Spunk Odds Reference

Plinko guide

Tsunami Plinko: when the house edge changes

The tsunami label is a page name only; the board is the same 50/50 model as everywhere on this site. This page asks how a pay table changes when the edge built into it changes, and what that costs over time.

The starting table

Hypothetical 13-row table used as the baseline
SlotPathsProbabilityIllustrative pay
010.0122%42.00x
1130.1587%13.00x
2780.9521%5.00x
32863.4912%2.40x
47158.7280%1.20x
51,28715.7104%0.80x
61,71620.9473%0.50x
71,71620.9473%0.50x
81,28715.7104%0.80x
97158.7280%1.20x
102863.4912%2.40x
11780.9521%5.00x
12130.1587%13.00x
1310.0122%42.00x

Expected return of this illustrative pay table: 98.46% of the stake. It is a teaching example, not any operator's schedule.

The baseline returns 98.46%, which means a house edge of 1.54%. The slot chances do not depend on the table, so changing the edge can only be done by changing the multipliers.

Scaling every multiplier

The simplest way to change the edge is to multiply every pay by the same factor. The return scales by exactly that factor, because the return is a sum of chance x pay and the factor comes out of the sum. The shape of the table, its symmetry and the ratio between any two slots stay the same.

Scaling every multiplier of the hypothetical 13-row table by the same factor
Scale factorWall payReturnHouse edgeExpected net, 500 drops
1.0042.00x98.46%1.54%-7.7
0.9841.16x96.49%3.51%-17.5
0.9640.32x94.52%5.48%-27.4
0.9439.48x92.55%7.45%-37.2

A factor of 0.96 looks like a small trim: the wall pay drops from 42.00x to 40.32x. But the edge more than triples, from 1.54% to 5.48%. Over 500 drops the expected loss grows by the same proportion. Small changes to the multipliers matter much more than they look.

Trimming only the middle

An operator could also lower just the most common slots. Cutting the two centre pays (slots 6 and 7) from 0.5x to 0.4x lowers the return by 0.1 x 41.89% = 4.19% of the stake. A cut of the same size on the walls would remove only 0.0024%. So two tables that look identical at the edges can differ a lot in cost, depending on decimals in the middle.

How to compare two tables