Risk: Medium
Fractional Kelly
Stake size is derived from the estimated edge and the price, then divided down (quarter Kelly) to absorb model error.
How it works
Full Kelly maximises long-run growth but is far too violent when the model is imperfect.
We take a fraction (quarter Kelly by default), which keeps most of the growth with a fraction of the swings.
When the edge is zero or negative the formula returns zero and the strategy simply declines the bet.
Formula
f* = (p × (odds − 1) − (1 − p)) / (odds − 1); stake = kelly_fraction × f* × bankroll
- p — Model probability of the selection winning
- odds — Decimal price taken
- kelly_fraction — Slice of full Kelly actually staked, e.g. 25%
Worked example from a €10 base stake
- Price 2.10, model probability 0.53 → b = 1.10, f* = (0.53 × 1.10 − 0.47) / 1.10 ≈ 0.10.
- Quarter Kelly: 0.25 × 0.10 ≈ 2.5% of bankroll.
- On a €1,000 bankroll that is a €25 stake; on a €400 bankroll it is €10.
If the bet wins: The bankroll rises, and the next Kelly stake rises with it — but only when a genuine edge is present.
If the bet loses: The bankroll falls and stakes shrink. Zero or negative edge returns a zero stake and the bet is declined.
Reset conditions: No ladder to reset. Each bet is sized independently from its own edge and price.
Historical robot performance
Settled bets
2
Won / lost
1 / 1
Strike rate
50.0%
Average odds
2.95
Total staked
€20.68
Net P/L
€13.82
ROI
66.8%
Max drawdown
€10.68
Sample size: 2 settled bets. This is far too small a sample to draw conclusions from. Figures are published for transparency, not as evidence that any strategy is profitable.
Simulated bankroll: €1013.82 from a €1000.00 start. Paper mode — no real money is staked.
Advantages
- Stake scales with the real edge
- Best theoretical long-run growth per unit of risk
Disadvantages and risk
- Very sensitive to probability errors
- Large stakes on high-edge bets need the caps
Overall risk rating: Medium.
When this strategy suits
- Situations where the probability estimate is reasonably trustworthy
- Maximising long-run growth per unit of risk
- Portfolios where edge size varies a lot between bets
