The Cash WinFall Roll-Down: How MIT Students and a Michigan Couple Legally Made Millions

Massachusetts Cash WinFall is the clearest documented case of a lottery that was, on certain days, a genuinely good bet. It was not hacked, rigged or defrauded. Its rules were published, and the rules were exploitable.

The rule that did it

Cash WinFall had a jackpot cap. If the jackpot reached $2 million and nobody matched all six numbers, the money did not roll over to the next draw. It rolled down — redistributed into the lower prize divisions, multiplying the payouts for matching five, four and three numbers.

That single design choice inverts the usual economics of a lottery:

• Lower divisions have far shorter odds, so a large bankroll can hit them reliably rather than hoping for a miracle.
• Lower-division prizes are small enough that few people share them.
• The redistributed money is large relative to those divisions' normal payouts.

On a roll-down draw, the expected value of a ticket rose above its price. Not by a rounding error — by enough to build a business on, provided you could buy tickets in enormous volume. The general mechanics are set out in roll-down mechanics.

Who worked it out

Several groups did, independently.

Random Strategies, founded by MIT student James Harvey, grew out of a class project examining the game's structure. Over roughly seven years the group wagered between $17 million and $18 million on Cash WinFall.

GS Investment Strategies, run by Jerry and Marjorie Selbee of Michigan, had already worked a similar roll-down structure in their home state before travelling to Massachusetts to play Cash WinFall.

A third group operated out of Michigan as well. All of them did the same unglamorous thing: on roll-down weeks, they bought hundreds of thousands of tickets, filling out betting slips by hand at a small number of hand-picked retail terminals.

What the Inspector General found

After Boston Globe reporting exposed the syndicates, Massachusetts Inspector General Gregory Sullivan investigated and published his report on 27 July 2012. The findings:

• High-volume players, including MIT undergraduates, spent about $40 million on tickets over seven years and won an estimated $48 million (WBUR).
• Random Strategies alone earned at least $3.5 million in profit between 2005 and 2012.
• The Lottery knew. Officials were aware of the syndicates' activity and were content with the increased sales, at times bending rules to let bettors buy hundreds of thousands of tickets at once.
• Nobody was harmed. The Inspector General found no evidence the game had harmed anyone — not small players, not taxpayers.
• The high-volume bettors did break lottery rules — by operating terminals themselves and outside regular hours — but that is a rules-of-play issue, not fraud.

That combination is what makes the story unusual. Ordinarily an exploit means someone lost. Here the state collected its cut on vastly increased sales, ordinary players faced unchanged odds on their own tickets, and the syndicates took the margin the rules had left lying on the table.

Why it was legal

Three things were true at once:

1. The rule was published. The roll-down was a documented feature of the game, printed in the rules the Lottery itself distributed.
2. The tickets were real. Every ticket was bought at face value through normal terminals.
3. Nothing about the draw was influenced. The syndicates could not affect which numbers came up. They simply bought enough combinations that the lower-division payouts, on roll-down weeks, exceeded their outlay.

This is the difference between exploiting a rule and compromising a system. For the latter, see the Eddie Tipton hack — the one case where a US lottery's random number generator genuinely was corrupted, which was a crime and produced a conviction.

Why it ended

Once the Globe published and the Inspector General reported, the game's economics were public and its political viability was not. Massachusetts wound Cash WinFall down; the last draws were in 2012.

Operators everywhere absorbed the lesson. Modern game designs avoid uncapped roll-downs into short-odds divisions, and where roll-downs survive they tend to be structured so that the redistributed money lands in tiers popular enough to be heavily shared. The bulk-buying countermeasures that followed are covered in why lotteries now cap or block bulk buying.

What it actually teaches

The romantic reading is smart people beat the lottery. The accurate reading is narrower and more useful:

• The exploit was in the rules, not the randomness. No amount of analysis of past draws would have found it. It came from reading the prize structure.
• It required enormous capital and enormous labour. Hundreds of thousands of hand-filled slips per roll-down week, for a margin in the single-digit percentages.
• It was available only on specific, announceable days. On every other Cash WinFall draw, the game was as negative as any other.
• It has been designed out. The specific structure that made it work is not on sale anywhere now.

If you want to check whether any live game is close to positive today, the jackpot tracker ranks the games we track by expected value per ticket, and positive expected value in a lottery explains why no jackpot game reaches it.