What is max drawdown?

Maximum peak-to-trough decline of bankroll over a series of bets.

How do you compute fair probability or fair odds?

Enter either fair probability (as % or 0–1) or fair odds; the tool uses p or 1/fair_odds accordingly.

What is Kelly fraction α?

We bet f = α × f*, where f* is the Kelly optimal fraction derived from edge and odds.

Max Drawdown Calculator

Simulates bankroll paths and reports the distribution of the maximum peak-to-trough drawdown. Accepts decimal comma or dot.

Odds format

Offered odds (decimal)

Fair probability (p) Give as % (60,2) or 0–1 (0,602).

…or Fair odds (decimal) If filled, p = 1 / fair odds. Typing here clears “Fair probability”.

Kelly fraction α (0–1) α = 1 full-Kelly, 0.5 half-Kelly, etc. We bet f = α × f*.

Average stake (€) Used to estimate starting bankroll B₀ ≈ avgStake / f (so drawdowns also shown in €).

Number of bets

Simulations

How to use

Select Odds format and enter Offered odds. You may also enter Fair odds — formats convert automatically.
Enter fair p (as % or 0–1) or Fair odds. Typing one clears the other.
Choose a Kelly fraction α (0–1). The model bets f = α × f* each bet.
Provide your average stake (€) to estimate starting bankroll (B₀ ≈ avgStake / f), so drawdowns can be shown in euros.
Set number of bets and simulations (capped to protect performance) and hit Run.

Kelly fraction f*

–

Using stake fraction f

–

Implied starting bankroll

–

Exp. log growth / bet

–

Median max drawdown

–

90th pct. (worse)

–

95th pct. (worse)

–

Probability DD ≥ 30%

–

Median DD (≈ €)

–

95th pct. DD (≈ €)

–

Histogram: distribution of maximum drawdowns (labels = % of sims). X-axis at 0 / 25 / 50 / 75 / 100%.

Equity paths: bankroll PnL% over bets (plots up to 300 paths if simulations > 300).

What is maximum drawdown?

Maximum drawdown (MDD) is the worst peak-to-trough decline of a bankroll over a sequence of bets. If B_t is bankroll after bet t and P_t = \max_{0..t} B_i is the running peak, the drawdown at t is DD_t = (P_t - B_t) / P_t. The reported value is MDD = \max_t DD_t.

Model & assumptions

Each bet uses stake fraction f = α × f*, where α ∈ [0,1] and Kelly fraction f* for decimal odds o and fair probability p is f* = ((o-1)·p - (1-p)) / (o-1) (clamped to [0,1]).
Win pays o on stake; loss forfeits stake. Bankroll updates multiplicatively: B_{t+1} = B_t · (1 + f·(o-1)) on a win, otherwise B_{t+1} = B_t · (1 - f).
We simulate S paths of length N with i.i.d. wins ~ Bernoulli(p).

Why we ask for average stake (€)

The tool estimates a starting bankroll as B₀ ≈ avgStake / f. This lets us convert percentage drawdowns into approximate euro amounts for planning and risk limits.

Expected log-growth per bet

For the chosen f we show E[log growth] = p·ln(1+f·(o-1)) + (1-p)·ln(1-f). Positive values imply long-run growth; negative values imply shrinkage.

Worked example

Suppose offered odds are 1.72 and fair probability is 60.2%. Then f* ≈ 4.92%. With half-Kelly (α = 0.5) we bet f ≈ 2.46% of bankroll each time. With 2,000 bets and 200 simulations the median maximum drawdown typically sits around 50%, while the 95th percentile can exceed 70%—use the chart and percentiles to set tolerable risk.

Related concepts

Convert odds formats with Odds formats, derive no-vig p with Hold/Overround, and compute stake sizes with Kelly calculator or expected value with EV calculator.

More tools

Edge Finder Max drawdown Implied probability Odds formats Hold/Overround Arbitrage Back/Lay hedge Kelly calculator EV calculator All tools