Does Daletou Draw After Seeing Everyone’s Bets? How the Draw, Odds and Prizes Work

A player asked: “Does Daletou wait until everyone has bet before picking the numbers? It’s like rock-paper-scissors where I throw first and it throws second. How could I ever hit the jackpot?”

No. In the BluffKing lottery center, Daletou’s winning numbers come from a public random value that only appears after sales close. When you buy, that value does not exist yet, and the server cannot compute it either, so it cannot “see your ticket and then throw.” The jackpot is rare because it is genuinely hard to hit. Below are the draw process, the odds and this release’s prize adjustment.

How one Daletou period is drawn

Daletou (a free gold-bean game on the platform, not affiliated with China Sports Lottery) draws every minute, on the minute. One period runs like this:

1. The rules are fixed when sales open. As soon as a period starts selling, the server creates and stores a draw plan: game, period, draw time, sales cutoff and which future round of public randomness it will use. The plan yields a hash commitment, and your browser saves it in a pre-purchase receipt before sending the purchase.
2. Sales close five seconds before the draw. After that, the server rejects new purchases for the period.
3. The random value appears only after sales close. The plan names the first drand beacon round at or after the draw time. drand produces a round every three seconds, generated jointly by nodes run by many organizations; unless most of them collude, no single party, BluffKing included, can know or choose a round in advance.
4. A public algorithm turns it into numbers. The server fetches that round, verifies its BLS signature, hashes it together with the commitment from step 1 (SHA-256) into a seed, and uses a fixed shuffle to draw five front numbers from 1–35 and two back numbers from 1–12.
5. Settlement starts only after the numbers are stored. Only then does settlement read players’ tickets and pay each one.

No step of the derivation reads your numbers, the site-wide spread of bets, your balance, membership or past results. If the planned beacon round is temporarily unavailable, the period waits for exactly that round; it never switches rounds or falls back to server randomness.

So it is not rock-paper-scissors

Rock-paper-scissors fails when the second player sees your throw. Here the server is not the one throwing: the result comes from a shared random value agreed in advance and revealed only after sales close. It is closer to two people agreeing to “throw whatever the public scoreboard shows in three minutes,” on a scoreboard neither of them controls.

At the cutoff the server does know what you bought, but it neither knows the numbers nor can change them: they depend only on the plan fixed before sales closed and the beacon round that appears afterwards.

Why it feels impossible to win

There are 324,632 ways to pick 5 of 35 front numbers and 66 ways to pick 2 of 12 back numbers: 21,425,712 combinations in all. One stake hits the jackpot with a probability of about 1 in 21.4 million, the same as the basic real-world game. Buying one stake every minute, you would wait about 40 years on average for one jackpot. A long run of misses does not make the next period any likelier.

With the prize table from this release, payouts per 100-bean stake (stake included) and per-stake odds are:

Across all tiers, about one stake in 15 wins something. Payouts are fixed amounts and are not split when many players win the same period. By this table, every 100 beans staked returns about 41.7 beans on average over the long run. A lottery loses money over time by design, just like a real one.

This release’s prize adjustment: prizes really were too low

Part of that feeling was right. On 24 October 2026 every lottery-center stake moved from 10 beans to 100, but the Daletou, SSQ, TOTO, Happy8 and Fucai3D prize tables were not adjusted. Theoretical returns fell from about 38%–57% to 1%–4%: in Daletou, 100 beans staked returned only about 4 beans on average, and the lowest prize was 25 beans, less than the stake itself.

From this release, new periods scale prizes to the 100-bean stake: Daletou, SSQ and TOTO by 10×, Happy8 by 25× and Fucai3D by 50×, restoring the earlier return level. The draw algorithm and the odds are unchanged. Periods already drawn keep the prize table they were sold under, and their original proofs still verify.

How to check for yourself

1. Buy from the same device and browser when you can; it saves a pre-purchase receipt containing the period’s commitment.
2. After the draw, download the period’s proof from My tickets and import it into Lottery fairness verification under Tools, or recompute it with the offline verifier in the public source. It checks the scheduled round, the beacon signature, the number derivation, and your ticket’s tier and payout.
3. Compare the commitment in the proof with the one in your pre-purchase receipt to confirm the draw used the plan you saw before betting.

The limits matter too: this shows the numbers came from the pre-agreed public randomness and the payout follows the published rules, but on its own it does not prove the server’s database and wallet credits were never altered, or that the deployed program matches the public source; missing historical proofs are never fabricated. For more on draw proofs see How does BluffKing draw prizes?, and for full gameplay the Daletou guide.