Why Liquor Store Clerks Often Win Lotto
This is really cool--and fascinating, too. A few days ago, riffing on this Wired story about Mohan Srivastave, a geological statistician in Toronto who figured out how to beat scratch-off lotto tickets, I wrote about my history as a callous, collegiate liquor store clerk and my own much more rudimentary system for beating lotto. Well, apparently I am not the only liquor store clerk to employ this system, which, when I think about it, is not really shocking. Mohan Srivastave himself read my item and dropped me a note to let me know that liquor store clerks--or lottery retailers, a slightly broader category, I'd imagine--actually win a disproportionate share of scratch-off lotto tickets. He was kind enough to let me share his email and data:
You're right that the number of winners per packet is a design specification of most of the provincial and state lotteries. They don't want to let this be truly random because then there would be some packets that had very few winners, and this would leave some retailers feeling like they'd been given crappy packets. As humans, we tend to have poor instincts for what is truly random, and therefore are not good judges of when we're lucky and when we're unlucky. So camping out, waiting for someone else's losing streak, is a good strategy. And it's one that can only be employed by lottery retailers ...
... which goes a LONG way to explaining the documented anomaly that lottery retailers win an amazing percentage of the $$ in instant scratch games. Here in Ontario, analysis of recent data suggests that about 1 in 10 of the instant scratch prizes is claimed by lottery retailers. There are about 11,000 - 12,000 retail outlets in Ontario, so maybe 40,000 - 50,000 people who would count as employees of lottery retail outlets. And that's waaaay less than one tenth of the lottery-playing public in Ontario. So people have struggled to find an explanation. The lottery corporation has settled on the observation that employees in retail outlets probably tend to play more than the average shmo on the street. But your strategy sounds like a much more plausible explanation. Even without doing any kind of formal analysis of winning/losing patterns, a lot of retailers will have started to realize that long losing runs are often capped by a better-then-average winnings.