|
On August 03 2012 23:36 Golden Ghost wrote:Show nested quote +On August 03 2012 17:53 sirkyan wrote:On August 03 2012 17:29 zalz wrote:On August 03 2012 10:50 smokeyhoodoo wrote:On August 02 2012 20:12 Ahelvin wrote:On August 02 2012 20:04 smokeyhoodoo wrote:On August 02 2012 19:39 professorjoak wrote:On August 02 2012 19:16 smokeyhoodoo wrote:On August 02 2012 19:09 professorjoak wrote:On August 02 2012 19:03 smokeyhoodoo wrote: When this went public people STOPPED playing this lottery? I'm baffled, how can people be that dumb? According to the report, Massachusetts Lottery Commission limited the number of tickets any one person could buy so the big players left. You needed a very sizeable number of tickets (see an above post of mine for a discussion on estimating this) in order to stand a good chance of hitting at least the expected number of partial match prizes and getting money. The report says it was more like ~20% profit, but that doesn't count taxes. That's kind of besides the point. If you already play the lottery and are taking on risk it's stupid to feel cheated and stop playing when you discover it has positive expected value. No, this is exactly the point. Suppose that being risk-tolerant or risk-averse isn't just a binary condition, it's a function of both expected value and variance. If you assume that your utility function increases sub-linear with regards to money this is automatically fulfilled. This lottery was special because it didn't just have positive expected value, it had low variance so these big bettors winning money in a very large percentage of all possible realizations. Suppose we simulate the lottery results 10000000000000000 times and sort the results by the amount that you made/lost. The expected value is the sum of all those individual results divided by the number of simulations (the integral of the curve in the limit of an infinite number of simulations). There are many different profitability curves that give the same integral (area under the curve). Most lotteries involve a profitability curve that has you losing money in almost all scenarios but then a smaller than struck-by-lightning chance of winning the entire lottery in which case those bars on your chart are astronomically high. (This also assumes that you picked your numbers independently of anyone else and don't share the prize.) You would much rather play a lottery where the profitability curve is a flat line at the expected value, because then you don't lose money ever. The more of these fictional simulated lottery results that are positive, the more likely you want to play the lottery. You would be stupid to play Powerball's $100+ million jackpot just because it has positive expected value because outcome given the (very high) conditional probability that you don't win the jackpot is that you lose everything you put into it. Having said that, I imagine that for the short while when the Lottery Commission had the limited buy-in rules, the total number of tickets bought would be much lower which would probably increase the expected value and thus lower the threshold for how many tickets one would need to buy to be most probabably profitable. They eventually closed the game due to bad press. I understand what's going on but you're missing my point. Usual lottery players already accept that risk. That they are overwhelmingly likely to lose their money regardless of positive expected value. However, positive expected value is still a bonus. Learning about it shouldn't deter you from the purchases you're already making, it should simply delight you as a pleasant surprise. Feeling cheated or something because these students gamed the system is stupid as well as it doesn't hinder your chances of winning at all. Lottery players do not accept the risk for the vast majority. The appeal of the lottery is intrinsically irrational and based on emotion and trustworthiness of the organizer. For this reason, discovering that somebody was able to have an edge over them and managed to beat the lottery suddenly makes it "unfair", and people are not willing to play anymore. Argghhhhhhhhhhhhhhhhh the MIT students did not have an edge over regular players! They did. Their trick couldn't be duplicated because it would then not be profitable for anyone. The entire calculation is based on the idea of winning the lottery, alone. If there are multiple winners, the tickets are no longer positive value, and you lose money. Did you even read? Their entire calculation was not winning the lottery (I'm not sure if you're referring to the jackpot as "winning", but I'll give you the benefit of doubt and so no), their strategy was having as many 2 to ideally 6 rights on roll-down draws in order to recieve money that people betting during regular draws have been stacking up. If there are multiple winners they get a bit less, but it's not that big of a deal. They still make money, just not as much as can be compared with their own orchestrated roll-down vs 'regular' roll-downs. You are actually both right and wrong at the same time. I just read the whole article (not just the OP but the considerably longer article in the link the OP is just the first 1/5 of, very interesting read if you have the time). Although you are right in that everybody could reproduce the same trick everybody would end up losing money (as stated by the big investment clubs (of wich MIT was only 1 of 4) themselves). Even now with 4 investment clubs participating in the roll-downs their expected winning percentage plummeted from when it was only 1 club. At first the statistical amount won with 5 correct numbers was ~80k. In 2011 that was down to ~22k. The reason for this is that the money in the roll-down is distributed amongs the tickets sold for that specific draw. And at first this would be about 400k tickets with ~7 tickets with 5 correct numbers. Later on this went to 1.2m tickets and alot more tickets with 5 correct numbers. Although overall this would mean that more money could be distributed it also means it had to be distributed between a lot more parties. All this would have lead to a profit percentage that was to low to offset the risk and the investors themselves stated that with only one or two extra big parties (or a lot of normal people just purchasing 1 or 2 tickets) they would eventually have to stop investing at all. Also the normal people who also played in the regular non roll-down weeks felt cheated because the investors waited until the roll-down and swooped in to cash in on their effort to even get the jackpot there. And rightfully so imo. And thus I guess they stopped buying tickets in the regular weeks, with the result the investors had nothing to invest in anymore as there were no (or significantly less) roll-down weeks, and the whole game had to cancel in the end. TLDR: Yes the investors did have an edge because the strategy they followed is only viable when purchasing a big amount of tickets. Yes it´s also true everybody could duplicate this strategy relatively easy but this would also have resulted in the strategy becoming not viable at all anymore.
I'm not sure if you responded to the wrong post. I said nothing about replicating their strategy.
However, I'll respond with my thoughts on your post anyway, if I may .
First of all, MIT was one of three in the farily early stage, where the other groups was the Bio-med researcher one and the one with Selbee. The Bio-med later split up and Dr.Zhang scaled his ticket purchasing waaay down, while his group kept going without him.
This is the information I got when reading, where did you get they were 4? Only three were interviewed, no?
Secondly, not everyone could replicate this strategy. Or, well, I guess anyone could _try_ but at that point it ceases to be an investment opportunity and thus groups will fall off and come back when it cools. This sporadic behaviour would disallow 'anyone' to replicate it in practice. Another problem with the strategy was the extreme hassle it caused. Dr.Zhang quit his research job because it took such an extreme amount of time to not only register the slips but also to refine the formula and predict roll-downs and keep up-to-date. Harvey was attending his last semester of MIT, he had little time for course-work because Cash WinFall took time.
You would also have to find a licensensed retailer to register the MANY MANY slips you had filled out by hand. Not to mention the potential loss.
I do realize 'normal' people would get less money when they actually won, but because of the betting clubs there were 4 to 5 more roll-downs per year, weren't there? I think this pretty much makes it a zero-sum if you play on every roll-down as a 'normal' person. Not to mention the potential jackpot who got boosted. Eventhough the jackpot was only won once during roll-downs occurring in the time-span discussed in the report, that jackpot winner (be it a club or an individual better) recieved a considerable boost in winnings, had the clubs not stepped in.
Thirdly. I don't feel the MIT-orchestrated roll-down should be a point of concern. It took them a bit over a year to organize it and it got fixed by the next drawing, it was a one time happening.
I do agree however with that people buying tickets during regular drawings might feel cheated.
Slight edit, my bolded is repeating what you've said. Just so we're on the same page, basically.
|
Can someone clarify this a little more? I am confused.
|
My friend and I tried calculating some odds for fun after Calculus class before for our local 4D thing but nothing interesting came up =(
But yeah holy shit reading that felt just like watching Oceans Eleven haha
|
Very interesting. But it seems like it wasn't profitable for them if they did all the work themselves. ~4 people working on it full time + a dozen that was part of the project. $3.5 million over 7 years is $500000 per year. Seems like they could have made that kind of money doing normal work as MIT graduates.
|
On August 04 2012 03:25 NeMeSiS3 wrote: Can someone clarify this a little more? I am confused. The OP has a pretty nice summary already, but I can give another overview.
Most lotteries work by adding part of your ticket to the jackpot, which keeps increasing until someone wins. So you end up with jackpots in the hundreds of millions that stay around for weeks or longer.
This lottery is special in that the jackpot never goes above $2 million. Whenever it does go over, they basically have a "clearance sale" where if no one wins, they just divide the jackpot among everyone who came close to winning or won any prize at all. That's what the rolldown is. You don't have to be a mathematician to realize that you will get way more money by playing during a rolldown.
What the MIT students did is figure out that the payout is so good during the rolldown that you will actually make money on average. So they got a bunch of guys together and pooled their money and bought thousands of tickets during the rolldown. Pretty simple. At some point buying too many tickets is overkill and you'd be better off putting the money into the next rolldown, so that's how they calculated a specific number of tickets to buy.
The real question is why anyone would play the lottery when it's not a rolldown. But people do, because the jackpot has to get to $2 million somehow. Now that everyone knows about this, there aren't going to be as many people playing when it's not a rolldown and that's probably the main reason they pulled the game.
|
On August 04 2012 04:33 starfries wrote:Show nested quote +On August 04 2012 03:25 NeMeSiS3 wrote: Can someone clarify this a little more? I am confused. The OP has a pretty nice summary already, but I can give another overview. Most lotteries work by adding part of your ticket to the jackpot, which keeps increasing until someone wins. So you end up with jackpots in the hundreds of millions that stay around for weeks or longer. This lottery is special in that the jackpot never goes above $2 million. Whenever it does go over, they basically have a "clearance sale" where if no one wins, they just divide the jackpot among everyone who came close to winning or won any prize at all. That's what the rolldown is. You don't have to be a mathematician to realize that you will get way more money by playing during a rolldown. What the MIT students did is figure out that the payout is so good during the rolldown that you will actually make money on average. So they got a bunch of guys together and pooled their money and bought thousands of tickets during the rolldown. Pretty simple. At some point buying too many tickets is overkill and you'd be better off putting the money into the next rolldown, so that's how they calculated a specific number of tickets to buy. The real question is why anyone would play the lottery when it's not a rolldown. But people do, because the jackpot has to get to $2 million somehow. Now that everyone knows about this, there aren't going to be as many people playing when it's not a rolldown and that's probably the main reason they pulled the game.
Ahh, this is legal? Because if it is, I think the most interesting part is the dilution of the product, it will become more interesting if the trend continues as more people try everyone starts losing money so they stop, then people (such as MIT) come in and repeat the previous model that worked in the first place until it dilutes again, rinse repeat etc...
Very clever though.
|
On August 04 2012 03:25 NeMeSiS3 wrote: Can someone clarify this a little more? I am confused.
tl;dr, in this specific game in the weeks where the prize money was bigger, it was profitable to play, so by buying a lot of tickets they would win often enough to not need years to see a return on their investment
|
inspiring story for mathematicians =P
|
|
|
|
|