Is Plinko rigged?

Plinko triggers the “this is rigged” instinct harder than almost any other casino game, and for an understandable reason: you watch a ball bounce down a board and it lands in the low-paying middle slots again, and again, and again, while the juicy edge multipliers sit there untouched. It feels steered. But on a provably fair site, “feels rigged” and “is rigged” are very different claims — and the second one is testable. So let's test it. We'll recompute a Plinko drop from its seeds, exactly the way the casino does, and see whether the math lines up.

Not sure which to choose? See what we recommend for each type of player.

How provably fair Plinko decides the ball path

A Plinko board has a fixed number of peg rows — Stake Plinko and Shuffle let you choose any count from 8 to 16, and this walkthrough uses 16. At each row the ball goes left or right, and after the last row its final slot is simply how many times it went right (0 rights lands in the far-left slot, 16 rights in the far-right, and everything else in between). Crucially, none of this is physics or pre-stored — it's a deterministic function of three inputs you can reproduce.

Before your bet, the casino commits a secret server seed by showing you only its SHA-256 hash. You set a client seed, and a nonce counts each bet. The game runs HMAC-SHA256 over these and reads off one random number per row; each number, compared against a 50/50 split, decides left or right at that row. Because the server seed was locked before you dropped the ball, the entire path was already determined — the casino can't nudge it mid-fall without the revealed seed later failing its hash check. That's the same commit-and-reveal lock explained in our provably fair primer.

The verification: a real drop, recomputed

Here's a full, reproducible 16-row Plinko verification using the same Stake-style algorithm our provably fair verifier implements. The seeds are illustrative teaching values rather than a captured live bet — but every number is real: enter the same inputs (game: Plinko, 16 rows) and you'll get the identical path and slot.

Step 1 — confirm the commitment. Hash the revealed server seed with SHA-256. It produces e5a2b6a4…217ebe — matching the hash the casino published before the drop. The seed was not swapped.

Step 2 — recompute the path. HMAC-SHA256 over dropzone:0:0 keyed on the server seed yields the per-row numbers. Mapping each to left (<0.5) or right (≥0.5) gives this exact 16-row path:

Step 3 — read the slot. Count the rights: there are 7. So this drop lands in slot 7 of 0–16 — just left of centre. If the casino had paid out on a different slot for these seeds, the mismatch would be cryptographic proof of tampering. It matches. The drop was provably fair. That's the entire verification, and you can run it yourself in under two minutes.

Why Plinko lands in the middle (and why that's fair)

Now the part that actually dissolves the “rigged” feeling. With 16 independent 50/50 bounces, the final slot follows a binomial distribution — and the number of distinct paths to each slot is wildly lopsided toward the centre. There is exactly one path that ends in the far-left slot (all 16 lefts) and one to the far-right (all 16 rights), but there are 12,870 different paths that land in the dead-centre slot 8. The ball isn't being herded to the middle — the middle simply has thousands of times more ways to be reached.

To make the scale concrete, here's how often each region comes up on a fair 16-row board:

That ~0.003% is why the top multiplier is so large — it has to be, because reaching the corner is staggeringly rare. The clustering you see toward the middle isn't the game cheating; it is the signature of an honest, unbiased board. A genuinely rigged board would actually be easier to spot: the slot frequencies would drift away from this binomial curve, and a few hundred verified drops would expose it. On a provably fair Plinko, every individual drop is checkable, and the aggregate matches the math.

The other way to tilt a board: the payout table

The seed check answers one question: was the slot fixed before the drop? It does not touch a second one: are the multipliers a fair price for those odds? A board could pass every hash check and still pay 80 cents on the dollar, because the edge of Plinko lives entirely in the numbers printed on the slots. That part is not hidden behind cryptography. It is arithmetic anyone can do from the paytable.

The return of a table is each slot’s multiplier times its chance, added up. Take the 8-row Low table that Stake and Shuffle both publish: 5.6x, 2.1x, 1.1x, 1x, 0.5x, 1x, 1.1x, 2.1x, 5.6x. Eight rows give 256 equally likely paths:

Multipliers: Stake payout table (1 Nov 2026) and Shuffle Plinko game page (12 Nov 2026). Chances and products are arithmetic.

That table returns 98.984%, an edge of just over 1%. Running the same sum over all 36 tables Stake publishes (nine row counts, four risk levels) gives returns between 98.906% and 99.160%, against a stated 99%. Shuffle’s 27 tables are the same Low, Medium and High numbers and cover the same range, against a stated edge of "around 1%". The small spread comes from rounding multipliers to tidy figures.

This is the test that would expose a board that is honest in its randomness and dishonest in its price. If a casino stated 99% and its paytable summed to 94%, the hashes would all verify and the game would still be taking five times what it claimed. On the tables above, the stated figure and the sum agree to within a fraction of a percentage point. The per-casino detail is on our Stake Plinko and Shuffle Plinko pages.

What four casinos publish about Plinko

"Provably fair Plinko" is not one algorithm. The casinos we cover document four constructions, and they differ in what a player can check and with which tool.

Stake: fairness documentation, payout table and game page, 1 Nov 2026. Shuffle: game page and help centre, 12 Nov 2026. Roobet: game page 1 Nov 2026, fairness page and verifier 12 Nov 2026. Gamdom: game page 1 Nov 2026, sample code 12 Nov 2026.

Stake and Shuffle work exactly as the walkthrough above: a number per row, doubled and rounded down. The demonstration in this article would give the same path on either site, because Shuffle’s help centre documents the same HMAC-SHA256 function and the same step rule as Stake.

Gamdom does not simulate a path at all. Its sample code computes one HMAC-SHA512 over the client seed and nonce, reduces it to a number below one million, and reads the slot from a cumulative table. For 8 rows the table starts 3,906, 35,156, 144,531, 363,281, 636,719. The gaps between those entries are 3,906, 31,250, 109,375, 218,750 and 273,438 per million, which are 1, 8, 28, 56 and 70 paths out of 256 to six decimal places. So the slot odds match the binomial odds of a real board; what you cannot do is replay the bounces, because there are none in the calculation.

Roobet adds something the others lack. Its fairness page says the Game Hash of the current game board changes periodically, and its verifier contains code that generates the Lightning mode’s payouts from that hash. A Lightning drop therefore has two things to verify, the ball’s path and the board it fell through, and both go through Roobet’s verifier. Our Roobet checker confirms only that the revealed server seed matches the hash shown beforehand.

One category is outside all of this. Plinko-branded games from outside studios, such as the "Magic Plinko" and "Triple Pot Plinko - Hercules" titles named in Gamdom’s list of games excluded from bonus wagering, are supplier games. They do not use your seed pair, and their fairness rests on the supplier’s testing and licence, not on a hash you can recompute. The question "is Plinko rigged?" has a checkable answer only for the in-house versions.

Sources: Stake fairness documentation, game events (1 Nov 2026) and Stake Plinko payout table (1 Nov 2026); Shuffle Plinko game page and Shuffle game events (12 Nov 2026); Roobet fairness page and its Plinko verifier (12 Nov 2026); Gamdom fairness page, its Plinko sample code and Deposit Bonus Guide (12 Nov 2026). Table returns are our arithmetic.

Verify your own Plinko drop in two minutes

Step 1 — record the hash. Before playing, copy the hashed server seed from the fairness panel. That's your receipt.

Step 2 — set your client seed. Type something unpredictable so the casino couldn't have precomputed a path against a chosen seed.

Step 3 — play, then rotate. After your session, rotate your seed pair so the casino reveals the retired server seed.

Step 4 — check both locks. Hash the revealed seed and confirm it matches the pre-published hash, then enter the server seed, your client seed and the bet's nonce into our provably fair verifier (set to Plinko, matching row count) and confirm the recomputed slot matches what you were paid. Our dedicated verify provably fair Plinko walkthrough covers each field, and the Stake Plinko strategy guide covers risk levels and row counts. Beware any “Plinko predictor” — the landing slot is unknowable before the drop, so every predictor is a scam, just like the Mines bots we debunked.

Frequently asked questions