Towers, shapes and favourite numbers
If you would rather play shapes, towers or favourite numbers, stack chips on a few streets and size the ladder to your bankroll: that stays ahead for 500 spins as well as anything we tested.
Chapter 6 of the spinner roulette research (chapters 0 to 9): Towers, shapes and favourite numbers.
Three ladders passed every test. a1 widens: $1 on street 1-2-3, then one more street after each loss, up to nine streets (27 numbers) at $288 each. b3 stacks: two streets, 1-2-3 and 16-17-18, with a tower growing from $2 to $383 on each over 29 steps. c9 does both: one street, then two, three and four (1 to 12), then stacks those four from $3 to $411 each over 16 steps.
All three start from $5,000, need $1 table minimums and stop when the ladder is lost. Over four seeds the typical session ends up about $1,500 to $1,700, the average is always negative, and a third to nearly half of sessions lose the ladder.
Open this chapter in spinner, with the interactive chartsTest your own system
The numbers
| Ladder | Typical session | Chance of ending ahead | Average | Worst 5 % of sessions | Lost the ladder |
|---|---|---|---|---|---|
| c9, hybrid | +$1,540 (4 seeds: +$1,537 to +$1,542) | 67.1% (4 seeds: 66.6% to 67.2%) | -$295.45 (4 seeds: -$326.25 to -$295.45) | -$4,820.13 (4 seeds: -$4,827.38 to -$4,819.17) | 32.7% (4 seeds: 32.6% to 33.2%) |
| b3, stack | +$1,679 (4 seeds: +$1,679 to +$1,698) | 60.8% (4 seeds: 60.8% to 62.1%) | -$320.90 (4 seeds: -$320.90 to -$245.78) | -$4,491.05 (4 seeds: -$4,491.05 to -$4,483.41) | 38.7% (4 seeds: 37.3% to 38.7%) |
| a1, widen | +$1,494 (4 seeds: +$1,491.50 to +$1,498) | 55.1% (4 seeds: 54.6% to 55.3%) | -$330.93 (4 seeds: -$351.14 to -$324.08) | -$3,380.37 (4 seeds: -$3,384.89 to -$3,377.19) | 44.9% (4 seeds: 44.6% to 45.3%) |
| Ladder | Levels | First stake per bet | Top stake per bet | Whole ladder costs |
|---|---|---|---|---|
| c9, hybrid | 16 | $1 | $411 | $4,919 |
| b3, stack | 29 | $2 | $383 | $4,574 |
| a1, widen | 9 | $1 | $288 | $3,444 |
Takeaways
- Stacking a narrow shape works as well as anything tested: b3’s typical session is +$1,679 to +$1,698.
- Widening across the board is the weakest of the three: a1 ends ahead least often (54.6 % to 55.3 %), though its worst 5 % is the mildest.
- A lost ladder costs $3,444 (a1), $4,574 (b3) or $4,919 (c9) in one go.
The math
The deficit recursion generalises to mixed layouts: after a loss at level k with deficit D(k), the stake at the next level must make the smallest-paying covered bet return at least D(k) + τ. A street pays 11 to 1, so with n streets at $a a hit nets 12a − n·a − D. a1 level 3: D = $11, four streets need 8a − 11 ≥ 11, so a = $3.
Widening raises the per-spin win chance (27 numbers at the top: p = 0.730) but puts most of the risk on the last level: $2,592 of a1’s $3,444 cap loss. Survivor parameters (q, B in cycle profits, k): b3 0.00591, 196.7, 1.17; c9 0.00356, 338.8, 1.21; a1 0.00494, 258.1, 1.28. Predicted P(no cap at 500) = e^(−k·x): 0.617, 0.672, 0.547; measured over three seeds 0.622, 0.672, 0.551. Medians across four seeds move by at most $19 (b3), $5 (c9) and $6.50 (a1); every mean is negative.
n streets at $a: 12a − n·a − D(k) ≥ τ
P(no cap) = e^(−k·p·S/B)
b3: e^(−1.17 × 0.162 × 500 / 196.7) ≈ 0.617