Choose your two hole cards, add the flop, turn, or river, and estimate your Texas Hold'em poker odds. The calculator deals the remaining cards thousands of times to show your winning chances and the most likely final hand combinations. Calculations run in your browser.
How to Calculate Your Poker Odds
- Select your two cards on the table to start the estimate.
- Add any community cards that have been revealed; the estimate updates after each card.
- Set the total number of players, including yourself.
The remaining cards are dealt at random without replacement. Each simulated deal completes the board and gives two unseen cards to every opponent. The calculator evaluates the best five-card hand available from each player's two hole cards and the five community cards, following standard Texas Hold'em rules.
Winning Chances and Expected Pot Share
Win means you finish as the sole winner. Tie means you share the best hand with at least one opponent. Lose means another player has a better hand. The expected pot share includes your fraction of every split pot: a two-way tie contributes one half, a three-way tie one third.
All opponents are assumed to play uniformly random hands. The estimate does not account for betting behavior, folded cards, hand ranges, bluffing, rake, or side pots. It is not a prediction of how a particular opponent will play.
Final Hand Probabilities
The combinations are sorted by their observed frequency in the simulated deals. They describe your best final five-card hand at the river, including community cards. A full house counts as a full house, not also as a pair and three of a kind. Royal flush is listed separately from other straight flushes. On the river, your hand category is already known, while winning chances may still vary with the opponents' unknown cards.
The gold frames identify the five cards that currently make your best hand, including any necessary kickers. This may use both hole cards, one hole card, or only community cards. Before the flop there are not yet five known cards, so no complete hand is highlighted. When several equally strong five-card selections exist, one valid selection is highlighted.
The Probability Model
Let \(b\) be the number of known community cards and \(r\) the number of opponents. Your two hole cards are known, leaving \(N = 50-b\) unseen cards. The board needs \(m = 5-b\) more cards. First choose those board cards, then choose an unordered pair for each distinct opponent seat. The number of possible completions is
\[|\Omega| = \binom{N}{m}\prod_{j=0}^{r-1}\binom{N-m-2j}{2}.\]Every completion has the same probability under the random-opponent model. Within a completion, cards are drawn without replacement, so a card cannot occur twice. The order of unrevealed board cards does not affect a showdown hand. Burn cards need not be explicitly simulated: after averaging over their unknown identities, the remaining visible cards still have the same uniform distribution.
For a final-hand category \(A\), the exact conditional probability is the fraction of completions that produce it:
\[p_A = \frac{1}{|\Omega|}\sum_{\omega\in\Omega}\mathbf{1}\{H(\omega)=A\}.\]Here \(H(\omega)\) is the category of your best five-card hand. The categories are mutually exclusive and exhaustive, so their probabilities sum to one. Opponent count changes your chance of winning, but not the marginal distribution of your own completed hand under this model.
Monte Carlo Estimates and Split Pots
Instead of visiting every completion, the calculator independently samples \(n\) random deals. It uses 5,000 deals on mobile-sized or touch-oriented devices and 10,000 on desktop devices. Cards may recur in different deals, but never twice within the same deal. With \(X_i=\mathbf{1}\{H(\omega_i)=A\}\), the hand-probability estimate is
\[\widehat p_A = \frac{1}{n}\sum_{i=1}^{n}X_i.\]For each deal, the three outcomes are an outright win, a tie for the best hand, or a loss. Their indicator averages give the corresponding displayed percentages. To calculate expected pot share, define \(W_i=0\) if you lose and \(W_i=1/k_i\) if you are one of \(k_i\) equally ranked winners. Then
\[\widehat e = \frac{1}{n}\sum_{i=1}^{n}W_i.\]Consequently, a tie contributes its actual share rather than being counted as either a full win or a loss. This assumes a single common pot and no rake; it does not model different all-in amounts or side pots.
Accuracy and Rare Hands
For a fixed category, \(X_i\) is a Bernoulli random variable. Independent deals give an unbiased estimate with
\[\mathbb{E}[\widehat p_A]=p_A\] \[\operatorname{Var}(\widehat p_A)=\frac{p_A(1-p_A)}{n}.\]A normal approximation gives a 95% margin of roughly \(1.96\sqrt{\widehat p_A(1-\widehat p_A)/n}\) when both outcomes occur often enough. The largest possible standard error occurs at \(p_A=1/2\), giving a worst-case margin of \(0.98/\sqrt n\): about 1.39 percentage points for 5,000 deals or 0.98 points for 10,000. These are pointwise approximations, not a simultaneous guarantee for all displayed probabilities. They do not measure error from incorrect assumptions about opponents.
Near zero or one, the normal approximation can be misleading. If a hand occurs zero times, the display reads 0%, but zero observed does not mean impossible. The chance of observing zero occurrences when the true probability is \(p\) is \((1-p)^n\). Solving \((1-p)^n=0.05\) gives the one-sided 95% upper bound
\[p_{\mathrm{upper}} = 1-0.05^{1/n}\approx\frac{3}{n}.\]Thus zero observations in 10,000 deals still allow a true probability near 0.03% at this confidence level. Likewise, a sampled 100% winning rate does not generally prove certainty. On the river, however, your own final hand category is fixed by the seven known cards; only the opponents' cards remain uncertain. Increasing the sample count reduces typical sampling error in proportion to \(1/\sqrt n\), not \(1/n\).
The estimates are conditional on the cards you select. They are calculated as soon as both hole cards are known and update with each community card. This tool is intended for learning and hand analysis, not a guarantee of profit.