A group of recent Princeton graduates has won more than $6M in 66 lottery wins | Hacker News Reader

A group of recent Princeton graduates has won more than $6M in 66 lottery wins

indystar.com

166 pointsedot89 comments

The story of the selbees, a couple who made ~8m for themselves and their community from gaming a lottery, is a great read [0]

Jon Wertheim: You saved all the losing tickets?

Marge Selbee: Saved them in big totes.

Jerry Selbee: Big plastic totes.

Jon Wertheim: There must have been millions.

Jerry Selbee: 18.

Jon Wertheim: $18 million worth of losing tickets. And you have those?

Jerry Selbee: Uh-huh just in case we had a physical federal audit.

Marge Selbee: We had the upstairs of the barn. I stored them in one end and in the other end. And then I thought, "Oh no, this floor is gonna fall through." So then we stored them down in the pole barn. And we had probably 60, 65 tubs of tickets.

[0] https://www.cbsnews.com/news/jerry-and-marge-selbee-how-a-re...

That story is awesome

My grandpa had 4 or 5 huge garbage bags full. On his death bed he whispered to me and I had to go out and buy the numbers for him.

Why would they need to keep them? Did they buy them with cash or something? Wouldn't the credit/debit transaction be good enough to prove you bought it?

Maybe not everywhere, but in some states you need to pay for lottery tickets with cash.

Quant guy here. This boils down to information about the game, and maybe tax laws.

You have a load of scratch cards out there, most of them duds. If you come to a situation where a prize hasn't been claimed and there's few cards left to be sold, it's like counting Blackjack cards, the odds might change in your favour, or in favour of everyone who buys any ticket.

If that happens you need to do a bit of optimizing risk, assuming you have a limited amount of money to potentially lose. Kelly Criterion will have something to say here.

It really depends a lot on the details of the game. Are prizes rolled over if nobody found a winning card? Do cards expire? Is there a bulk buying discount? What is the nature of randomization? Does every region get its own guaranteed share of winners?

And crucially, is there information released about the progress of the game? This is maybe the toughest bit to figure out. With quant hedge funds you get this thing where you have to do a lot of asking questions about how the data was collected, to ensure your assumptions hold.

Tax I mentioned since it seems they formed a limited company. That's a whole game in itself, played by many many corporate entities in differing situations. You'd need to understand that as well. I know some traders whose entire jobs depend on understanding the difference between the tax codes of different countries.

I have no doubt you can get information that drastically increases your odds.

A long time ago, I worked for a retailer who sold lottery tickets. In order to sell tickets, I had to attend a government training session, at the actual government lottery center, which educated me on the different games and how they worked. It was probably the most interesting thing about the retail job.

The people doing the training were the ones that either designed or managed the games and payouts. Some were near the end of the their career and were happy to share the history of the games and the ups and downs of managing them.

I distinctly remember a couple tidbits where even my teenage mind went "well that's pretty relevant information to someone looking to game the lottery".

1. There was one game where the gov't actually decided to end the lottery because nobody ever won. The number of digits, the total possibilities and the number of people who actually bought tickets meant nobody ever won. So the gov't shut it down, but then players complained that they loved the game so much, so they started it back up. Still nobody won.

2. Scratch off tickets were sold in foil sealed packets of 20 tickets. The trainer dropped that there was always one winning ticket per pack, and only one. I had assumed that winning tickets were randomly distributed - some packets might have none, some might have 2 or 3. Nope, just one per packet. Now as the person selling the tickets, that's kind of important. I never mustered the energy to actually pursue it, but you could imagine how that could drastically increase your odds of a winning ticket.

1) I guess you're saying "don't play this game"

2) That could be used quite easily if you're the guy in the shop and you encourage people to scratch right there in front of you. If it's one out of 20 you barely need any info to come up ahead, since the odds/payout is probably within a few % of breakeven and each card starts at 5% of the pack. One or two losers and the rest are worth +ve, alterntively a winner and the other 18/19 are worth even less -ve.

1) Now that I think about it, that information doesn't tell you that much about the odds. The odds can be calculated independently. Whether or not people win is kind of irrelevant to the odds. It has more to do with the number of people who play (so the odds of a winner in the group is low). I just found it interesting that the gov't was like "jeeze, nobody is winning, let's just shut it down. And the players (who never win) are like "no!".

2) Yeah, the 1 in 20 winner really jumped out at me. Open a new pack and someone scratches a winner and you know with 100% certainty the other 19 are losers. Or, you sell 15 out of a pack and all are losers (most buyers scratched right away) and of the remaining 5, you know there is 1 winning ticket.

I worked at a bar that had those stacks of pull tab cards. Random cards in the stack would give you cash that you exchanged the card for.

The manager of the bar would know which ones were redeemed so if it got low and he knew the big money cards weren’t won yet, he’d go play the machine until he got it.

> If you come to a situation where a prize hasn't been claimed and there's few cards left to be sold

You'd have to also account for cases where the prize-winner just hasn't scratched their card yet? (Or, as you say, whether and how this has been accurately reported)

> If that happens you need to do a bit of optimizing risk, assuming you have a limited amount of money to potentially lose. Kelly Criterion will have something to say here.

It seems to make sense that they've set up as an "investment fund", since they need a bunch of cash to buy the tickets in the first place, and even more cash to reduce the fund's risk enough to get an OK from the Kelly Criterion.

(In extreme cases full-Kelly can advise you to borrow cash at your own risk, but the odds have to be wildly in your favour - if they were dependent on one ticket for a net profit, that doesn't appear to be advisable)

What a poor excuse for news.

> "The four participants have not been accused of any crime. In fact, lottery officials in Indiana, Missouri and the District of Columbia told IndyStar that Montori and Davinroy, like all major prize winners, were vetted before receiving their winnings.

> "We have not conducted an investigation that fell outside our standard clearance processes for big winners." Wendy Baker, communications manager for the Missouri Lottery, said in an email to IndyStar last week.

And yet the IndyStar has publicised a heap of personal information about the winners online. Why?

It can be interesting without being a crime.

Interesting enough to dox them?

In many state some of your personal details become public info in order to claim your prize.

The norms for doxing in media are going to have to change. It's a tenet of traditional journalism that everyone in a story must be doxed. A local newspaper wouldn't write a story about someone being burgled without describing them along the lines of Julie West, 53, unmarried, formerly of Westford but now living in Sackville Close, Eastford... I don't know the underlying reasons, I think it derives from principles of reporting truthfully combined with NOT having a principle of not giving personal information. But it seems weird as hell compared to more recent views on privacy an online anonymity.

To me it seems like this is also a very US American thing. In Germany for instance it is common, even for the most tabloid papers to never publish full last names or clear pictures. Most of the time in pictures faces are made blurry etc.

In Germany privacy is a real thing, in the US it's a figment of everyone's imagination.

I think you're asking a very good question. Thinking more deeply upon it, the angle of this journalist is that this group is sort of a reverse Robin Hood, as in theyre taking public funds in a way from being distributed to the sort of the mathematically challenged. And instead Distributing it to some mathematical wizards that are jetting in from the East Coast. It's sort of highlighting The Divide between the East Coast Elites and rural America.

It's often the law to publicize the winners names and the location the winning ticket is purchased; this helps the losers have confidence that the winners aren't insiders.

Because events other than crime can be newsworthy? Pro athletes make the news for winning games every week.

> She did not respond to a phone message or attempts to reach her through family and social media.

uh yeah, these grifting creeps posing as journalists are exactly what people expect when they win lottery tickets and claim them non-anonymously.

these students are clearly fronts for limited partners that found a way to stay anonymous

>uh yeah, these grifting creeps posing as journalists are exactly what people expect when they win lottery tickets and claim them non-anonymously.

Is there a reason you consider them not journalists?

https://news.ycombinator.com/item?id=25225484

They're not journalists because they published too much info? It seems a lot more like journalism than a lot of stuff I see published such as "Top 10 x for you to y with", or simply restating a bunch of facts from a different news article. This involved independent research.

They doxxed people who did nothing wrong.

In my mind doxxing means displaying private information. Also in my mind Linkedin information isn't private.

Lately there’s been a burst of intense anti-journalist sentiment generated by various tech plutocrats who would prefer to be free of public investigation or criticism. This sentiment trickles down to professional class workers with aspirational visions of becoming plutocrats, where it is enthusiastically aped and amplified, and applied indiscriminately to basically any journalistic endeavor, under the principal that the wealthy and influential should never have to be exposed to the public eye on anything other than their own terms

someone winning the lottery and getting their family members bombarded on Facebook messenger by strangers has absolutely nothing to do with professional class tech workers echoing their amoral leadership

kthxbye

The recent outburst of hostility to journalism in the tech sector is absolutely the result of a lot of tech and tech-adjacent sacred cows coming in for public examination and criticism for the first time, and a driving factor is absolutely driven by the tendency of tech sector workers to personally identify with their employers. If this is not apparent to you, then you are

But regardless, if you come up with any kind of system to generate millions of dollars, whether it's a business or a way to game the lottery, you should absolutely expect to be subject to public scrutiny and it's ludicrous to frame a Facebook DM as an unacceptable invasion of privacy (making this argument while working in the surveillance-tech sector elevates the absurdity to self-parody)

It is unacceptable to try to track someone down about money through their parents.

You are conflating this with that other conversation you keep trying to force happen.

This conversation has nothing to do with mirroring tech leadership or personally identifying with employers. I am not an employee anywhere and you failed to predict what I do as well. Moving on before your next ad hominem finishes rendering, they could consider using states and game combinations where anonymity of winners is protected and maybe they already are. There is an argument to be had about anonymous winners.

I'm not talking about you personally, obviously I don't know or care what you do or who you are (although if you suddenly came up with a way to make $6 million, perhaps I’d take an interest). I'm merely making a correct observation about a phenomenon of which you are a small part.

It is of course even more ludicrous to expect anonymity when you have solved the lottery in states that do not permit anonymous winners.

> It is unacceptable to try to track someone down about money through their parents.

As a blanket statement, that's nonsense. Stalking someone's parents to collect an unpaid medical would certainly be unacceptable. A polite DM or e-mail requesting an interview with someone who has apparently come up with a foolproof method to generate millions of dollars is entirely reasonable.

That is how journalism traditionally worked. The names were always part of it.

would the person being bombarded by strangers be able to distinguish?

does it matter if they could distinguish if the person reaching out is saying they're going to put all their personal information online in an even more digestible format so that more people will harass their parents on Facebook Messenger?

grifting creeps.

> would the person being bombarded by strangers be able to distinguish?

If it's via email you could see the domain in the email address. Back before email, journalists might contact people via phone, and I'm not sure if it's so easy to tell then. I assume the email/phone call/facebook message starts off with an introduction saying something like "I'm a journalist with such and such a newspaper..."

> does it matter if they could distinguish if the person reaching out is saying they're going to put all their personal information online in an even more digestible format so that more people will harass their parents on Facebook Messenger?

Do you think the personal information is irrelevant to the story? Do you think this is a story that shouldn't be covered? I think it's an interesting story, and giving some info about the characters in the story helps make it more interesting. It's generally considered proper form to contact the subjects of your article, so I would think it would actually be worse if the journalists didn't contact the people in the story. The purpose of putting the info about the characters is certainly not so that people will harass their parents on Facebook Messenger. It's not like they included their email address, phone number, link to a Facebook profile, or physical address. I don't consider stuff on my Linkedin profile to be creepy. My Linkedin profile is my brag page; it would be great if journalists were writing about stuff on my Linkedin profile.

> grifting creeps.

Are they worse than journalists 50 years ago? Would you consider all journalists grifting creeps?

It’s fairly amazing to me that the person being accused of being a grifting creep is the journalist who tracked down quotes in several states from witnesses to a notable event instead of people who are clearly running a game on multiple lotteries.

Norms have really changed in the last decade or so.

the journalist is acting like a stalker wanting something from a lottery winner, and is indistinguishable from the other people out there that would be doing the same thing

someone playing a lottery in a way that no lottery authority disagrees with would never be called grifting.

> is indistinguishable from the other people out there that would be doing the same thing

Except for their email address that is ReporterName@local-newspaper.com...

This reminds me of the WinFall lottery [0], where if no one got a grand prize, the prize money was rolled over to the next week. As a result, if there were enough prizes rolled over to the next week the lottery could actually have a positive expected value, so groups of people started spending millions on tickets in these weeks. The lottery organizers didn't really even care much, since the people spending millions on tickets just increased their profitss.

[0] https://www.theatlantic.com/business/archive/2016/02/how-mit...

Of course they shoukdnt care, because, if I remember correctly, it was only a positive payout if someone didn't win. Basically what it was doing was spreading the rollover along all winners" rather than growing the main jackpot. So the organizers were paying out the same amount of prize money.just under a different distribution.

An idea for lottery payouts that are fun for a lot more people. Set a top-prize limit - say $50M. Now suppose the jackpot has risen to $200M. When someone claims the top prize, pay them the $50M. Then select 150 more winning numbers, randomly picked from the numbers sold, and pay each $1M. Or pay out 1500 $100,000 winners.

I read a similar story about someone in North Carolina a bit ago. He identified a scratch-off lottery game that was near its end, with an unclaimed top prize, and bought all the tickets left at gas stations in the less-populated parts of the state. He found the winning ticket.

>Even if the group developed a system for using public information to improve their odds, experts expressed skepticism that it could result in reliable profit.

My hunch is that winning tickets are not in fact equally distributed, and they have discovered something regarding this. Going by already claimed prizes gives some information, and I am sure there's a decent way to model the relationship between the set of claimed tickets to that of the unsold winning tickets (to account for the "safe-deposit box" cases mentioned in TFA, which strikes me as a bit of a red herring).

Actually the winning tickets ARE equally distributed. They are not randomly distributed. Every roll/batch has the same number of small and mid-prices, let say under 5k$ to 10k$. As for the big prices they take care that they are evenly distributed across the country and over the lifetime of the game.

A smaller scale trick is to social engineer the cashier to know which rolls have already released their prices, and buy from other rolls. Cashiers often tend to naturally do this with there best customers.

On the tax side, it is favorable.

Under default tax rules, each losing ticket is deductible against the winning ticket's tax.

Under a professional gambler election, each losing ticket is deductible as a Net Operating Loss, which allows deductions against everything and rolls over (and sometimes backwards) as a deduction into different years. Sadly, the tax reform law has severely limited this until 2026.

Thank you for clueing me into professional gambler election. This changes the calculus.

The trick with taxes is to put every specific interest into its own business entity that just does that

That business entity picks from a catalogue of alternate tax codes

The entity type doesnt matter, so even if profits flow directly to you personally in a disregarded entity or partnership, the money and actions pass through an entity level force field that changes their tax treatment

“The people” will never get this, its fine

That is not correct. The entity type does matter; if an entity of disregarded it is treated as not existing for tax purposes (hence the label). There is no magical force field that changes the tax treatment of tax items for disregarded or pass through entities.

Net Operating Losses generated in a disregarded entity are passed through to the individual level. You can convert capital losses to NOLs, you can convert gambling losses to NOLs, and a few other conveniences.

Pick whichever source/year combination you choose to respect more. Or present a conflicting source.

https://www.carnahanlaw.com/net-operating-loss-nol/

https://ttlc.intuit.com/community/taxes/discussion/can-a-cal...

https://www.mwellp.com/deducting-pass-through-business-losse...

No, you have misread your sources. Capital losses offset capital gains, and if any losses remain, up to 3000 can offset any other type of income, which does not require conversion to an NOL.

Corporations can only use capital losses to offset capital gains.

In all cases unused capital losses carry over to future years.

Source for in depth discussion. https://www.journalofaccountancy.com/issues/2001/nov/getmaxi...

Section 475 elections convert capital losses to NOLs, and therefore information about capital treatment is not relevant or accurate as that is a parallel default tax regime.

Other professions have other elections.

You do the elections at the entity level so as not to mess up your own tax regime.

Instead of your article from 2001, you could pick some part of my articles, such as the parts where it says disregarded or pass through entities work as expected. Your article from 2001 doesn't mention those magic words at all, only discussing partnerships in the non-business tax context.

From my third source using post-2017 tax regime:

> For business losses passed through to individuals from S corporations, partnerships and LLCs that are treated as partnerships for tax purposes, the new excess business loss limitation rules apply at the owner level. In other words, each owner’s allocable share of business income, gain, deduction or loss is passed through to the owner and reported on the owner’s personal federal income tax return for the owner’s tax year that includes the end of the entity’s tax year. [and carried forward to the next year, explained in the following paragraph]

sigh, are we done yet

I guess I'll have to tell the partners that some guy on the internet says we've been doing things wrong...

You're still misreading the laws. If you refuse to accept that than we are done. Section 475 are the mark to market rules, not capital gain conversation rules. It applies specifically to dealers of securities (who would otherwise be unable to take deductions for their normal business losses.)

I charge $700/hour. Most of my work now involves cleaning up messes by guys who read one paragraph out of context and got themselves audited when they thought they found a new loophole.

Lottery expenses do not qualify for this deduction, and never have.

Moreover, even under this deduction, gambling expenses may only offset gambling winnings, so in years when losses exceed winnings you cannot use the deduction to offset non gambling income.

Only table games qualify? Or “skill” games or what?

Generally any game you would find in a casino. Basically anything you could do as a profession which is why the lottery and raffles don't count.

This source links to an indepth source which is unfortunately not available for free online but which is probably the best resource outside of reading IRS regulations and cases. https://www.journalofaccountancy.com/issues/2016/oct/taxes-f...

thats helpful to know thanks

It hints at the method right here:

> Lottery officials said there is nothing wrong with players using data about the number of prizes outstanding during the lifetime of scratch-off games, which is regularly updated online

So not really news. It's just a group of people had the nerve and initiative to do it and hope that the math works they way they expect.

And the money to go buy some pricey tickets; if I had a dad who was a rich businessman who could probably support the first round, I'd totally do something like this too.

An (IMHO) more interesting story of somebody actually "hacking" the lottery (well, scratch cards): https://www.wired.com/2011/01/ff-lottery/

My guess is they figured out who prints the lottery tickets, then gathered data on the ticket ID numbering system produced by that printer in the past(winner vs loser ticket ID's), then trained that data via deep learning to figure out probabilities of jackpot likelihood for each ticket ID # (or by range of numbers), then figured out the historic distribution patterns in relation to the ticket ID #'s, then pinpointed the specific places actually selling those ticket ID's, then drove to West Bumblefuck and bought all those tickets.

That's the nefarious solution.

The brute force solution seems to be use published redeemed winning ticket numbers to estimate sold ticket numbers? And then run expected value calculations of the unclaimed prizes vs the unsold tickets; if positive enough, go out and buy as many tickets as you can, process them and claim the winners (and enter any second chance drawings).

Of course, if people delay claimimg prizes, it makes the estimates inaccurate.

My guess would be employee fraud. Nothing against the people mentioned in the article, just where I'd start looking. There are umpteen posts in this thread already about potential interesting ways to game lotteries, but the historic way to get an edge is to have an insider.

Rules to one of the mentioned lotteries. [0] It looks like prize counts are fixed in advance, so any information about redemption changes odds. Blacklight also shows that the lottery website is riddled with trackers, so it is possible they are inferring redemption from advertising data.

[0] https://hoosierlottery.com/getmedia/33f6f7bf-f0e6-4265-bcbe-...

There’s a whole chapter in the popular mathematics book “How to Not Be Wrong” by Jordan Ellenberg that covers successful lottery strategies and the math behind them. As I recall from the book, the basic strategy was very simple and had a few groups of practitioners. Optimizing for increased incremental returns was far more sophisticated, though. There were also the operational challenges of buying and tracking so many tickets.

The game is coming to an end, and the big prizes are still out there. Tickets cost $20 to $30, so fewer players.

>The manager said they paid with cash or cashier's checks, and they weren't just buying at Break Time.

Couldn't they pay with credit cards and get credit card rewards to help their margin?

Generally credit card companies consider lottery purchases to be cash advances, if they allow them at all.

Probably just another scheme for rich daddies to launder money to their spoiled kids without paying tax.

Making money these days is all about having a rich daddy, rich mommy, sugar daddy or sugar mommy.

Gold diggers are the closest we have to 'self-made' these days.

Maybe just boring old school money laundering?

Why would you use Princeton grads? And money laundering this way has got to be hella inefficient, no?

It's definitely not money laundering, but are Princeton grads too good to be involved in illegal activities? What a weird statement.

Rich people don't need to risk their wealth by signing up to be mules.

Too good to be the mules, yes certainly.

There was a Columbia undergrad that was charged with joining ISIS, this is a silly argument.

NCSU grads may have done illicit things too.

Money laundering would rather work different way. Someone who gets winning lottery ticket instead of claiming the prize himself sells it for a double to other guy who needs to legalize his income.

In other variant the ticket is sold to someone who later uses it to pay a neat bribe...

This is a rational move if one wants to increase his wining and possibly to make it tax free in return for taking a little additional risk.

This way of profiting doesn't need money laundering to work. But of course if you were the mob it would make sense to hire these guys, as you could regularly launder some money, and they would get paid even more than the winning tickets.

This kinda reminds me of MIT Blackjack Team

(Deleted)

This and they have a large number of tickets to examine for patterns along with locations to look for patterns in physical ticket distribution.

They are scratch offs

Alternatively, they could do the math and figure out when the expected value exceeds the purchase price. That's how these games have been won in the past. It's explained a bit in the book, "How Not to Be Wrong: The Power of Mathematical Thinking" by Ellenberg, Jordan

Such a pity young capable people spend their most productive years doing something entirely unproductive. And one more thing. Are they winning with scratch cards only? Then it could be that they are utilizing some equipment to look through the paint, with accomplices in a ticket factory or distribution chain.

There are endless things I could do with my free time that would be far more productive & useful to society than anything I could realistically be paid to do. Having lots of money would enable me to do those things.

Making (probably) millions you find unproductive?

economically unproductive I think is what they meant, you could say society is a lot when their talents are used for gambling vs something like scientific research

What about high-frequency traders?

Profitable, technically challenging, but in a very real sense entirely pointless and contrived.

I meant to add finance as a good example of what I meant, including HFT. basically just saying that while individually people benefit from working in these schemes, the broader economy arguably benefits little to even negatively

HFT isn't my cup of tea, but the rise of HFT has certainly narrowed spreads and contributed to zero comissions; and I apprechiate those effects.

Reply on news.ycombinator.com