ГСЧ randomness lab

Poker

Coin-flip duel

Two players, different chances of winning, and hundreds of rounds — a clear visualization of the advantage.

~0.5 s server-side calculation without reboot
Player 1 has a 55% advantage on each roll, but in a single round can lose: the result of the round is the sum of a hundred random rolls, not a single roll.
Duel results Waiting for the launch
—player 1's win rate

Set the bias, the number of throws and rounds, then start the simulation. The results will appear here without reloading the page.

— player 1 wins
— victories of player 2
— draws
— player 1's win rate
— player 2's win rate
— average edge 1
— average edge 2
— max. win streak 1
— max. win streak 2
The calculation has not been started yet.
Accumulated victories by rounds
Distribution of series series length → how many times
Player 1 —
Player 2 —

A streak is consecutive rounds won. A draw or an opponent's victory interrupts it. The longest streaks are grouped in the "24+" bucket.

Last 15 rounds new on top
Round Player 1 Player 2 Winner Edge
Run the simulation — the last 15 rounds will appear here.

Coin duel: how probability turns into outcome

"Duel on a Coin" is the oldest site simulator. Two players agree that the first one has a slightly higher chance of winning each toss, and they flip the coin many times. The task is to see how a small advantage behaves over the long run: where luck ends and mathematics begins.

What happens in one round

A series of independent tosses is performed in a round. The player with more 'heads' wins; if the numbers are the same — it's a tie. The difference between 'heads' and 'tails' is the round advantage: 6:4 is an advantage of 2, 7:3 is an advantage of 4.

The more throws in a round, the less frequent ties occur and the more consistently the player with the advantage wins. With a single throw, the round is just that one throw, so the proportion of wins almost matches the given advantage.

Why a 55% edge does not guarantee a win

An advantage of 55% means that, on average, player 1 gets 55 "heads" out of 100. But in a particular round, he may get 48, 52, or 61 — this variance. Randomness does not have to coincide with expectation in the short term: this is why a player with an advantage sometimes loses several rounds in a row.

The simulator also shows the reverse side: over the long distance, the advantage almost always manifests, while a 'losing streak' can be longer than it intuitively seems.

How to read the results

  • Win rate — how many rounds each player won out of the total number.
  • Average edge — how large the rounds were won on average.
  • Maximum series — the longest winning streak in a row.
  • Distribution of series — how many times series of each length occurred.
  • Accumulated victories — a graph showing how the gap grew.

Compare the scenarios: 55% on 10 throws and 55% on 1000 throws — these are two different games in terms of result stability.

Frequently Asked Questions
Why does player 1 win more than half of the rounds with a 55% advantage?
A 55% edge applies to each individual throw. A round consists of a hundred throws, and the winner is the one who gets more "heads": over a long series, a small advantage accumulates, so the share of rounds won ends up being noticeably higher than 55%. Make a single throw in a round — and you will see exactly the percentages that were set.
What does 'average advantage' mean?
This is the average difference between "heads" and "tails" in the rounds that the player won. For example, a value of 3.4 means that on average he won with a margin of 3.4 tosses. Ties and lost rounds are not included in this number.
Can a player with an advantage lose several rounds in a row?
Yes, and the simulator shows this. The distribution chart of streaks answers the question 'how often does a streak of 5, 10, 15 losses in a row occur.' The shorter the round, the longer the streaks tend to be: losing 10 times in a row in a single throw is a common occurrence.
How many rounds should be counted?
For a stable picture, 500–1000 rounds of 100 throws each are enough. The more rounds, the closer the actual percentages are to the theory: this is how the law of large numbers works. But even with 2000 rounds, there is still a variation, and that is normal.
Is this a gambling game?
No. 'Coin Duel' is an educational simulator of probability theory and poker variance. Bets are not accepted, a prize fund is not formed, winnings are not paid out, and the results have no monetary value.
↑↓ navigation Enter to open Esc close