Not sure which to choose? See what we recommend for each type of player.
What Stake publishes about Wheel
Wheel is a one-spin game with two settings. You choose a risk level, low, medium or high, and a number of segments, 10, 20, 30, 40 or 50; the wheel stops on one segment and pays the multiplier printed on it. Stake's game page states the price in one sentence: "Wheel features a theoretical return to player percentage of 99%, meaning this game has a house edge of just 1%" (1 Nov 2026).
Unlike most games, the whole paytable is in the fairness documentation. Stake's game-events page prints an array of multipliers for each of the 15 combinations, and the same arrays are what the game itself loads. Segment 0 is the first entry, and the arrays are the wheel, in order.
The fifteen wheels
Low risk is one pattern of ten segments, repeated: 1.5x, 1.2x, 1.2x, 1.2x, 0x, 1.2x, 1.2x, 1.2x, 1.2x, 0x. A 20-segment wheel is that pattern twice, a 50-segment wheel five times. On Low the segment count therefore changes nothing but the picture: every Low wheel pays 1.2x on 70% of spins, 1.5x on 10% and nothing on 20%.
High risk has a single paying segment, the last one. It pays 9.9x on 10 segments, 19.8x on 20, 29.7x on 30, 39.6x on 40 and 49.5x on 50, which is 0.99 times the number of segments in every case.
Medium risk alternates losing and paying segments, so half of all spins pay nothing, and the paying half differs by wheel:
| Segments | Paying segments (the rest pay 0x) | Top multiplier | Chance of the top |
|---|---|---|---|
| 10 | 1.5x on 2, 1.9x on 1, 2x on 1, 3x on 1 | 3x | 10.00% |
| 20 | 1.5x on 2, 1.8x on 1, 2x on 6, 3x on 1 | 3x | 5.00% |
| 30 | 1.5x on 6, 1.7x on 1, 2x on 6, 3x on 1, 4x on 1 | 4x | 3.33% |
| 40 | 1.5x on 8, 1.6x on 1, 2x on 7, 3x on 4 | 3x | 10.00% |
| 50 | 1.5x on 13, 2x on 8, 3x on 3, 5x on 1 | 5x | 2.00% |
Stake Wheel payout arrays, fairness documentation and game data (1 Nov 2026).
Averaging the multipliers of any one wheel gives its return, because every segment is equally likely. All fifteen come to exactly 99.00%. The odd-looking values on Medium, 1.9x on the 10-segment wheel, 1.8x on 20, 1.7x on 30 and 1.6x on 40, are there to make each sum land on 0.99 times the segment count.
| Wheel | Spins that pay | Top multiplier | Chance of the top | Standard deviation per spin | Return |
|---|---|---|---|---|---|
| Low, 10 segments | 80.0% | 1.5x | 10.00% | 0.50 | 99.00% |
| Low, 30 segments | 80.0% | 1.5x | 10.00% | 0.50 | 99.00% |
| Low, 50 segments | 80.0% | 1.5x | 10.00% | 0.50 | 99.00% |
| Medium, 10 segments | 50.0% | 3x | 10.00% | 1.06 | 99.00% |
| Medium, 30 segments | 50.0% | 4x | 3.33% | 1.09 | 99.00% |
| Medium, 50 segments | 50.0% | 5x | 2.00% | 1.13 | 99.00% |
| High, 10 segments | 10.0% | 9.9x | 10.00% | 2.97 | 99.00% |
| High, 30 segments | 3.3% | 29.7x | 3.33% | 5.33 | 99.00% |
| High, 50 segments | 2.0% | 49.5x | 2.00% | 6.93 | 99.00% |
Arithmetic from Stake's published arrays. Standard deviation is in units of the stake.
This is the whole of the decision. Risk level and, on Medium and High, segment count set how results are spread: Low has a deviation of half a stake per spin, High with 50 segments nearly seven stakes. The expected cost is one cent per dollar on every one of the fifteen.
How long High can go without paying
On a High wheel the chance of a paying spin is one in the number of segments, so dry runs are the normal state.
| Segments | Pays | No win in 20 spins | No win in 50 spins | No win in 100 spins | Spins for an even chance of one win |
|---|---|---|---|---|---|
| 10 | 9.9x | 12.2% | 0.52% | 1 in 38,000 | 7 |
| 20 | 19.8x | 35.8% | 7.69% | 0.59% | 14 |
| 30 | 29.7x | 50.8% | 18.4% | 3.37% | 21 |
| 40 | 39.6x | 60.3% | 28.2% | 7.95% | 28 |
| 50 | 49.5x | 66.8% | 36.4% | 13.3% | 35 |
Arithmetic: each spin is independent, so the chance of n losses in a row is (1 − 1/segments) to the power n.
With 50 segments a player has only an even chance of a single 49.5x within 35 spins, and about one session in eight sees none in 100. Those runs do not mean the wheel is behind schedule, and a paying spin does not mean the next is less likely; the segment for each spin is computed from the seeds and the nonce alone.
How the segment is chosen
Stake's documentation gives the step as PAYOUTS[segments][risk][float * segments]: take one random number between 0 and 1, multiply by the segment count, round down, and read the multiplier at that position in the array. The number comes from the standard generator: HMAC-SHA256 with the server seed as key and clientSeed:nonce:0 as message, the first four bytes turned into a fraction. Wheel is on Stake's list of games that use a single cursor.
Two consequences are easy to check on any revealed bet. The risk level does not change the random number, only the array it indexes, so a spin that landed on position 7 would have landed on position 7 at any risk. And on a High wheel the bet wins only when the number falls in the top slice: at 10 segments, 0.9 or above.
Stake shows the server seed as a SHA-256 hash before the bet and reveals it when the seed pair is rotated. Its calculation page lists Wheel with risk and segment selectors. Our provably fair verifier has no Wheel mode; it confirms that the revealed seed matches the hash, and the position itself has to be worked from the step above or with Stake's page.
Wheel at other casinos
Four of the larger casinos covered on this site run a house Wheel, and they do not price it alike.
| Stake | Shuffle | Roobet | BC.Game | Rainbet | |
|---|---|---|---|---|---|
| Stated price | 99% RTP | No figure on the game page | House edge 2% in the game configuration | 98 in the game data; "just 1%" edge in the help article | "Max RTP 96.40%" |
| Return from the published tables | 99.00% on all 15 | 99.00% on all 15 | 98.00% on all 15 | 98.00% on the tables in use; the verifier also holds 99% and 96% sets | 96.4% for one spin on Low |
| High risk, 10 segments | 9.9x on 1 segment in 10 | Shown in the game | 10x, hit on 9.8% of spins | 9.8x on 1 segment in 10 | Different game design |
| How the edge is taken | In the multipliers | In the multipliers | A share of spins is sent to a 0x segment | In the multipliers | On every spin of a chained round |
| Stated limits | $0.0001 minimum; profit cap per coin | $15,000 bet, $200,000 win | Bet 0.01 to 50,000; profit 3,000,000 | Shown in the bet panel | Maximum payout 10,000x |
Shuffle game page, limits and fairness articles (13 Nov 2026) and tables (1 Nov 2026); Roobet game configuration and Wheel verifier (13 Nov 2026); BC.Game help centre, game data and Wheel verifier (13 Nov 2026); Rainbet game page and verifier (1 Nov 2026).
Roobet takes its 2% in an unusual way. Its tables are whole numbers that average exactly 1: the Low pattern is 2, 1, 1, 1, 0 and the High wheel pays 10x, 20x, 30x, 40x or 50x. The verifier then diverts a slice of outcomes: "a house-edge-sized slice of the draw is diverted to a 0x segment, and the remainder is rescaled across all segments". So a High spin on 10 segments pays a full 10x and wins 9.8% of the time, where Stake pays 9.9x and wins 10%. The verifier's comments record a change: High-risk bets "placed before 2026-09-18T15:25:00.000Z" carried a 4% edge, and 2% since.
BC.Game describes its wheel with the same multipliers as Stake's, "One segment with a 9.9x multiplier" on High and a "house edge of just 1%", in a help article dated 28 November 2025. Its game data lists 98, and the verifier it links contains three complete sets of tables selected by a house-edge parameter: the 99% set, a 98% set (1.4848x, 1.1879x and 9.8x in place of 1.5x, 1.2x and 9.9x) and a 96% set (1.4545x, 1.1636x, 9.6x). The table the game was using on 1 Nov 2026 was the 98% one. The help article describes a price the game no longer offers.
Rainbet lets a round chain up to ten spins, each multiplying the last. Its verifier code applies the edge on every spin, so the return of a round on Low is 0.964 raised to the number of spins: 92.9% for two, 83.3% for five and 69.3% for ten (arithmetic from the code, 1 Nov 2026). The "Max RTP" on the page is the one-spin figure.
Smaller sites reviewed here list a wheel among their in-house games too (all as of 11 Nov 2026). BitStarz prints 96% on its Wheel page, with a €100,000 maximum win. Betpanda's default table gives its Wheel 98%. Wild.io says its Wheel's RTP varies by risk mode and lists a 48x top multiplier, and Winna publishes an $8,000 maximum bet and $400,000 maximum win without a return figure. 500 Casino's Wheel of Fortune is a different game, with bets on four colours and a stated average edge of about 5.01%.
Limits on Stake
Stake's help centre sets the minimum Originals bet at "$0.0001 USD" and gives a maximum bet and maximum profit per coin, for example 100 BTC and 20 BTC, or 25,000,000 USDT and 2,500,000 USDT. The top Wheel prize is 49.5x, a profit of 48.5 stakes, so a High 50-segment bet above the profit cap divided by 48.5, about 0.41 BTC or 51,546 USDT, has a top prize larger than the cap. On Low, where the best segment pays 1.5x, the maximum bet is the limit that applies.
Systems, and what is left of strategy
No way of sizing or ordering bets alters the 99%. A plan that doubles after each loss on a Low wheel wins small amounts on most sessions and loses a large amount on the session where the 20% losing segments repeat; summed over all sessions the cost is 1% of the total wagered, the same as flat betting. On High, the appeal of "chasing" the single segment after a long dry run rests on an idea the mechanism rules out, since each spin's number is generated from a new nonce.
What is left is choosing the spread you want for a known price: Low for many small results, High with many segments for rare large ones, and a session budget sized to the dry runs in the table above. The Martingale simulator and the bankroll calculator put numbers on both. Between casinos, the stated price matters more than any setting: the same High spin costs 1% at Stake and Shuffle, 2% at Roobet and BC.Game.
Related: Plinko at Stake · Dice at Stake · Lowest house edge games · Casino Originals compared · Stake review
Sources: Stake Wheel game page (1 Nov 2026); Stake fairness documentation, game events and implementation (1 Nov 2026); Stake help centre, Crypto: Casino Betting Limits (article dated 24 August 2026, as of 13 Nov 2026); Shuffle Wheel game page, limits and game events (13 Nov 2026); Roobet Wheel game configuration and Wheel verifier (13 Nov 2026); BC.Game help centre, How to Play Wheel, game data and Wheel verifier (13 Nov 2026); Rainbet Wheel game page and verifier (1 Nov 2026). Returns, hit rates and streak chances are our arithmetic from the published tables.