How This Simulator Works

Draws use the browser's crypto.getRandomValues, the same generator used for security purposes, rather than the weaker Math.random. Random integers are produced with rejection sampling so every value in a range is exactly equally likely, with no modulo bias. The 80 numbers are shuffled with the Fisher-Yates algorithm and the first 20 become the draw, which is statistically identical to drawing balls from a hopper without replacement.

Published odds come from the combinations formula. The probability of exactly h hits on an n-spot ticket is C(n, h) times C(80 minus n, 20 minus h), divided by C(80, 20), where C(80, 20) equals 3,535,316,142,212,174,320. We compute these counts with arbitrary-precision integers, so figures like the 1 in 8,911,711 solid ten are exact rather than rounded folklore. The same functions power the odds calculator, the odds tables, and the expected returns on the paytables page.

Every simulator run is driven by a random seed, and the Share Last Run button encodes that seed in the link. Anyone opening the link replays the identical draw sequence and gets the identical result, on any device. This makes results checkable rather than take-our-word-for-it.

An automated test suite runs against the math on every change. It verifies known combination values, confirms that hit probabilities sum to one for every pick count, checks draw validity across thousands of draws, and confirms that no paytable preset returns more than it takes over the long run. In live checks, ten thousand draw sessions land within a fraction of a percent of the theoretical win rate.

A simulator shows how randomness behaves. It cannot predict any specific future draw, here or in any real game, and nothing on this site improves real-game odds. Real keno games also use their own paytables, which change the money math even though the draw odds stay the same.