Free Tool

Kelly Criterion Calculator

The mathematically-optimal fraction of capital to risk per trade — and why almost no professional desk in India uses full Kelly. The output of this tool is the maximum Kelly value; the recommended sizing is half-Kelly or quarter-Kelly for reasons covered below.

Win rate (%) Percentage of trades that close above breakeven, measured over at least 30 trades.

Reward-to-risk (R) Average winner divided by average loser. 2.0 means winners are twice the size of losers.

Account size (₹) Total capital available for trading.

Full Kelly

—

Half Kelly

—

Recommended

Quarter Kelly

—

Professional default

—

Why almost no professional desk uses full Kelly

Kelly assumes a known edge. Real edges are estimated. A win rate measured on 30 trades has a 95% confidence interval roughly 17 percentage points wide; Kelly computed with the central estimate over-sizes against the lower-bound estimate.
Kelly produces deep, normal drawdowns. Full Kelly's expected maximum drawdown over 1,000 trades is roughly 50% of capital. Half Kelly's is ~25%. Quarter Kelly's is ~12%.
Kelly assumes serial independence. Real markets cluster — winning trades tend to follow winning trades within a regime; losing trades cluster within a regime shift. Serial correlation makes Kelly more aggressive in practice than in theory.
Behavioural drift. The trader who computed Kelly is not identical to the trader executing it under drawdown. Half or quarter Kelly leaves room for the inevitable execution drift.

Reading the output

The institutional default of 1% per trade typically lands well below quarter Kelly for any system with a positive edge. This is deliberate — institutional sizing is not optimised for return; it is optimised for survival across uncertain edge estimates and unknown regime shifts.

Get the Kelly sizing report

An email-delivered PDF covering the Kelly derivation, the Indian-market drawdown profiles for full / half / quarter Kelly, the seven reasons institutional desks cap below quarter Kelly, and the rebuild protocol when realised drawdown exceeds modelled drawdown.

Bharath Shiksha does not spam. Unsubscribe instructions in every Academy email.

Sent. The report will arrive within minutes.

Formula: f* = (W × R − L) / R where W is win rate (0 to 1), L is 1 − W, R is reward-to-risk ratio, and f* is the optimal fraction of capital to risk per trade.

Source: Kelly, J. L. (1956), "A New Interpretation of Information Rate", Bell System Technical Journal. Adapted for binary-outcome trading systems. Full treatment in Stage 3 Volume 2 (Advanced Risk Management).