Which Plinko setting has the best odds?

None of them, in expectation: computed from the published tables, the nine boards return between 98.91% and 99.12% of the stake, all within a hair of 99%. Rows and risk choose where that return sits. Low risk on 16 rows pays at least 1× on 80% of drops and tops out at 16×; high risk on 16 rows pays below 1× on 79% of drops and tops out at 1,000×.

What are the odds of the 1,000× bucket in Plinko?

The 1,000× sits at both edges of the 16-row high-risk board. Reaching one edge needs all sixteen bounces to go the same way: 1 in 65,536. Either edge is 1 in 32,768 per drop. On 8 rows the edge pays 29× and is hit 1 in 128 drops either side: fewer rows means a smaller top prize reached far more often.

Does switching rows or risk after a loss help?

No. Every drop uses fresh HMAC-SHA256 floats and has no memory of the previous one, and every board has the same expected return. Switching changes the shape of the next drop, not its worth. The only thing that reliably changes what you lose is how much you stake in total.

Plinko guidePublished 2026-09-27

Plinko strategy: the honest maths

There is no Plinko strategy that changes the average, and here is the computation. The ball takes 8, 12 or 16 coin-flip bounces, the bucket is the count of rights, and the payout is a lookup in a published table. Every one of the nine boards returns between 98.9% and 99.1% of the stake on average. What rows and risk do change is the shape: how much of that return is concentrated in rare edge buckets and how often a drop pays back less than it cost. This guide computes all of it from the same tables the free Plinko game uses, and then works through what staking plans do on top.

How Plinko pays

Each row consumes one HMAC-SHA256 float: below 0.5 the ball goes left, above it right. With n rows there are n + 1 buckets and the chance of bucket k is C(n, k) ÷ 2^n. The multiplier is the published table for that row count and risk. Return is the sum of chance times multiplier over the buckets, and the table below prints it for every board.

What rows and risk do

Every Plinko board: return, top multiplier, chance of the edge and how often a drop pays below 1×
Rows	Risk	Return (RTP)	Top multiplier	Either edge, one in	Centre bucket pays	Drops below 1×	Std. deviation
8	low	98.98%	6×	1 in 128	0.5×	27.3%	0.56
8	medium	98.91%	13×	1 in 128	0.4×	71.1%	1.24
8	high	99.06%	29×	1 in 128	0.2×	71.1%	2.67
12	low	98.98%	10×	1 in 2,048	0.5×	22.6%	0.39
12	medium	98.99%	33×	1 in 2,048	0.3×	61.2%	1.29
12	high	99.12%	170×	1 in 2,048	0.2×	85.4%	4.38
16	low	99.00%	16×	1 in 32,768	0.5×	19.6%	0.33
16	medium	98.99%	110×	1 in 32,768	0.3×	54.6%	1.47
16	high	98.98%	1,000×	1 in 32,768	0.2×	79.0%	6.57

Return is computed from the tables and lands between 98.91% and 99.12% on every board. The standard deviation of the multiplier per drop runs from 0.33 (16 rows, low) to 6.57 (16 rows, high): a 20-fold difference in swing for the same average.

More rows spread the same return over more buckets, so the edges get rarer and bigger: the top prize goes from 29× at 1 in 128 (8 rows) to 1,000× at 1 in 32,768 (16 rows). Higher risk takes payout out of the middle and puts it in the edges: on 16 rows the centre bucket, which lands 19.6% of the time, pays 0.5× at low risk and 0.2× at high. That is why high risk feels like losing: 79% of drops return less than the stake, and the average is rescued by edges you will rarely see.

What staking plans do on Plinko

Martingale needs a clean win-or-lose to double against, and Plinko does not give one: a drop can return 0.2×, 0.5×, 1×, 1.1× or 1,000×. The nearest version people play is “double after any drop below 1×”. On 16-row high, a drop is below 1× 79% of the time, so with a 1,000 bankroll and 10 base the 7th consecutive sub-1× drop, which empties the plan, shows up in 100% of 200-drop sessions. On 16-row low it is 0.2%. And the “win” that ends the doubling might be a 1× or 1.1×, which does not recover the earlier losses at all.

The simulator models binary games, so the closest honest stand-in is Dice at the chance of a 16-row low drop returning at least 1× (80.4%): run martingale on that and compare flat betting. It flatters Plinko, because the real 1×-or-better drops mostly pay far less than the 1.232× a Dice win at that chance pays. The ruin rate you see there is a floor, not a ceiling, and the realised return on turnover is the same −1% either way.

Changing risk after a loss, alternating rows and waiting for a cold board are the other popular ideas. Each drop is independent and each board has the same expected return, so all three change which distribution the next drop is drawn from and nothing about its worth.

Plinko has settings, not strategies. Rows and risk pick a distribution; every distribution on offer returns about 99% of the stake. A staking plan picks the next stake before the bounces are known, so its expected loss is 1% of the total staked, whatever pattern it makes. Nothing you choose moves the −1%.

Understood the maths? Craze runs the real thing.

Same published odds, real stakes. PlinkoLab is sponsored by Craze; whether to play for money is entirely your call.

Play at Craze

Questions

Which Plinko setting has the best odds?: None of them, in expectation: computed from the published tables, the nine boards return between 98.91% and 99.12% of the stake, all within a hair of 99%. Rows and risk choose where that return sits. Low risk on 16 rows pays at least 1× on 80% of drops and tops out at 16×; high risk on 16 rows pays below 1× on 79% of drops and tops out at 1,000×.
What are the odds of the 1,000× bucket in Plinko?: The 1,000× sits at both edges of the 16-row high-risk board. Reaching one edge needs all sixteen bounces to go the same way: 1 in 65,536. Either edge is 1 in 32,768 per drop. On 8 rows the edge pays 29× and is hit 1 in 128 drops either side: fewer rows means a smaller top prize reached far more often.
Does switching rows or risk after a loss help?: No. Every drop uses fresh HMAC-SHA256 floats and has no memory of the previous one, and every board has the same expected return. Switching changes the shape of the next drop, not its worth. The only thing that reliably changes what you lose is how much you stake in total.

Keep going

Last reviewed 2026-09-27