"Things are meant to be for a reason"
June 7, 2013 5:45 AM   Subscribe

'It could have been us, but things happen. Sometimes it's better to be patient than right." Mindy Cradell has little regret about letting an older woman in front of her, while standing in line for the lottery. That woman, Gloria Mackenzie, became the sole winner of the $590 million Powerball jackpot, the largest in the history of American lotteries.
posted by Brandon Blatcher (64 comments total) 4 users marked this as a favorite
 
Ironically - the lottery winner I know was the one who WON because she let an impatient guy into the line in front of her, so she was the one who got the winning ticket. Sometimes it really does pay to be patient.
posted by EmpressCallipygos at 5:53 AM on June 7, 2013 [2 favorites]


so a nationwide computer system that spits out 3000 random tickets per minute would have given this woman the same exact numbers as the lady in front of her? When was the last time an American news organization actually found some news instead of just manufacturing it? Also, wtf is this shit about people not picking their own numbers, how long has that been going on? I'm a big believer that you can gauge the health of any society by how many people play the lottery.
posted by any major dude at 5:56 AM on June 7, 2013 [24 favorites]


I'm not clear how this matters except in a brush-with-fame sort of way; Tickets win or lose based on matching numbers that you manually pick or the numbers the machine randomly generates for you. The timing of the ticket purchase is irrelevant.

It's like saying you could have been a movie star instead of Brad Pitt if you had cut in front of him in line at a movie audition.
posted by ardgedee at 5:57 AM on June 7, 2013 [11 favorites]


Interesting, I had no idea that's how the Powerball lottery was played, thought it spit out numbers for people.
posted by Brandon Blatcher at 6:01 AM on June 7, 2013


Yup. Powerball machines do not dole out a predefined winning numbers for one random ticket. They generate random numbers for each ticket.

More, with most computers, random numbers are not actually random. They are generated from a number of factors, called seeds, including time - to the second, not the minute. If she hadn't let the little old lady go ahead, she still would not have received the winning number unless the machine had printed the ticket at the exact same instant, which is laughably unlikely.

So, her generosity allowed something magical for someone else, and it did not rob her in the least. Win/Win.
posted by Slap*Happy at 6:01 AM on June 7, 2013 [31 favorites]


I understand why the lottery commissions would want to publicize the names of winners - get people excited and head off accusations of possible blatant corruption - but sometimes it seems like... I don't know, akin to painting a target on their forehead. I would think that this woman (the winner, not the line-cut-letter-inner) is going to be living out the rest of her days constantly looking over her shoulder in fear, and that's probably a best-case scenario.

I was particularly struck by an article that I had read earlier, which, in addition to being plastered with her name and where she lived and such, noted that she "begged reporters for privacy".
posted by Flunkie at 6:02 AM on June 7, 2013 [4 favorites]


allowed something magical

Probably worth checking back with the winner in a few years to see if winning all that cash really was a good thing. Often lottery wins ruin people's lives.
posted by aught at 6:03 AM on June 7, 2013


Probably worth checking back with the winner in a few years to see if winning all that cash really was a good thing. Often lottery wins ruin people's lives.

I think this particular winner isn't going to have that problem, one way or another.
posted by Etrigan at 6:05 AM on June 7, 2013 [4 favorites]


I prefer to think that she dodged the bullet on living out her short final years as the coke-addled milch cow for everyone she has ever met or anyone even vaguely related to her, adrift in a world where no-one's motivation can be trusted. I guess I'm just an optimist.
posted by biffa at 6:06 AM on June 7, 2013


Ah yes, "things happen for a reason", that happy mental refuge. It's a tremendous comfort until you actually try and discern the reason, which is "because the universe is vast, unthinking, and essentially random."
posted by seanmpuckett at 6:09 AM on June 7, 2013 [9 favorites]


I've heard that the randomly created numbers are also not random in that they attempt to pick combinations that haven't been sold for that drawing yet; in other words, a set of numbers that noone else has chosen.
posted by ceribus peribus at 6:09 AM on June 7, 2013


Often lottery wins ruin people's lives.

I just want to let the vast, indifferent universe know that the lottery can ruin my life anytime.
posted by Elementary Penguin at 6:11 AM on June 7, 2013 [18 favorites]


I really hope lottery PRNGs are seeded a bit better than gettimeofday() but stupider things have happened so who knows.
posted by Skorgu at 6:13 AM on June 7, 2013 [1 favorite]


I've heard that the randomly created numbers are also not random in that they attempt to pick combinations that haven't been sold for that drawing yet; in other words, a set of numbers that noone else has chosen.
That seems utterly unlikely to be true.
posted by Flunkie at 6:13 AM on June 7, 2013 [3 favorites]


I think this particular winner isn't going to have that problem, one way or another.

Check in with that son in ten years then.
posted by aught at 6:13 AM on June 7, 2013 [1 favorite]


I am guilty of watching some of TLC's TV show "Lottery Changed My Life", and boy-oh-boy, those poor people.

As far as I can tell, every person on that show was worse off afterwards. Why?
Suddenly overnight being a millionaire is a heavy burden.
Money doesn't buy you class. If you were a drinking-fighting-gambling-lying-trouble maker before you won, you probably will be after you won, the only difference is there are millions more bullets in the gun. Dangerous!
posted by QueerAngel28 at 6:15 AM on June 7, 2013 [1 favorite]


That seems utterly unlikely to be true.

And yet it is! It's what I heard.
posted by ceribus peribus at 6:19 AM on June 7, 2013


As far as I can tell, every person on that show was worse off afterwards. Why?

Because the people that are doing it "right" know better / have better things to do than go on some TV show.

(Though the one episode of that show that I've seen had an older couple who purchased a nice RV to tool around in and seemed to be doing fine)
posted by ghharr at 6:20 AM on June 7, 2013 [3 favorites]


As far as I can tell, every person on that show was worse off afterwards. Why?

I haven't watched this show, but would suspect there's a lot of selection bias going on:

1. The overwhelming majority of the viewing audience are non-lottery-winners and it's more satisfying to see people doing poorly after winning the lottery than thriving. Sour grapes and all.

2. "Look! This person won a billion dollars and is miserable" is a more dramatic story than "this person won a billion dollars and is happier than ever."

3. Prodigal winners are more likely to accept an offer to go on a show because, well, they need the money.
posted by payoto at 6:21 AM on June 7, 2013 [4 favorites]


As far as I can tell, every person on that show was worse off afterwards. Why?

One argument, made above, is that if a jackass wins they're still a jackass. A related point is that there are certain advantages in convincing Proles that their superiors are actually not as happy and fortunate as they may seem. No, you must pity the wealthy, and be happy with your poverty!

I propose a simple experiment: everyone here, give me all of your worldly assets. Cash, investments, belongings, real estate. All of it. I will also receive 100% of any income you possess, but I will permit you to retain 10 sets of clothing and two pairs of shoes per year; after the first year I will supply you with those articles of clothing at no cost.

After five years we'll see who is happier.

It's a worthy test, and we must do it For Science. Not me, mind you. No, not me. I'm not being selfish, I'm really just here for the Science. When you think about it, I must be some kind of hero to take this sort of risk. A true hero.
posted by aramaic at 6:25 AM on June 7, 2013 [2 favorites]


I would guess that people who have trouble after winning the lottery are in the same boat as athletes who are impoverished a few years after retirement: If you have always lived paycheck to paycheck, that is a hard habit to break just because your paychecks are bigger. There's a reason that young athletes get financial advisers now. If everyone who won the lottery used some of the winnings to hire someone to manage their money (like Mrs. MacKenzie seems to have done) you wouldn't hear of so many ruined lives.
posted by Elementary Penguin at 6:33 AM on June 7, 2013 [2 favorites]


I was immediately reminded of Jack Whittaker (previously). The link for the article in the FPP is dead now, but you can still get to it here. The article's quite well-written and quite sad: The day would come when many West Virginians recalled the story of Jack's Powerball Christmas with a shudder at the magnitude of ruination: families asunder, precious lambs six feet under, folks undone by the lure of all that easy money...

tl;dr: Time, BusinessWeek, and the New York Post all have shorter versions of the Whittaker story.
posted by orthicon halo at 6:33 AM on June 7, 2013


Really, what would be the motivation for going on a reality tv show after you won the lottery? Probably the same as any other reality show and tt generally wouldn't be because your life is going well. You win the lottery and are happy? Stay the fuck away from TLC or anyone else pitching you horrible ideas.
posted by edgeways at 6:36 AM on June 7, 2013 [1 favorite]


The lottery corporation in my neck of the woods sent out a press release when a woman voided a ticket which would've won $30 million had she kept it. It was very irresponsible, because it became clear who that woman was, and it could've easily led to a suicide (which luckily it did not). They did it to show "you can't win if you don't play" which is only a half-truth IMHO.
posted by Brodiggitty at 6:50 AM on June 7, 2013 [1 favorite]


so a nationwide computer system that spits out 3000 random tickets per minute would have given this woman the same exact numbers as the lady in front of her?

Yeah, and Crandall might have been planning to select her own numbers, in which case the whole "did that elderly woman steal my lottery winnings?" question would be moot.

The lottery corporation in my neck of the woods sent out a press release when a woman voided a ticket which would've won $30 million had she kept it.

Damn, that's a disgusting move on the part of the lottery company and the newspaper reporting that story; the woman apparently asked for a $10 ticket and the clerk accidentally printed a $27 ticket, so she returned it. The lottery waving the episode in a press release to drum up business and the paper not only biting but leading with "N.L. store patron blows chance to win $30-million jackpot by returning ticket" are both shitty, desperate things to do.
posted by mediareport at 7:03 AM on June 7, 2013 [4 favorites]


I've seen news stories about how lottery winners end up being targets for lawsuits. Basically, if it is known that you have a big reserve, it will be a target for those who want a piece of it.

If you won a lottery, I would think that step 1 would be to get a decent umbrella policy and step 2 would be to retain a lawyer, finally step 3 hire a financial adviser.
posted by plinth at 7:15 AM on June 7, 2013


Step 4: Hire someone to follow you around and say "No, that's stupid," when you are about to do something stupid. This would be particularly good for professional athletes, who start their "rookie" seasons after years of being the most important person in the room -- having someone around who is specifically not dependent on telling you how awesome you are would easily pay for itself.
posted by Etrigan at 7:18 AM on June 7, 2013 [1 favorite]


Also, wtf is this shit about people not picking their own numbers, how long has that been going on?

Almost since the start of these large lotteries. Probability-wise, a random pick is better than a user-picked choice. Why? Because most people pick birthdays and other favorite numbers, which tend to range below 31 while the number space of these lotteries tend to go into the 50s and higher.

Lottery officials LOVE it when numbers higher than 31 come up, that cuts down the chance of someone winning and raises the chance of a rollover.
posted by JoeZydeco at 7:18 AM on June 7, 2013 [2 favorites]


The family I knew who won decent-sized lottery jackpot dis something similar to plinth's suggestion, and didn't announce that they had won until after they changed and unlisted their phone number. They actually went on a weekend snow trip with a boy scout troop, and the father of that family helped my dad put on snow chains, crawling under our car at 9pm on the side of a snowy road. After that trip, they claimed their prize, much to the surprise of everyone who went on that trip.

They bought a bigger house with a money-losing orchard (they didn't bother harvesting it) for a tax write-off, bought the building where they had their little business, gave all their employees raises and bonuses, and went on working for a while, at least. I don't imagine you'd see them on a TV reality show any time soon.
posted by filthy light thief at 7:20 AM on June 7, 2013 [7 favorites]


I think it was pretty irresponsible to publicize the winner's home address; her life may be endangered.
posted by Renoroc at 7:20 AM on June 7, 2013 [1 favorite]


It's like saying you could have been a movie star instead of Brad Pitt if you had cut in front of him in line at a movie audition.

I went to grade school with Kiefer Sutherland. My mom was watching TV when it was announced that he was set to marry Julia Roberts and she turned to me and said, "Why the hell aren't you marrying Julia Roberts?! If Kiefer can do it..."

People's brains work in weird ways. Of course later when I told her that they hadn't married, she said that was besides the point. "She wouldn't have left no Dobbs standing at the altar. You crazy? No way!"
posted by dobbs at 7:28 AM on June 7, 2013 [1 favorite]


My advice to the lucky lottery winner is the same advice I offer all lucky lottery winners: Zeppelins.
posted by notyou at 7:31 AM on June 7, 2013 [3 favorites]


Probability-wise, a random pick is better than a user-picked choice. Why? Because most people pick birthdays and other favorite numbers, which tend to range below 31 while the number space of these lotteries tend to go into the 50s and higher.

Well then, it seems to me a canny player would pick numbers above 31 to reduce the likelihood of others also picking the same numbers.

I used to know an impressively lazy guy who kept himself in groceries playing bingo. In these parts, at least, bingo games are regulated: they have to keep to a fixed schedule and have set prizes. He paid attention to traffic patterns for each day of the week, and factored in what kind of dent a rainy day or extreme temperatures made in attendance, and always went to games where the attendance was likeliest to be low. He reasoned that they still had to hand the same $1500 (or whatever) out whether there were four hundred people in the room or thirty.
posted by ricochet biscuit at 7:40 AM on June 7, 2013 [6 favorites]


I've bought lottery tickets a few times. I'm too cheap to buy them regularly, and I don't buy them under the belief I'm going to win. What I do is buy a ticket on a Sunday morning, or on a Wednesday afternoon. I'll pick some silly numbers, in fact, I think I chose 1,2,3,4,5 and 41 once.

For two days, I fantasize about what I would do if I actually were to win. Hey, I could get a little place with solar panels and a windmill. Or, I could move to Denmark and open a kite store. Maybe, my wife and I could buy a nice boat, and travel around North America (no more than 3 miles from the shoreline, of course).

Around Valentine's day, my wife found a half dozen of them in the cigar box. I guess I won $10 on one. We took the kids out for ice cream, and my daughter sneezed while she was eating it and looked at us with a confused face covered in white cream and sprinkles. That alone was worth the $1.

Summary - I'm turning into one of those old-man types.
posted by Bathtub Bobsled at 7:40 AM on June 7, 2013 [6 favorites]


winners end up being targets for lawsuits

I've heard that too, and I believe it. It's why my strategy (why do I have a strategy, when I don't buy tickets?) would be to establish a trust and have the trust claim the prize, assuming I bought the ticket from one of the states that permit it. Next, begin insulating all of my assets and income streams through shell corporations and offshore holding companies.

After that, it's jurisdictional arbitrage ahoy!

I won't buy anything. Instead, I will be the beneficiary of the fact that MiniTrue Corporation just happened to buy a nice place downtown, and the fact that IngSoc7 Co. has a contract with a limo service for whomever happens to be staying at the MiniTrue Corp. condo. Did I buy that meal? No, that was BigBro9 Inc., all taxes paid.
posted by aramaic at 7:43 AM on June 7, 2013 [2 favorites]


Well then, it seems to me a canny player would pick numbers above 31 to reduce the likelihood of others also picking the same numbers.

A canny lottery player is not playing the lottery.
posted by Etrigan at 7:44 AM on June 7, 2013 [3 favorites]


They are generated from a number of factors, called seeds, including time

I wonder how these machines actually do generate their random seeds. Ideally, one would want each machine to get a truly random seed each time it boots. Using the time for a seed is common practice in programming, but would be a disaster for a distributed lottery: people buying tickets across the state would get the same numbers.
posted by qxntpqbbbqxl at 7:44 AM on June 7, 2013 [1 favorite]


I'm not sure the ticket machine itself needs to be truly random, given that the drawing appears to be truly random. I'd think that the ticket numbers just need to be random "enough".
posted by aramaic at 7:48 AM on June 7, 2013


Yeah, I'm not sure they really need to be seeding for each ticket. But if they did want to do that, a trivial first step would be something like "concatenate the current time to the millisecond with the ID of the terminal that the ticket was purchased from, perform some hash on the result, and then use the result of that as the seed".
posted by Flunkie at 7:56 AM on June 7, 2013


"Sometimes it's better to be patient than rich," correct?
posted by aturoff at 8:01 AM on June 7, 2013




There was an interesting article making the rounds back when the Powerball jackpot hit umpty gazillion dollars, pointing out how fundamentally having millions of dollars can change your relationship to your friends, family and community even if you try to do everything responsibly (set up a trust to claim the winnings, invest it all and live comfortably off of the interest) because now when your friend who's living paycheck to paycheck needs a new water heater, resentment can start poisoning that relationship... because what would a few hundred dollars be to someone who's just sitting on millions? It puts everyone in a lousy place; if you help everyone out with a few hundred bucks here and there you'll just wind up dribbling your money away. If you decide to alot a sustainable amount of money per year for friend/family emergencies, then you wind up having to make the call of who gets how much. If you try to avoid the issue altogether you may end up alienating everyone anyway.

Everyone likes to think that they would find a way to make it work, but I imagine you'd always be running into unforeseen consequences and complications. In this case the number of predatory, opportunistic scumbags coming out of the woodwork is probably even higher than usual given her age.
posted by usonian at 8:18 AM on June 7, 2013 [2 favorites]


The only answer if you win millions is to completely disappear; move to another state, take another name, become another person completely. If you want to help people you knew in that old life, hire lawyers to deliver cheques anonymously.

I would not be even slightly surprised if there were a half dozen law firms who have made it their primary business to help "lucky" people disappear in this way.
posted by seanmpuckett at 8:27 AM on June 7, 2013 [1 favorite]


People who believe in this kind of thing are the ones who complain about blackjack players "taking their card" if they make a bad decision. It's great that they are being magnanimous about "losing" their shot at millions, but I feel bad that they are walking around thinking they actually would have gotten that winning ticket had they done something differently.

One of my lottery winning daydreams is to set up a profit sharing fund kind of thing where everyone I know gets a piece of a trust that invests the money and gives them an income stream.

One of my other daydreams is to come up with the most convoluted ways to hide the money from opportunists. "What do you mean I won a million dollars? I had to pay a third of it in taxes, and then I paid off my family's mortgages. There's nothing left!" (Which in that case, would almost literally be true. It doesn't take too many mortgages to add up to $750k...)
posted by gjc at 8:30 AM on June 7, 2013


Long time ago there was a multimillion dollar lottery winner in Illinois who started by founding a trust as an LLC with no family name attached, and set things up in such a way as to have the wealth held in the trust, and block anyone suing for the money. (The winner was a pharmacist.) They did show up at the press conference, but wearing a goofy disguise.

One of the problems is that privacy seems to be only protected by state law, and most states haven't passed the laws that favor privacy over transparency. For example, my state allows corps to collect on behalf of winners, but there are no laws against finding out from the lottery agency who owns the corporation.

I would disappear as thoroughly as possible, but the temptation to use that money, for good or bad, would overwhelm me sooner or later, and I'd resurface.
posted by ZeusHumms at 8:45 AM on June 7, 2013


For example, my state allows corps to collect on behalf of winners, but there are no laws against finding out from the lottery agency who owns the corporation.

With the right set of shells, I suspect that you could be fairly impervious to this sort of thing. Yeah, they'd find out that IngSoc6 claimed the prize, only to discover that IngSoc6 was a subsidiary of MiniTrue9, which was itself incorporated in the Isle of Man and which is in turn owned by CompuMatoTran, a Panamanian division of Hairy Mole Fireworks of the Bahamas.

If you're not trying to evade taxes, the govt won't even mind. Like everything else in the world, it's not about being impossible to find, it's just about being harder to find than other targets. As the duly-employed agent to Blue Heron Crane Technologies, why of course you get to ride in a limo to lovely restaurants at your leisure! And that nice condo you have on the Riviera? Well, that's just a perk of being the trusted agent for Maximal Conveyor Systems Inc.

(Seriously. Look into One Hyde Park, in London. Check out "who" owns the flats and remember that you'll be far less interesting than any of those entities.)
posted by aramaic at 9:01 AM on June 7, 2013 [2 favorites]


Ideally, one would want each machine to get a truly random seed each time it boots. Using the time for a seed is common practice in programming, but would be a disaster for a distributed lottery: people buying tickets across the state would get the same numbers.

Time can also be measured in a number of ways by the PRNG. Instead of canonical time (the day and time on the clock), it could be using the length of time the process has been running, or since the machine's been booted, or since the last ticket was issued - or all of the above. If it's really fancy, it will have a PNRG chip that uses time differentials caused by by detecting various physical phenomena at play... you will not get the same random number at another moment in time.

In any event, computer generated random numbers don't queue up waiting to be taken in turn.
posted by Slap*Happy at 9:43 AM on June 7, 2013


With the right set of shells, I suspect that you could be fairly impervious to this sort of thing. Yeah, they'd find out that IngSoc6 claimed the prize, only to discover that IngSoc6 was a subsidiary of MiniTrue9, which was itself incorporated in the Isle of Man and which is in turn owned by CompuMatoTran, a Panamanian division of Hairy Mole Fireworks of the Bahamas.

It depends on where you are.
In quite a few states, lottery winners are not allowed to be "anonymous".

No lawyers on behalf of a client, no representative of a trust, etc.
The person who bought the ticket(or signed the back) has to show up, in person, to claim it.
It's mostly presented as a way to reduce insider shenanigans.

So, if you're planning on doing this, buy your ticket in a state with winner privacy.

(A lot of states allow you to skip the big check press conference though.)
posted by madajb at 9:44 AM on June 7, 2013


I think it's more common for PRNGs to emit a conceptually infinite sequence of random numbers independent of the time those numbers are requested. This property is desired when PRNGs are used in simulations, so that a simulation can be repeated again with the same seed. In this transcript, I took no special care to make sure that the same intervals passed between entering each call to random.random():

>>> import random
>>> random.seed(1)
>>> random.random()
0.13436424411240122
>>> random.random()
0.84743373693723267
>>> random.seed(1)
>>> random.random()
0.13436424411240122
>>> random.random()
0.84743373693723267

There certainly do exist time-based PRNGs (though I've got to reach way back for one that I know of: the Commodore 64 BASIC had RND(positive) for a seed-based PRNG and RND(0) for a time-based PRNG; RND(0) or RND(negative) would seed the RNG)
posted by jepler at 9:58 AM on June 7, 2013


This property is desired when PRNGs are used in simulations, so that a simulation can be repeated again with the same seed.

That's not a desirable property when it comes to a lottery, where you don't want consistent repeatability. Note again the entry for the hardware random number generator... lottery machines are a common application.
posted by Slap*Happy at 10:26 AM on June 7, 2013


I just want to let the vast, indifferent universe know that the lottery can ruin my life anytime.
posted by Elementary Penguin


Just so you know: when a mathematician wins the lottery, they revoke your PhD.
posted by benito.strauss at 10:58 AM on June 7, 2013 [1 favorite]


But they'll gladly name a campus after you anyway.

ahem, ahem.
posted by seanmpuckett at 11:03 AM on June 7, 2013 [1 favorite]


It's important for the numbers to be random because people track winning numbers and notice if it's not random. One of the lotteries was hacked when the numbers were chosen by picking ping pong balls.
posted by theora55 at 1:52 PM on June 7, 2013


I'm a big believer that you can gauge the health of any society by how many people play the lottery.

What does this even mean? Is it some sort of "Lol lottery is a tax on the stupid" type of point? Is a measure of the hopelessness of the average persons situation?

It's one of those things that sounds really zingy in some Stephen Colbert type of way without really meaning much.

So yea, what's your point with this quip?
posted by emptythought at 3:02 PM on June 7, 2013


I dunno. I'm pretty resistant to magical thinking, but if I let Brad Pitt cut in front of me in line then he gets hit by lightning three times in succession out of a clear blue sky, I'd be a little freaked out.
When something that unlikely happens in close proximity, its SOOO incomprehensible - but you just saw the damn thing happen, you figure (wrongly) there had to be some kind of meaning to it.

It's an affront to reason for the universe to show its naked ass like that to a conscious being.
posted by Smedleyman at 4:11 PM on June 7, 2013


benito.strauss: "Just so you know: when a mathematician wins the lottery, they revoke your PhD."

I love my job, but for $300 million after tax they can have it.
posted by Elementary Penguin at 4:13 PM on June 7, 2013 [2 favorites]


I'm not sure the ticket machine itself needs to be truly random, given that the drawing appears to be truly random. I'd think that the ticket numbers just need to be random "enough".

I disagree; here's why. Suppose your machine is seeded based on time + location + some other local phenomenon with a few bits of entropy, using this seed to generate all of the random numbers. It's still not random enough. I can enumerate all of the pseudorandom lottery numbers that this station could possibly produce for a given lottery, given a few reasonable assumptions about things like the maximum rate at which people buy tickets and the resolution of the clock. This represents a tiny swath of the actual n-dimensional space that the winning numbers are sampled from. So when I go and buy a ticket from that machine, I am not actually getting a uniformly random number, and worse yet, there are many potential winning numbers which I actually have a zero percent chance of matching.

Maybe they bootstrap the seed with something like "wave the mouse cursor around for a while to generate entropy" that key generation schemes use. Dunno.
posted by qxntpqbbbqxl at 4:40 PM on June 7, 2013


A few years ago a woman with a baby stroller was walking down the sidewalk near an intersection a few steps from my house.

At the same time, an elderly man who was late for a doctor appointment and not paying attention to his driving ran a stoplight and collided with another vehicle in that intersection, sending it spinning across the road, over the curb, through the verge and onto the sidewalk.

It didn't hit the woman, or the baby stroller.

But, if she'd been walking a little slower that day, it would have.

What I am saying is this stuff is happening all the time, according to the laws of physics and the statistics of chance. Most of the time, passers by don't get hurt by vehicles spinning out of control, and most of the time, people don't win the lottery. Sometimes, they do.
posted by seanmpuckett at 6:44 PM on June 7, 2013


Suppose your machine is seeded based on time + location + some other local phenomenon with a few bits of entropy, using this seed to generate all of the random numbers. It's still not random enough. I can enumerate all of the pseudorandom lottery numbers that this station could possibly produce for a given lottery, given a few reasonable assumptions about things like the maximum rate at which people buy tickets and the resolution of the clock. This represents a tiny swath of the actual n-dimensional space that the winning numbers are sampled from. So when I go and buy a ticket from that machine, I am not actually getting a uniformly random number, and worse yet, there are many potential winning numbers which I actually have a zero percent chance of matching.
The winning numbers aren't being picked by a computer's RNG. The winning numbers are being picked by bouncing ping pong balls. The only things being chosen by a computer's RNG are numbers for tickets purchased by customers who don't bother explicitly picking their own numbers. If you could generate all such numbers, why on earth would you care? It's not like you can use that to make a better prediction for what number will win. I guess you could use it to choose numbers to avoid due to there potentially being more winning tickets with those numbers, like the "always choose numbers above 31" crowd, but that's extremely theoretical at best.

And it wouldn't matter if there were in fact many potential winning numbers that you had zero chance of getting on your ticket. As an example, let's make the lottery very, very simple: Let's say there's one ping pong ball, and it has numbers one through six (with equal probability). The computer generating ticket numbers has a very bad random number generator, which only ever generates the numbers one or two (with equal probability). I purchase a ticket generated by that RNG. What is my chance of winning?

Well, let's count. Here are all possible outcomes, all of which are equally likely:
  • Ticket 1, Winner 1
  • Ticket 1, Winner 2
  • Ticket 1, Winner 3
  • Ticket 1, Winner 4
  • Ticket 1, Winner 5
  • Ticket 1, Winner 6
  • Ticket 2, Winner 1
  • Ticket 2, Winner 2
  • Ticket 2, Winner 3
  • Ticket 2, Winner 4
  • Ticket 2, Winner 5
  • Ticket 2, Winner 6
Count them up. There are twelve possibilities, and again they're all equally likely. Of those twelve, two wind up with me winning. That's a 1 in 6 chance of me winning.

Which is exactly the same as it would have been if the RNG were able to generate any of the numbers 1-6.

"There are many potential winning numbers which the RNG cannot possibly choose" is completely irrelevant to your chances of winning.
posted by Flunkie at 7:28 PM on June 7, 2013 [1 favorite]


The winning numbers aren't being picked by a computer's RNG. The winning numbers are being picked by bouncing ping pong balls.

Right. I understand that. My point is this: if random-number tickets are generated based on an insufficiently random seed, then they are not being sampled from the same distribution used to choose the winning number. In fact, there would be many winning numbers chosen by ping-pong balls that cannot possibly be won by people who bought a PRNG ticket. Now, if that RNG ticket was the only way to play the lotto, this would still be fair... but people can write in whatever number they want, and they would improve their expected payout by writing in a number that the PRNG machines can't generate. Which is not fair.
posted by qxntpqbbbqxl at 8:26 PM on June 7, 2013


Did you read the rest of my response, past the first two sentences which you quote? I don't mean offense; it just seems like you didn't.

Again, "There are many potential winning numbers which the RNG cannot possibly choose" is completely irrelevant to your chances of winning.

As long as the winning numbers are chosen in a fair and equally probable way, the way in which the numbers on your ticket were chosen has literally zero effect on the chance of the numbers on your ticket winning.
posted by Flunkie at 8:32 PM on June 7, 2013


Editing your response to change meaning is frowned upon.
posted by Flunkie at 8:35 PM on June 7, 2013


Caught red handed! Twenty seconds is an eternity around here.
posted by qxntpqbbbqxl at 8:38 PM on June 7, 2013


I can't find any information about how the ticket selling machines generate numbers—whether via seeded PRNG, time-dependent PRNG, or some sort of hardware generator which is supposed to be a "true" random number generator.

It is trivial to design a seeded PRNG which will be able to eventually output all powerball combinations. There are only 175,223,510 combinations, which is less than 28 bits. So even the (comparatively weak) lrand48 generator would output them all. (proof: a properly-constructed LCNG eventually outputs all numbers 0..M, so if M ≥ 175223510, all values 0..175223509 will be output, and in no more than M trials). (for lrand48, M = 231 > 175223510). Here's such a program:
#include <stdlib.h>
#include <stdio.h>

#define N 175223510
unsigned char seen[N];

#define Q (2147483647/N*N)
long pick() {
    long r;
    do r = lrand48(); while(r >= Q);
    return r % N;
}

int main() {
    unsigned long n=0, m=0, k=1;
    srand48(time(0));
    while(n < N) {
        long p = pick();
        m++;
        if(m == k)
        {
            printf("m=%lu n=%lu [%.1f%%]\n", m, n, 100. * m / n);
            k *= 10;
        }
        if(seen[p]) continue;
        n++;
        seen[p] = 1;
    }
    printf("m=%lu n=%lu [%.1f%%]\n", m, n, 100. * m / n);
    return 0;
}
In two trials it took 3181257916 and 3578575031 pick() calls (i.e., 3.1 billion) to produce all 175223510 distinct values. Furthermore, in two runs after 100000 picks I'd picked 99960 and 99969 distinct tickets. Modifying the program to use high-quality random numbers (from the freebsd /dev/random, which gives cryptographic-quality numbers with input from entropy sources) the number of distinct values after 100000 picks was 99959 and 99972 respectively, so for this task even a "dumb" algorithm gives results that eyeball like a good cryptographic algorithm. (I was too tired to wait for the 'final result' of the /dev/random-using program, as that generator is not fast—it's intended for seeding PRNGs, not acting as a RNG directly)

So I'm not entirely sure what my point is, except that just by looking at "random" tickets generated by machine you're not likely to be able to detect by statistical methods whether they're from a weak PRNG, a strong PRNG, or a TRNG. Either analysis of the machine/its code or a claim by the people who run PowerBall would be needed to have a better idea what is actually happening.
posted by jepler at 9:19 AM on June 10, 2013


« Older Pacific Heights is just a movie, right?   |   Where Looks Don’t Matter and Only the Best Writers... Newer »


This thread has been archived and is closed to new comments