What is the house edge in Mississippi Stud?

4.9149%% of the ante, solved exactly here by backward induction over every deal. Published analyses give about 4.91%%.

How often should you fold the first decision?

On 412 of the 1,326 two-card hands, or 31.07%. Folding immediately is correct almost a third of the time.

When should you raise 3x on the first decision?

On exactly the pocket pairs, all 78 of them, and on nothing else. That was verified by naming the hands, not inferred from the count.

When is the 2x raise correct?

Never, at the first decision. It is optimal on zero of the 1,326 starting hands: the right answer is always 3x, 1x or fold.

Is Mississippi Stud worse than Let It Ride?

Per initial bet, yes: 4.9149%% against 3.5057%%. Both are far more expensive than blackjack at 0.5153%%.

Mississippi Stud: The House Edge, Computed

The house edge is 4.9149% of the ante, against about 4.91% in published analyses. The interesting part is not the total. It is that correct play folds 31.07% of starting hands and never raises 2x on the first decision.

Why this one needed a tree and Let It Ride did not

Let It Ride collapses into two averages because its three bets pay independently on the same final hand. Mississippi Stud does not, and the reason is a single design choice: the pay table applies to the total staked. A later raise multiplies the value of every earlier one, so the decisions are coupled and the game has to be solved backwards from the last card.

This page also exists because a note in this repository claimed the solve was "a much larger backward induction" and deferred it. That was written from intuition. Counted, the expensive part is the same shape as the Let It Ride pass that already ran: a mean over the 48 cards completing each of the 270,725 four-card states, just under 13 million evaluations. Everything above it is arithmetic. Same cost, not a larger one.

What perfect play actually does

First decision	Hands	Share of starting hands
Raise 3x	78	5.88%
Raise 2x	0	0.00%
Raise 1x	836	63.05%
Fold	412	31.07%

Two results here are worth more than the headline percentage.

The 2x raise is never correct. Not rarely, not situationally: on zero of the 1,326 starting hands. The bet sizes offered are 1x, 2x and 3x, and one of the three is decoration at this point in the hand. That follows from the shape of the decision rather than from any subtlety: the pay table multiplies the whole stake, so if the expectation is positive you want the most on and if it is negative you want the least, and the middle option is never the answer to either question.

Folding immediately is correct 31.07% of the time. A game marketed on three chances to raise asks you to give up before the first community card on close to a third of hands.

The 3x raise is correct on exactly the pocket pairs. All 78 of them, and nothing else. That is not read off the count, which would only be suggestive: the hands were named and checked.

And no folded hand contains a card above a ten. The highest card appearing in any of the 412 folds is a ten, so holding a jack, queen, king or ace means playing on, every time. Between the two results the whole opening decision fits in three lines: raise 3x with any pair, never fold with a jack or better, and otherwise raise 1x or fold. The 2x button is never the answer.

How the arithmetic works

The ante is one unit. Two hole cards are dealt, then three community cards one at a time, and before each the player folds or adds 1x, 2x or 3x the ante. Because the pay table pays on the total staked, the value of a decision node with stake S and expectation m per unit is the better of folding, which forfeits S, and raising by k, which is worth m times (S+k).

Working backwards: the mean payout over the 48 cards that could complete each four-card state gives the last decision; averaging those over the 49 possible fourth cards gives the second; averaging those over the 50 possible third cards gives the first. The expectation per ante comes out at -53223/1082900, which is 4.9149% of the ante.

The pay table this is computed from

Hand	Pays on the total staked
Royal flush	500 to 1
Straight flush	100 to 1
Four of a kind	40 to 1
Full house	10 to 1
Flush	6 to 1
Straight	4 to 1
Three of a kind	3 to 1
Two pair	2 to 1
Pair of jacks or better	1 to 1
Pair of sixes through tens	push, stake returned
Pair of fives or lower, or nothing	loses

The push line is the unusual one. A pair of sixes through tens returns your money and nothing else, which is neither a win nor a loss, and it is the only zero on the table. A pair of fives or lower loses like any other weak hand. Pay tables vary between casinos; every figure above describes the one shown.

Frequently Asked Questions

What is the house edge in Mississippi Stud?: 4.9149%% of the ante, solved exactly here by backward induction over every deal. Published analyses give about 4.91%%.
How often should you fold the first decision?: On 412 of the 1,326 two-card hands, or 31.07%. Folding immediately is correct almost a third of the time.
When should you raise 3x on the first decision?: On exactly the pocket pairs, all 78 of them, and on nothing else. That was verified by naming the hands, not inferred from the count.
When is the 2x raise correct?: Never, at the first decision. It is optimal on zero of the 1,326 starting hands: the right answer is always 3x, 1x or fold.
Is Mississippi Stud worse than Let It Ride?: Per initial bet, yes: 4.9149%% against 3.5057%%. Both are far more expensive than blackjack at 0.5153%%.

Every figure on this page is produced by mississippi_math.py, which solves the game by backward induction and refuses to write its result unless the edge lands in the band published analyses occupy. It does: 4.9149% against about 4.91%. The hand evaluator underneath it is checked separately against the exact five-card frequencies before anything here is computed. Last computed August 2026.