Plinko guide
Poseidon Plinko: the bell curve behind the board
Poseidon is the page's title word, not a version of the game. The page compares the exact slot chances on a 16-row board with the smooth bell curve that a plinko board is often used to illustrate.
Why plinko looks like a bell
The slot number is the sum of 16 independent bounces, each worth 0 (left) or 1 (right). The central limit theorem says a sum of many independent pieces tends toward a normal (bell-shaped) distribution. A plinko board is a physical picture of this, which is why similar pegboards are used in classrooms to show the idea.
For R rows with p = 0.5, the slot has mean R / 2 and standard deviation the square root of R / 4. On 16 rows the mean is 8 and the SD is 2. The normal curve with those values gives a density for each slot that can be set next to the exact binomial chance.
| Slot | Exact | Normal approx. | Gap (percentage points) |
|---|---|---|---|
| 8 | 19.6381% | 19.9471% | 0.3091 |
| 9 | 17.4561% | 17.6033% | 0.1472 |
| 10 | 12.2192% | 12.0985% | -0.1207 |
| 11 | 6.6650% | 6.4759% | -0.1892 |
| 12 | 2.7771% | 2.6995% | -0.0776 |
| 13 | 0.8545% | 0.8764% | 0.0219 |
| 14 | 0.1831% | 0.2216% | 0.0385 |
| 15 | 0.0244% | 0.0436% | 0.0192 |
| 16 | 0.0015% | 0.0067% | 0.0052 |
Where the approximation is good
Near the centre the two columns agree closely. For slot 8 the exact chance is 19.638% and the normal curve gives 19.947%. Within two slots of the middle the gap stays under a few tenths of a percentage point. Ranges need more care. Slots 6 to 10 sit within one SD of the middle, and the exact chance of landing there is 78.99%. The textbook one-SD figure, 68.3%, is far too low, because the end slots 6 and 10 lie right on the boundary and count in full.
Where it fails: the walls
The normal curve has thin tails, but it does not know the board ends. At slot 16 it gives 0.00669%, while the exact chance is 0.00153%. In relative terms that is off by a factor of about 4.4. This matters because pay tables put their biggest multipliers exactly there. Pricing an edge slot from a bell curve would give the wrong answer, so always use the exact binomial for the outer slots.
A note on continuity correction
The normal curve is continuous; slots are whole numbers. A standard fix is to treat slot k as the interval from k - 0.5 to k + 0.5. For slots 6 to 10 that means the interval 5.5 to 10.5, which is 1.25 SDs either side of the mean, and the normal curve then gives 78.87%, very close to the exact figure. Reading the density at each slot number, as the table does, is a cruder shortcut that works because the SD of 2 slots is not tiny. On boards with fewer rows the fit gets worse, since the bell has fewer slots to spread across.