Smart Score Called Pot of Gold (1695) Before the $500,000 Top Prize Claim

On May 23, 2026, a player claimed a $500,000 top prize on NY scratch-off game 1695 — the $5 ticket sold under the Pot of Gold family. The top-prize counter dropped from 3 to 2.

Heading into that claim, Pot of Gold was not the loudest game on the board. It was not the biggest jackpot, it was not the newest release, and it was not the most heavily advertised. In our overall list of every active NY scratch-off, it sat in the mid-teens.

But on the rankings that actually matter for a player with a $5 bill in their hand — the $5 dollar-tier rankings — Smart Score had Pot of Gold as the #1 pick three days in a row, and #1 or #2 every day for nearly a week before the claim.

The game: Pot of Gold (1695)

Game 1695 is a $5 NY scratch-off in the long-running Pot of Gold family. The headline tier pays $500,000 and the game originally printed three top-prize tickets. Heading into the May 23 claim, all three were still alive, sales were running at a healthy clip, and the game had pulled into the high-velocity portion of its lifecycle.

Nothing about that headline is unique. There are larger jackpots on the $20 and $30 boards, and there are higher-rated $1 games on tighter EV margins. What made 1695 stand out to the model was the combination of three structurally favorable signals at the same time:

• All 3 top prizes still in the field — the rarest configuration for a game this far into its run
• Claim intensity on mid-tier prizes accelerating week over week
• Realized payout rate trending above the headline rate — a sign the actual ticket pool is paying a little better than the brochure promises

The receipt: Smart Score's daily rank for 1695

Every row below was pulled from the production history/games_YYYY-MM-DD.json snapshot we write to S3 each day. No edits, no hindsight, no retraining after the fact. The "tier" column shows the rank within the $5 ticket tier — the actual list a $5-bill player would scroll.

The pattern is hard to miss: five of the six pre-claim snapshots had 1695 in the top 3 of the $5 tier, with three of them at #1. The May 22 dip to tier rank #3 reflects a tier-membership change (several $5 games were marked ended in the overnight refresh), not a deterioration in 1695 itself.

Why "dollar per ticket" is the lens that matters

This is the part that often gets missed in scratch-off coverage. The number that gets reported — the overall jackpot — isn't actually the right comparison for a player. The right comparison is this question:

"With the dollar amount I'm willing to spend per ticket, what is the best game I can buy at that price?"

A $20 ticket with a $1,000,000 top prize is not the same product as a $5 ticket with a $500,000 top prize. They have:

• Different print runs — $5 games typically print 5x to 8x as many tickets as $20 games
• Different active store counts and pack-deployment patterns
• Different claim velocity profiles — cheaper tickets sell and claim faster
• Different pull risk — how long the lottery leaves a game on shelves after a top prize is claimed

You cannot meaningfully sort these games against each other on one giant overall leaderboard. A cross-tier ranking will almost always over-reward the games with the biggest headline numbers, because the model has to penalize "only a half-million-dollar top" against "a million-dollar top still alive." The structurally good $5 ticket gets buried, even when it is the better buy for the player who is buying a $5 ticket.

This is exactly what almost happened to Pot of Gold. On the all-games leaderboard heading into May 22, 1695 was in the high teens to low thirties — a perfectly ordinary number. In the tier its buyers actually shop, it was #1.

What we shipped because of this claim

Looking back at the snapshots, the model did its job: Pot of Gold was correctly elevated within its tier before the claim. But the audit also surfaced a few things we wanted to tighten so the next case study is even cleaner. Today we deployed five surgical model improvements:

1. Heat-lag bonus. When a game's claim-velocity acceleration percentile crosses into the top 5 and at least two top prizes are still alive, the heat factor now gets a small immediate bump so it stops trailing the underlying signal by a day or two.
2. Burn-regime reconciliation. When seven-day claim intensity is in the top 10 percent but the regime-shift factor still reads cool, we reconcile the two so a hot game does not get falsely flagged as in a cool-down regime.
3. Top-prize-in-field floor. The cross-sectional jackpot-size penalty is now floored at 25 whenever the game still has at least two top prizes alive. Smaller-but-reachable jackpots no longer get over-penalized just because the cohort happens to include a $5,000,000 headline.
4. Top-prize hazard per dollar. A new post-score adjustment that rewards games with the highest ratio of (remaining top prizes × top-prize amount) per dollar of ticket price. This is exactly the structural advantage that made 1695 a strong $5-tier pick.
5. Public audit endpoint. We launched /recent-winners-vs-ranks, a public API that takes every top-prize claim in the last N days and looks up each game's Smart Score rank in the seven daily snapshots immediately before the claim. This is the only honest way to ask "where was the model when the prize was actually claimed," and we now expose it so anyone can run the audit themselves.

None of these changes were retro-fits to make 1695 look better. The Pot of Gold ranks above are exactly what we published live, before the claim, with the prior model. The new code starts taking effect on tomorrow's daily snapshot.

What this case study does NOT show

The same honesty section we ran in the $10K a Week for Life case study applies here:

What to take away as a player

1. Decide the dollar tier first, then rank within it. Pick the ticket price you are comfortable spending per ticket, then look at the top of that tier. The all-games leaderboard is for context, not for picking.
2. Persistence beats spikes. A game that holds the top of its tier for a week is a much stronger signal than a one-day jump, which is often just a competitor selling out.
3. Multiple top prizes alive is the cheat code. Games with all of their original top-prize tickets still in the field carry disproportionate hazard mass. That was the single biggest contributor to 1695's score going into May 17.
4. Re-check after a claim. When a top prize is claimed, the tier rankings re-weight immediately. The $5 tier is going to look different tomorrow morning than it did yesterday.
5. Play within your means. No model erases the house edge. Even the best-ranked $5 ticket has a negative expected return per single ticket. Treat this as entertainment with an information edge, not as income.

Related Articles

• Smart Score Called the $10K a Week for Life Top Prize (Case Study)
• Best $5 Scratch-Offs in NY
• NY Scratch-Offs Running Out of Top Prizes (Hazard Model)
• Scratch-Off Expected Value Explained
• Do More Expensive Scratch-Offs Have Better Odds?

Data sourced from the official New York Lottery prize reports at nylottery.ny.gov. All ranks and scores in this article are exactly as published on the day shown and pulled from the production daily snapshots in S3. For entertainment and informational purposes only. Please play responsibly.