Skip to content

Bankroll Simulator

Run the same edge ten thousand times and look at where the seasons actually land — including the ones that go badly.

$
%

Break-even at this price is 52.38%. Your edge is 2.62%.

$

Median ending bankroll

—

Running…

Finished profitable
—

Even a real edge loses money over a season more often than people expect.

Hit a 25% drawdown
—
Hit a 50% drawdown
—
Went broke
—

Could not cover the next flat stake.

Median worst drawdown
—

The typical peak-to-trough fall along the way. This is what a season feels like from the inside.

Runs entirely in your browser · nothing is sent or saved

How it works

Expected value tells you the average. It does not tell you what a single season looks like, and a single season is all you get. This simulates thousands of them from the same edge, the same price and the same staking plan, and shows the spread.

The spread is always wider than people expect. A genuine 55% bettor at −110 over 500 bets is a serious edge, and a meaningful share of those seasons still finish down. Not because the edge was fake — because 500 bets is a small number and variance does not care that you were right.

Read the median, not the mean. The mean is dragged upward by a handful of extraordinary runs that you are unlikely to have, particularly under percentage or Kelly staking where good runs compound. The median is the season in the middle, and it is the one to plan around.

The staking plans behave quite differently. Flat staking is the only one where strict ruin is possible, because the stake does not shrink when the bankroll does. Percentage staking recalculates from the current bankroll, so it can approach zero without reaching it — for those, the drawdown figures are the real risk measure rather than the ruin probability. Kellysizes from your edge, which grows fastest and swings hardest; the multiplier is there because full Kelly’s drawdowns are brutal even when the edge is real.

The drawdown numbers are the most useful output on the page. Not because they predict anything, but because seeing that a winning strategy hits a 25% drawdown in a large share of seasons is what stops you abandoning it the first time it happens.

The formula

for each of N simulated seasons:
  for each bet:
    stake = fixed amount                 (flat)
          | bankroll × fraction          (percentage)
          | bankroll × kelly × multiplier (Kelly)

    if stake > bankroll → ruin, path stops

    win  (probability p): bankroll += stake × (D−1)
    loss (probability 1−p): bankroll −= stake

    track the running peak and worst fall

Kelly fraction f* = (b·p − q) / b,  b = D − 1

runs are seeded, so identical inputs give identical output.

Paths are simulated one bet at a time because percentage and Kelly stakes depend on the bankroll at that moment — the path is the calculation.

A worked example

A 55% bettor at −110 over 500 bets, flat $100 on a $10,000 bankroll. Break-even is 52.38%, so the edge is 2.62 points — real, and roughly what a strong bettor achieves.

The median season finishes at $12,500, up 25%. But the 10th percentile lands just below the starting bankroll, and about one season in nine finishes down. A genuine, sustainable edge — and a losing year is a normal outcome, not evidence that anything broke.

The drawdowns are the part worth internalising. The typical season dips 13.8% below its peak at some point along the way, and roughly one in twelve falls 25% or more. Everyone who quits a working strategy quits during one of those, which is the reason to look at this chart before the season rather than during it.

Common questions

Why does my winning strategy lose money in so many simulations?
Because 500 bets is a small sample against a two-or-three-point edge. Variance dominates at that scale, and losing seasons are a normal outcome for a genuinely profitable bettor. That is the single most useful thing this page shows.
Should I read the mean or the median?
The median. The mean is pulled upward by a few extraordinary runs, particularly under compounding staking plans. The median season is the one in the middle, and it is what you should plan around.
Which staking plan is best?
Kelly grows fastest if your edge estimate is right, and punishes you hardest if it is not. Percentage staking is a reasonable middle. Flat is the most predictable and the only one where strict ruin is possible. Run all three with your own numbers and compare the 10th percentile, not the median.
Why is the ruin probability zero on percentage staking?
Because the stake shrinks with the bankroll, so it never fails to cover the next bet. The bankroll can fall to almost nothing without technically going broke — which is why the drawdown figures matter more than the ruin figure there.
How accurate is this?
It is exactly right about the variance of the model it is given, and the model is simplified: constant edge, constant price, independent results. Real betting has none of those, so the real spread is wider. Treat this as a floor on the uncertainty.
Why do I get the same result every time?
The random number generator is seeded, so identical inputs always produce identical output. That makes the results reproducible and stops the figures twitching while you adjust an input.

The guide behind this calculator

Related calculators

For informational and analytical purposes only. These tools do not predict outcomes and do not recommend wagers.