"1 in 4" is the strongest line on the MegaPot page. It is even true: 26.4999% of tickets win something, which is 1 in 3.774 at B = 10. It says nothing about money either, because the chance of winning and the size of the win are two different numbers, and the page is selling the first one.
Two more promises sit next to it. The documentation says roughly 77.5% goes back to players as prizes. The jackpot board says the chance is 1 in 1,425,060. I checked against the contracts, and all three numbers describe one game, just not in the way a reader expects. A $1 ticket returns about $0.79. How the draw itself works and why it can be verified is in the first piece in the series.
Documentation versus the chain
The MegaPot documentation says roughly 77.5% goes back to players as prizes, roughly 12.5% goes to backers and roughly 10% goes to referrers. The three numbers add up to 100%, and that is the only thing about them that holds together.
On chain, lpEdgeTarget() is 17.5%, not 12.5%. The referral fee of ten percent comes out before the margin, not after it. Read either way, the three numbers do not add up. I do not repeat the 77.5% as a fact: it is documentation text, not a contract parameter.
What that means in money. The documented 77.5%, multiplied by what actually reaches the pool, 0.825 out of every dollar on a direct purchase, gives $0.639 per ticket. That is a structural estimate built from the documentation. The actual $0.792164 from drawing 178 is higher, because the guaranteed minimum tops up thin rounds. The gap between the documented backer margin of 12.5% and the on-chain 17.5% is five points, or $0.05 per ticket. Against the structural $0.639 that is 7.8%, about an eighth of the return.
What "1 in 4" means
The contract has twelve tiers, meaning twelve prize levels. A tier number is 2 × (numbers matched) + (bonusball matched or not). Two of the twelve tiers pay nothing: tier 0 (no numbers, no bonusball) and tier 2 (exactly one number, no bonusball). Those two tiers take in 73.5% of all tickets. Of that share, 33.6% match no numbers at all and 39.9% match exactly one and get zero.
The guaranteed minimum of $1.111112 sits in ten tiers. Two of those tiers pay nothing but the minimum: tier 1 (zero numbers plus the bonusball, 3.7283% of tickets) and tier 4 (two numbers, 14.5257%). Together they are 18.25% of tickets. Another 8.25% get the minimum plus a premium. The remaining 73.5% get nothing.
That is the whole "1 in 4": the probability is there, but the money behind it mostly is not. The page sells the chance to win something, not the size of the payout.
The board and the dynamic bonusball
A ticket costs $1 in USDC and is issued as an ERC-721. Each ticket holds five numbers from 1 to 30 and one bonus ball (bonusball) from 1 to B. B is not a constant. The contract recalculates it every drawing.
How the probability and the payout are calculated
bonusballMax = ceil(minNumberTickets / C(30,5)), where minNumberTickets = prizePool / ((1 - lpEdgeTarget) * ticketPrice).
C(30,5) is 142,506. lpEdgeTarget() on chain is 17.5% and ticketPrice() is $1. The pool in the DrawingState.prizePool snapshot taken when drawing 179 opened was $1,131,304.46, so minNumberTickets = 1,131,304.46 / 0.825 = 1,371,278.1. Divide by 142,506 and round up, and B = 10.
The probabilities come from the contract formula: W_i = C(5,m) × C(25,5-m) × (B-1) for a tier without the bonusball and 1 for a tier with it, P_i = W_i / (C(30,5) × B).
The payout rule in the contract is a formula, not a table. First the guaranteed minimum: minimumPayout() is $1.111112, and every winning ticket receives it in full, duplicates included. The getMinPayoutTiers() list contains tiers 1, 3, 4, 5, 6, 7, 8, 9, 10 and 11, and leaves out tiers 0 and 2. The rest of the pool is the premium, split by weight: tier 3 takes 20%, tier 5 takes 16%, tier 6 takes 16%, tier 7 takes 4%, tier 8 takes 4%, tier 9 takes 4%, tier 10 takes 3%, tier 11 takes 33%. That adds up to 100%.
Board size: 142,506 × 10 = 1,425,060 combinations. That number is not fixed. It is tied to the pool. More money in the pool means a bigger B, a bigger board and worse jackpot odds. Which is why the jackpot probability and the board size move with the pool: they are not a setting of the game, but a consequence of it.
All twelve tiers
The payouts in the table are the actual values from drawing 178, where the pool was $1,128,895. The snapshot behind the board is $1,131,304.46, a difference under 0.3%, so the money below can be treated as current.
| Tier | What matched | Probability | Payout per ticket, drawing 178 | 1 in |
|---|---|---|---|---|
| 0 | nothing | 33.5544% | $0 | 3.0 |
| 1 | 0 numbers + bonusball | 3.7283% | $1.111112, minimum only | 26.8 |
| 2 | 1 number | 39.9457% | $0 | 2.5 |
| 3 | 1 number + bonusball | 4.4384% | $3.354, minimum and 20% of the premium | 22.5 |
| 4 | 2 numbers | 14.5257% | $1.111112, minimum only | 6.9 |
| 5 | 2 numbers + bonusball | 1.6140% | $6.045, minimum and 16% of the premium | 62.0 |
| 6 | 3 numbers | 1.8947% | $5.314, minimum and 16% of the premium | 52.8 |
| 7 | 3 numbers + bonusball | 0.2105% | $10.568, minimum and 4% of the premium | 475.0 |
| 8 | 4 numbers | 0.0789% | $26.330, minimum and 4% of the premium | 1,266.7 |
| 9 | 4 numbers + bonusball | 0.0088% | $228.08, minimum and 4% of the premium | 11,400.5 |
| 10 | 5 numbers | 0.0006316% | $2,365.41, minimum and 3% of the premium | 158,340 |
| 11 | 5 numbers + bonusball (jackpot) | 0.0000702% | $234,066.85, minimum and 33% of the premium, about a fifth of the pool | 1,425,060 |
Two matched numbers is the most common paying outcome: a $1 ticket gets $1.111 at a probability of 14.5257%. The jackpot on the same day is worth $234,066.85, and the chance of hitting it on one ticket is 1 in 1,425,060.
Why the jackpot does not pay out the whole pool
A tier's premium is divided by the number of winning combinations in that tier, and the denominator counts combinations nobody bought. That is the rollover mechanic: there is no separate rollover variable in the contract. It also explains why the advertised jackpot is the pool while tier 11 pays only a share of it.
In drawing 177 (September 17, 2026) the jackpot paid $233,090.86 against a pool of $1,125,957. In drawing 178 (September 18, 2026) it paid $234,066.85 against a pool of $1,128,895. I pulled both figures from the contract, from the getDrawingTierPayouts function. The jackpot in 178 is about a fifth of the pool, not the whole pool. Tiers 1 and 4 pay exactly the $1.111112 minimum, because those tiers carry no premium weight, only the guarantee.
What a ticket actually returns
EV is the sum of the contributions of all twelve tiers. I take the probabilities from the table at B = 10 and the actual payouts from drawing 178.
How $0.792164 comes out
EV = Σ P_i × payout_i across all twelve tiers. I work out each tier separately and add them up: tier 1 gives $0.041425, tier 3 $0.148858, tier 4 $0.161397, tier 5 $0.097566, tier 6 $0.100688, tier 7 $0.022248, tier 8 $0.020786, tier 9 $0.020007, tier 10 $0.014939, tier 11 $0.164251. Tiers 0 and 2 give nothing.
The sum across the twelve tiers is $0.792164 for every dollar staked. For drawing 177 the same sum gives $0.790068.
So a $1 ticket returns about $0.79, and the margin is 1 - 0.792164 = 20.8%, about 21%.
Here is where the return comes from. Tier 4, where two numbers are enough, gives $0.1614. Tier 3, one number plus the bonusball, gives $0.1489. The jackpot, at 1 in 1,425,060, gives $0.1643, more than any other tier. The two tiers that do not pay give nothing, and they hold 73.5% of tickets.
More tickets do not help
The average across tickets does not change, while the spread around it narrows. One ticket with the drawing 178 payouts has a mean of $0.7922 and a standard deviation of $196.18. A single ticket's mean tells you almost nothing: 20.73% of the whole EV sits in the jackpot, at 1 in 1,425,060. What that produces for syndicates, and why money still circles the game, is in the third piece.
| Tickets | Spent | Expected return | Delta |
|---|---|---|---|
| 1 | $1 | $0.79 | minus $0.21 |
| 100 | $100 | $79.39 | minus $20.61 |
| 10,000 | $10,000 | $7,938.64 | minus $2,061.36 |
| 100,000 | $100,000 | $79,386.44 | minus $20,613.56 |
| 1,425,060 (the whole board) | $1,425,060 | $1,131,304.46 | minus $293,755.54 |
Standard deviation and the probability of profit
E[S_N] = N × E[X], Var(S_N) = N × Var(X), so the standard deviation grows as the square root of N. The return in the table for small N is N × $0.7922 from the actual drawing 178 payouts, while the last row is the whole pool, because the full board takes every tier at once: the two estimates differ by less than 0.3%. SD is √N × $196.18. The probability of profit comes from a normal approximation, and it is generous here, because the distribution is skewed.
| Tickets | SD | SD as % of return | P(profit) |
|---|---|---|---|
| 1 | $196.18 | 24,764% | 49.96% |
| 100 | $1,961.80 | 2,476% | 49.58% |
| 10,000 | $19,618.00 | 248% | 45.78% |
| 100,000 | $62,038.90 | 78.3% | 36.88% |
| 1,425,060 (the whole board) | $0.00 | 0% | 0% |
Buy the whole board and you take the entire pool, losing the difference with certainty: stake $1,425,060, receive $1,131,304.46, lose $293,755.54. That is 20.6% and exactly zero variance. More tickets do not change the average: they only squeeze the distribution around a negative one.
A big jackpot does not mean a better bet
A bigger pool looks like a better bet. In MegaPot it is not, and the reason is the dynamic B. Doubling the pool doubles minNumberTickets, which doubles B, which doubles the board, which halves the jackpot probability. At the same time it doubles the premium fund, and with it the jackpot's share, those same 33% of the premium. The two effects cancel each other almost exactly.
Why the two effects cancel
P_jackpot(B) × payout_jackpot(B) ≈ 0.33 × 0.20 × pool / N_design, where N_design is the sales volume the pool is sized for.
This is the opposite of a traditional lottery with a rollover. There a missed jackpot carries into the next drawing and makes the next ticket better value. Here a bigger jackpot does not make the ticket better, because the board grows with the pool and eats the gain.
The only lever: where the referral fee goes
A player controls none of the probabilities, the weights or the pool size. One thing is in their hands: who gets the referral fee. The contract withholds 10% of the ticket price and hands it to whoever the player names as referrer. That can be someone else or their own wallet.
That changes only the size of the loss. Through someone else's link the margin is about 29%, with your own referrer 13% or 11%. The sign stays the same: a $1 ticket returns about $0.79, which is less than a dollar.
There is no strategy
There is no strategy in MegaPot. Neither the choice of numbers, nor the number of tickets, nor the size of the jackpot changes the EV. Quick pick and manual selection give the same distribution. More tickets do not raise the chance of profit, and the whole board guarantees a loss of $293,755.54.
A $234,066.85 jackpot against a pool of $1,128,895 looks like an argument. It is not one. It is about a fifth of the pool, and the chance of taking it on a single ticket is 1 in 1,425,060.
The math is not on the player's side, and winning takes a lot of luck.
P.S. Want to run your own numbers? My referral link gives you 2 bonus tickets to play for free: https://megapot.io/r/9JV79U
