A Loophole on a Gas Station Brochure
Jerry Selbee spent his career as a mathematician at Kellogg's, the cereal company, working statistical models. When he retired to Evart, Michigan — a small town of about 1,800 people — he figured that chapter of his life was closed.
Then one day in 2003, he stopped at a gas station and picked up a brochure for a new Michigan Lottery game called Winfall. He read the rules on the drive home. Within a few minutes of doing the math, he'd found something remarkable.
Winfall was a draw lottery with a rolling jackpot — standard enough. But it had an unusual rule: when the jackpot exceeded $5 million and nobody claimed it, the prize money didn't keep rolling upward. Instead, it "rolled down" and was redistributed into the lower prize tiers. On those weeks, the 3-, 4-, and 5-number prizes paid out dramatically more than usual.
Jerry ran the numbers. On a roll-down week, the expected value of a $1 Winfall ticket could exceed $1. The house edge, which normally works against players in every lottery game, temporarily flipped. Buying a large enough volume of tickets on the right drawing — not any single ticket, but the statistical aggregate of thousands — had positive expected value.
ℹ️ What "expected value" means here: If you buy a single $1 ticket, you might win nothing. But if you buy 100,000 tickets on a roll-down drawing, the mathematical distribution of 3-, 4-, and 5-number prizes across those tickets should, on average, return more than $100,000. It's not a guarantee per drawing — it's a statistical edge that compounds over many drawings. Jerry understood this distinction precisely because of his background in industrial statistics.
He showed the math to Marge. She checked it. Then they drove to a convenience store and spent $3,400 on Winfall tickets for the next roll-down drawing. They won $6,300.
They went back the next roll-down week. And the next. Within a few months they were buying tickets in bulk — eventually hundreds of thousands of dollars worth per drawing, driving to multiple stores across the region to stay under per-store purchase limits. They brought in family members and eventually formed a lottery club — GS Investment Strategies LLC — to pool funds and split winnings.
Entirely Legal — and Entirely Overlooked
The most striking part of this story is how long it went unnoticed. Jerry and Marge ran their operation for nearly nine years. They weren't hiding. They bought tickets through normal retail channels, paid for them with checks, and collected prizes through the standard claims process. The Michigan Lottery was aware of bulk buyers — they actually welcomed them, since bulk purchases increased revenue — but nobody at the lottery connected the dots about what was happening.
Part of why it worked so long is that the Selbees were methodical and quiet. They didn't brag. They kept meticulous records. They drove long distances to spread purchases across many stores. They ran it like a small business, because it was one.
Eventually, a Boston Globe investigation into a nearly identical scheme running in Massachusetts (where MIT students had independently discovered the same roll-down mechanics in a game called Cash WinFall) brought national attention to the concept. Michigan cancelled Winfall in 2005, but by then the Selbees had already shifted to playing the Massachusetts game instead, driving to New England on roll-down weeks to continue their strategy.
No fraud. No hacking. No inside information. Everything Jerry and Marge did was based entirely on publicly available game rules. The loophole wasn't a secret — it was printed in the lottery's own promotional materials. They just read more carefully than anyone else.
What They Did With the Money
This is the part the movie gets right, and it's also the part that makes the story genuinely worth telling.
Jerry and Marge Selbee were not interested in a bigger house or a boat. They were people who had spent their whole lives in Evart, and they watched their hometown slowly hollowing out — businesses closing, young people leaving, the kind of slow decline that hits small rural communities across North America.
So they used the money to fight it. They bought the local general store — Evart's Market — when it was about to close, renovated it, and kept it running as a community hub. They invested in local businesses. They donated to the volunteer fire department, the local food pantry, and community organizations. They helped finance a building renovation project that gave the town a proper gathering space.
They weren't trying to turn Evart into something it wasn't. They were just trying to make sure the things that made it a real community didn't disappear. Jerry has said in interviews that if the town had already been thriving, they probably would have just donated the money elsewhere. The need was specific and visible, and they were in a position to address it.
About the Movie
Jerry and Marge Go Large (2022) stars Bryan Cranston as Jerry and Annette Bening as Marge, and it's a genuinely warm watch — one of those rare films that manages to be feel-good without being saccharine. Cranston captures Jerry's dry, methodical personality well, and the film does a good job of conveying the specific satisfaction of finding a real mathematical edge and executing it cleanly.
Where the movie compresses or softens: the timeline is condensed, the scale of the operation is understated (the film makes it feel more modest than it was), and the community-revitalization storyline is given more screen time relative to the actual lottery mechanics than real life warranted. The MIT student subplot — which in reality was a parallel operation the Selbees weren't really aware of until late — is dramatized into more of a direct rivalry than it actually was.
None of that undermines it as a film. But if you watch it and think "that seems almost too good to be true," the real story is in some ways even better: a retired couple in their 60s, doing the math that a state lottery had accidentally published, quietly building a winning operation for nearly a decade, and then giving most of it to a town that needed it.
ℹ️ Where to watch: Jerry and Marge Go Large is available on Paramount+. It was released in June 2022 and runs about 1 hour 36 minutes. Rated PG-13.
The Lesson Worth Taking
Jerry Selbee wasn't a genius in the Hollywood sense. He didn't break any rules or exploit any secret. He just read the terms of a lottery game carefully, did some arithmetic, and recognized that a condition existed — temporarily, under specific circumstances — where the math favored the player.
That's the same logic that underlies what Scratchers Edge does for scratch tickets. We're not predicting winning tickets — that isn't possible with properly designed games. But the prize-remaining data that WCLC, OLG, BCLC, and Loto-Quebec publish is real, public information about the current state of each game. Most players never see it. A game where 90% of top prizes are already claimed is a different proposition than one where the full prize pool is intact — and you can check that before you buy.
Jerry would have checked. That's the whole point.
If you want to read more about the other people who found real edges in lottery systems — including the Toronto statistician who predicted OLG scratch tickets with 90% accuracy without scratching them — check out our piece on The People Who Actually Cracked the Lottery.
⚠️ Responsible gambling reminder: Scratch tickets are a form of gambling, and no amount of analysis changes the fundamental math — the house wins in aggregate. Information about prize availability helps you avoid games that are already picked clean, but it doesn't create positive expected value. Only spend what you can afford to lose. If gambling is affecting you or someone you care about, visit responsiblegambling.org or call 1-800-522-4700.
Sources
- The Selbee Lottery Story — HuffPost
- How a Michigan Couple Won $26 Million in the Lottery — CBS News
- Cash WinFall Lottery Investigation — The Boston Globe