http://highline.huffingtonpost.com/a...lotto-winners/

Cliffs:

In certain situations certain US lotteries were +ev. An old man (and also others independently, including MIT students as usual) figured it out and won $7.75M after taxes and expenses over 9 years.

Extremely long read but I enjoyed it, thanks for posting. I had kinda heard about this as it ended but never really knew the details.

Would have been a more accurate thread title. But +1 on enjoyable read and thanks for posting.

Leave it to dumb degens to complain that a game is beatable. Probably some of those readers were unknowingly beating it already (in a too-small sample size to see results). But thanks to that misleading Globe article and the attention it got, there might never be another beatable US lottery.

That's a really good article. There are just 2 points the author tries to make that I'm not understanding.

1. The big bankroll players taking advantage of the little guys....... It seems like anyone buying tickets before the winfall is making a bad bet, and anyone buying during the winfall is making a good bet. Why does bankroll size matter? The little guys buying during the winfall have an edge that week too. And when the large bankrollers force the winfall earlier than expected, then those little players dumb enough to buy tickets in the non-winfall part got a benefit too that week

2. That the state is benefitting from all the extra tickets bought......It doesn't add up that the state is getting 30 cents, the retailer gets a cut for selling, and the player is getting more value than they paid for the ticket during the winfall. It just doesn't add up. Why isn't the state losing money during the winfall?

I think he meant that the little guy buys tickets day in and day out, not fully aware that most of the time it's -ev and only some of the time it's +ev, so overall they are subsidizing the big bankroll players who are smart enough to only buy when it's +ev.

I believe that the state IS actually losing money during the winfall, but that loss comes from accumulated profit from non-winfall days. Kinda like a badbeat jackpot. So over a full year the state still makes money.

Basically the state runs a jackpot that is, say, 40% return for 5 times, then a jackpot that is 110% return once. And during the +ev jackpot the total sales is only double a normal -ev jackpot. Rinse and repeat.

The state isn't losing money on winfall days. The jackpot just shrinks because all those 10x prizes suck most of the pool dry. In theory the state could lose money if there were enough prizes to completely drain the jackpot and still not have enough to pay out, but I don't think that's ever happened. (I have no sources to back this up)

18000 rpm is right, the big advantage players are obviously taking equity out of the prize pool which otherwise would be distributed to nonadvantage players. I actually think the article does a poor job of explaining why/how advantage players are taking equity from the fish but whatever.

Its more than that. If the edge comes from a big carryover such that buying one ticket for each combination costs less (but not a lot less) than the total being paid out to the players, a mega syndicate would have an edge but you wouldn't. The syndrome shows up more in Horse Race Pick Sixes than in lotteries but the theory is the same. Do you see why? (Stanford Wong didn't.)

I don't see why (but I still more than 50% believe you). If buying N>1 tickets is +EV then how is buying one ticket not also +EV? If anything, shouldn't your edge be greater buying one ticket since you're not picking duplicate combinations?

Prepare to get mad at yourself. Suppose that there are a million combinations and they cost a dollar each to bet on. Only you and I are aware of this contest. The prize is 1.5 million so it seems we both have an edge. You bet a ticket. I bet a million tickets. But my million bets are NOT RANDOM. They are chosen with the requirement that they don't duplicate. You are risking a dollar getting 749,999-1 on a 999,999-1 shot. I am risking a quarter of a million dollars, to win a half a million and will lose once in million.

Obviously real life scenarios are not this extreme but they don't have to be for syndicates to have an edge via guaranteed non duplication, compared to individual bettors (unless they saw a list of the other numbers bet).

Ok I thought I remembered the article saying these syndicates were picking duplicates. My mistake.