 0.5× — to reduce variance at the cost of some long-run growth.
Fractional Kelly is the practical answer to full Kelly's variance problem. At 0.25× Kelly, expected log-growth is reduced by about 25% but variance per bet drops by roughly 16×. Most professional bankrolls run between 0.25× and 0.5× Kelly. Glitch Edge's default is 0.25× with a hard 2% bankroll cap per single bet on top.
Why 0.25×, not 0.5×, not 1×
- 1× Kelly: optimal log-growth, brutal variance, 50%+ drawdowns expected
- 0.5× Kelly: ~75% of optimal growth, ~25% of variance, drawdowns more manageable
- 0.25× Kelly: ~50% of optimal growth, ~6% of variance, drawdowns rare
- 0.1× Kelly: barely above flat sizing — leaves a lot of growth on the table
Most operators settle at 0.25× because the growth tradeoff feels acceptable against the variance reduction.
What the cap on top does
Even at 0.25× Kelly, a model overestimating edge by 5× would size a bet far too large. The 2% bankroll cap is the second guard — even if Kelly says size at 8% of bankroll, the platform clips to 2%.