I am a former professional gambler and now I don’t gamble much, because I have neither the time nor interest to find good wagers, but what I do is on predictit. Being a great gambler mostly means not gambling until you have an edge, and what he says about and market being less accurate than 538 is true in my experience. I’ve been betting on there every couple years for maybe ten or so, it’s held up well.
As many mentioned here, the fees are high. That contributes to market inefficiency for sure.
Arbitrage opportunities pop up there all the time, but ones that are profitable after the fees less so. They do happen though.
He is neglecting how long contracts take to settle. There’s a lag of two months between when the election and when president is sworn in. It’s not unreasonable to sell your winning position at 98% just to be able to bet on more things.
It was the first time in 20 years of hoping for such an opportunity that I managed to get to bet on something that had already happened. That’s every gambler’s dream and I’m unaware of many instances of it happening without breaking laws or trickery. But even then, you bet 93 cents to win 7 and then pay 5% on cashout so it wasn’t exactly making people rich.
I guess the people on and other side of the wager would have agreed with you. I don’t know what odds there were of the election being overturned but I am sure they are much smaller than the odds those people were laying on it happening.
you took on the settlement risk and got paid for it, if you look into financial markets and central clearing and OTC contracts, central clearing/exchanges effectively get rid of that issue.
> I managed to get to bet on something that had already happened.
In less liquid in play sports betting exchanges have this happen all the time - generally you are getting $1 for every $1000 wagered and settlement will take anywhere from minutes to days depending upon the event.
Predictit during the 2020 election was just about the closest thing I've ever seen to free money. Shares for Texas going blue were still at $0.30 on election night. So many people wagering with their emotions and tribalism, it was unreal.
> Shares for Texas going blue were still at $0.30 on election night.
You think that's bad? Every month or so for like 6 months, there was another bet that the FBI was going to arrest Hillary by the end of the month and odds would be like between 7% and 13%. Every month, I bought more shares it wouldn't happen. The comments were all QAnon crap about the "The Kraken" coming. It was hilarious. Some of the comments were from me because I didn't want the gravy train to stop. Wasn't a full 7% return, though. Definitely lots of fees on Predictit.
> It was the first time in 20 years of hoping for such an opportunity that I managed to get to bet on something that had already happened. That’s every gambler’s dream and I’m unaware of many instances of it happening without breaking laws or trickery.
There was a Chinese lottery scandal in the semi-recent past. Someone noticed that the winning numbers were posted before the close of ticket sales, and bought a winning ticket. And he completely got away with it.
Can't speak from personal experience, but OP laid it out pretty clearly. You wait till you have an edge, then you bet. There are various books that try to capture it. Doyle Brunson's super system is like way outdated, but there are some funny bits about being a Texas road gambler getting run out of town. Annie Duke has a book I believe. My window into that world is pretty narrowly focused on poker.
I'll always recommend _Positively Fifth Street_ by James McManus. He doesn't really lay out how to do it step by step, but takes you along for a ride as to what the world looks like from that perspective. kinda.
Another big aspect of professional gambling is actually being able to place bets once you have an edge. This is obviously not an issue in poker but is in most other games:
The casino will ban you if you count cards in blackjack.
The sportsbooks will limit your action or ban you if the figure out you have an edge (or just see that you're winning or betting in certain patterns).
The most open information is about poker. See the forums. Many of the concepts there apply to other sports. The idea that you don’t know if you are any good
until you played 10000 hands etc. Trying to make money off lots of little edges. In a sport you probably need to model every game very well, use AI/ML (mostly linear regression though). Honestly it is boring frustrating work IMO.
10,000 hands in poker is nowhere even close enough to get a significant sample, unless you're playing extremely weak players where you just have a girnomous edge. But those sort of games are very uncommon. There were plenty of instances of strong, winning players going through swings that lasted hundreds of thousands of hands. A more reasonable baseline sample would be ~500,000 hands.
Frustrating is definitely a great way to describe the game, and emotionally crushing. Like imagine one is playing a 5/10 no limit hold'em game - so people are generally buying in for about $1000, unless they're from Spain. An extremely good player might have a longterm winrate of around 3bb/100 there. In 100 hands, there's every chance you'll be involved multiple pots, easily for thousands of dollars. And your long-term expectation is about $30 out of all of that. So you'll go through short term swings where you're up tens of thousands of dollars in very short periods of time, swings where you're down tens of thousands of dollars - all to get your $30 per 100. And you never become completely numb to the money, even after playing millions of hands.
You are probably right. I could have misremembered the number of hands, the main point is it is alot more than most think. One to ten games of soccer is enough
to tell roughly how good someone is, by comparison.
I mostly played poker, where you can have an advantage over other players that is greater than the rake the casino takes. But I also found all sorts of random opportunities here and there where I had an edge. Where a casino comp system gives you more in rewards than the game takes from you for instance. (That’s both common and legal but they will cut you off because you are essentially taking the comps they want to go to the people they make more off of so it’s a lot harder to make a lot at than it used to be.)
Or when an online casino is not careful enough about how you earn the bonuses they give you. Things of that nature. There are people also who can do it at things like horse, racing or sports betting, but I never knew enough about either of those or wanted to learn.
My Predictit wagering has mostly been on the theory that exactly what this article says is true: people are betting who they like and the polling aggregators are better. I really just look for arbitrage opportunities or discrepancies between the market and 538 that are large enough to overcome the fees.
The fees are high enough to make it not happen a ton, but it does happen.
The IRS has various rules in place that determine whether gambling is your business / profession. If you qualify as a professional gambler, you can deduct losses against your winnings (just like casual gamblers can) but you can also deduct your "business expenses" (e.g. cost of travel, hotels, meals, etc that you spend in pursuit of your business).
2017 tax reform took all of that away for gamblers. they are on as equal footing as casual gamblers, at least for when net losses exist, there are no more NOLs to carry forward (or back) which is what made that status the most useful
You can grind out poker games for a living, especially when there is a steady stream of rubes that think they’re hot stuff because they clear out friendly home games. In the long term those kind of people always lose to the pros.
I grinded online poker games for a summer in college. I observed opponents betting behavior and classified them into one of a half dozen strategies and then ran a strict algorithm based on who I was betting against. I ran three tables at the same time, finding success in rapid low stake games. My edge at the time was varying my play based on my opponent. Many of the other professional low stake players at the time were running simpler strategies on more tables.
Haha it was not. While I goofed around with that in the late 90’s/early aughts I never believed it was a long-term profession and never really won or lost much at it.
like r/wallstreetbets. Some people making consistent huge $ over there. You look for a setup which has a edge, like buying 0-day SPY calls if the market is down a bit. Backtest strategy and use something like the Kelly formula to find optimal betting amount. if you can make a living at this, then I guess that would make you a professional.
Also, there is no need to gamble when you can just park your
$ in leveraged tech funds or even something like QQQ and make a >30% ROI every year, which has been the long-term return since 2009. This is best for large amounts of cash. The huge bull market in tech, like Nvidia and Meta stock, made other type of speculation unnecessary. A lot of people doing this now.
The author does not understand bid-ask spreads, does not understand distortions created by fees, does not understand that the last trade price is not the same as what you can get at present, and just generally does not understand the realities of trading in markets. This article is a zero-information void.
> average predicted likelihood is frequently illogical
What is being pointed out is that is not illogical when you fully understand it.
I will let you bet me that Taylor Swift is not the current president of the United States. For every $1 you bet, I will return $1.05 to you*. You can bet as much as you like and as often as you want - my email address is in my profile.
But those things at least partly explain the apparent irrationality.
PredictIt charges 10% fee on profits and another 5% fee on withdrawals. So bets that have the potential to earn a few cents on the dollar aren't necessarily worth it. That's before considering how long you have to lock up your money in this contract.
> The implied probability of Brexit a few days before was around 25%.
It was one of the 25 days out of 100 not one of the 75 days out of 100.
> these dislocations
The market puts the chance of flipping a coin and getting heads at 50%, when it comes up tails it doesn't mean that the market was irrationally pricing heads.
So? 25% doesn't mean it can't happen. I wouldn't cross the road if there was a 25% of being hit by a car.
The real way to measure predictive accuracy is to find _all_ the times they gave something a 25% chance of happening. If the predictions are accurate then roughly 25% of those things should have happened.
If I am asked to predict whether it will rain every day for a year, and every day I blindly predict 30% because on average it rains 30% of days in this location, am I correct? Maybe, but I’m not a very useful forecaster.
What you should optimise instead is something like log loss between the given probability and the true outcome (0 or 1). That way you’re rewarded not only for being right, but for being confident and right.
If anything then that means that they work! Situations with a 25% chance do occur sometimes! If prediction markets show 75% odds and are right 3/4 times, they are doing their job well - if they show 75% odds and right 4/4 times, they are doing their job badly!
Meanwhile, both polls and analysts got those elections completely wrong, much more so than prediction markets. (IIRC, 538 was the only outlet that gave Trump a >20% chance to win, and they got a lot of bad words for their prediction pre-election.)
Unfortunately, people keep falling for this. You can always cherrypick elections from 7 years ago as "evidence" against prediction markets, but if you look at the bigger picture, you will see that they get a vast majority of elections right. But when they don't, even just once in a decade, they get all the blame.
538 got screwed in both directions: before the election they were derided for having Trump too high. After the election they were lambasted for "only" giving him 20-30% chance.
I'm forever puzzled by the lack of rigorous coursework in probability and statistics in the American public school system, at all levels. I mean, heck, it's pretty much the only mathematical subject for which the question, "When are we ever going to use this in real life?" is trivially answered! And all the other more traditional math subjects (algebra, calculus, etc.) fall right out of it anyway. It's the perfect foundation.
Honestly, why hasn't probability-centric mathematical pedagogy taken the world by storm?
> The implied probability of Trump beating Clinton a few days before was around 20%.
Unclear if you think this is incorrect?
Obviously it happened, but most reasonable prediction methodologies put it between 20 and 30%. Notably 538 and Superforecasters (via The Good Judgement Project at the time) both had it in that range (as did prediction markets). There were some outliers - in particular some newspaper(?) had a statistical methodology that gave Clinton a 95%+ chance. In a two horse race that is obviously wrong!
I was working in the forecasting field at the time, and that seemed about right. It was an unlikely but not impossible outcome.
The polls implied about a 20% chance of Trump beating Clinton too. I guess I'm not sure what you're getting at. That seems like the perfect example of not a sure thing.
The author claimed that 90% was illogical; that it should have been closer to 98-99%. What the author failed to realize is that there are overhead costs in betting -- 10% rake, 5% withdrawal, etc. that cut into that 98-99% margin.
Also isn't there a small bet size limit on predictit? That would screw up price discovery because a single informed trader can't correct any inefficiencies.
My cynical observation here is that betting odds[1] aren't looking great for the incumbent. The author, Jeremiah Johnson, is the founder of a Washington NGO for the center-left [2] and probably has some skin in the game in keeping the status-quo in the White House come November.
On Betfair, which has miniscule fees compared to predictit and effectively unlimited wagering caps, the same "You can bet on Joe Biden to win the election weeks after Joe Biden won the election" phenomena existed.
Political Markets are filled with dumb money and not enough sharps. They are absurdly profitable to bet in bit are not frequent enough for pro gamblers to focus on.
So I buy a "no" contract for 98 cents, and sell it for $1 in most of a year. Maybe someone who has used the platform can chime in, but the terms of service seem to suggest I don't get any interest on my contract money meanwhile. So I'm underwater already probably, then I pay 5% to the platform when I withdraw. Obviously the time-value thing gets better as the end approaches, but you'd expect most low-probability events to decline in price too.
You are absolutely correct. It doesn't make sense to buy that contract unless you already have the money in your PredictIt balance, you want to keep it there as "dry powder", and you really can't think of a better market to park it in (unlikely). Time value of money being what it is... contracts like that have a pretty dead market. There is no reason to hold either the longshot YES that's realistically less than 1% likely OR the tie-up-your-dollar NO.
Of course if you already have a balance on PI, you just won in another market, and say it's Nov 6th and you've got nothing better to do... you might as well buy 99c shares because the $1 payout could be in as little as a couple days and what else are you doing with the money? So the high priced shares DO eventually sell but basically anything that's months away dies out in trading volume before it reaches the 99c mark.
Personally if I still had money sitting in PI I'd probably park my dry powder in NO on "Joe Biden resign in his first term", which has been as high as 10c this month, but even assuming I could get 90c NO shares that is not worth me moving money from my bank back into PI just to skim those pennies.
The article is basically asking why the smart money hasn't pushed the No contract to trading for 99.9 cents or something. I'm saying the smart money would walk away at 96 cents or so because it's a lose-lose at that point.
The 2020 elections were such a shitshow on PredictIt. So many trades were skewed by partisan bias that it was shockingly easy to make money. My favorite trade that year was for some reason the "market" was pricing in a lower turnout in 2020 vs 2016. I grabbed a strip of the over bets at big discounts. I still to this day don't understand what people were thinking with their bets there. A hugely contested election during a pandemic when there was nothing else to do, of course there was going to be huge turnout!
I really enjoyed the market "Will either Biden or Trump concede within 2 weeks of the election?"
As I recall "no concession" was priced at like 40c, at the same time that Biden winning the election was priced at something like 55c-60c. I was pretty sure Trump was not going to concede, period, so this was just a substantially cheaper version of betting "Biden wins". And as it turned out, it also had the advantage that it closed much, much sooner than the actual election market -- thus freeing up my funds to go into those other markets that hung at 85c or so till December.
The article briefly mentions causes for market inefficiency such as the time value of money and platform fees, but then inexplicably goes on to ignore them as potential reasons for inefficient markets.
Plenty of arbitrage opportunities exist on PredictIt that can’t be taken advantage of either because the platform’s cut of any profit you make on a bet will negate the arbitrage or because resolution is so far off that it would be more profitable to just invest in a CD.
I would expect that platform risk would be even more significant than fees and time. I haven't looked into any of the popular prediction market platforms, but by default I would be more concerned about the risk of them resolving markets in ways I fundamentally disagree with or simply ceasing to exist.
That is not how I read it at all. To me, it reads as "there are likely many reasons these markets aren't operating efficiently".
The time value of money and fees certainly are part of it, but they don't explain things like, for example, the weight of money in the Trump-by-280+-EV contract.
The point (which the author misses) is that irrational actors (such as those putting money in Trump winning by 280+) wouldn't significantly distort market odds if the market weren't severely crippled by so many restrictions and fees.
> We know that this tidal wave of money is dumb money because we can see intense tribal behavior on these sites and research tells us that the more one identifies with a “team,” the worse one’s predictions for that team are. People will bet for candidates and parties not because they have an evidence-based analysis supporting their bet, but as an expression of identity.
This first sentence is completely unsupported. How do you know that the volume of comments tells you the amount of money bet?
Actually, sports betting does have a fan bias. I doubt many 49'er fans are going to bet on the Chiefs in the Super Bowl. It could well be that prediction markets give the wrong answer, but I'll still give them as much weight as FiveThirtyEight. Or more. FiveThirtyEight is hardly non-ideological, cold-eyed realism.
It's also true that smart money is not feasting on the dumb money, as a market really needs. The low volume and high costs of betting do make prediction markets less than perfect.
There's no problem here at all; except perhaps the wildly absurd fantasy of "perfectly rational (and/or intelligent) markets," which the popularity of e.g. lotteries and casinos immediately disproves. Not much to see here.
The magic of a prediction market with a reasonable sampling of people is that the players don't need to be rational, for things that are publicly decided. Imagine every person who went to vote, also bet in a prediction market based on nothing except who they wanted to win - the prediction market would be right 100% of the time. And it'd similarly be right most of the time, with a good random sample.
That said, even for things that aren't publicly decided - the 'wisdom of the masses' [1] is very real and kind of weird phenomena. If you get people to try to guess how many beans are in a large jar, you'll get some people making really stupid guesses like 100, or a million. Yet when you average out the result over all answers, it tends to be extremely close to the right answer. At least unless you let people collaborate, then the effect disappears and we become stupid again. Kind of explains a lot of things.
>
There's no problem here at all; except perhaps the wildly absurd fantasy of "perfectly rational (and/or intelligent) markets," which the popularity of e.g. lotteries and casinos immediately disproves. Not much to see here.
The behaviour of customers of lotteries is indeed clearly not perfectly rational if you consider the monetary expected value of your wins. On the other hand, the customers' behaviour gets much more rational if you consider the customer's utility function: a huge lottery win has a disproportional benefit for the customer with respect to the customer's utility function (since the lottery win will hopefully change the customer's life). In other words: what the customer optimizes for is the expected value of his utility function.
I have heard this argument before but I don't get it. Especially since evidence seems to show people generally have concave utility for money. (Under the assumption of concave utility, insurance is a better deal than it seems, but lotteries a worse deal than they seem -- even when one diversifies.)
What is the exact shape of the convex utility function you theorise these people have? And what data supports that?
> (Under the assumption of concave utility, insurance is a better deal than it seems, but lotteries a worse deal than they seem -- even when one diversifies.
Indemnity insurance and lotteries are actually symmetric to each other: indemnity insure is about preventing big losses in a rare event, and lotteries are about winning big in a rare event.
The central difference is that the behaviour of people with respect to wins and losses is not symmetric, that is why insurance policies and lotteries are marketed quite differently.
>
What is the exact shape of the convex utility function you theorise these people have? And what data supports that?
I am very sure that big lottery companies have data available to answer this question, but they won't make the data available. So I honestly admit that I don't know the answer, but I can propose an experiment how one might get a partial answer:
Create some lottery offers, paying out, say, 1000, 10.000, 100.000, 1.000.000, 10.000.000 $ with probabilities such that the expected payout is 5 $ for each of them. Now sell/auction the lottery tickets and look what prices customers are willing to pay. This should give a rough idea how the utility function might look like.
You mean the expected payout is -$5, right? Otherwise the rational person with decent means -- even with a very concave utility -- would pay near $5 for that ticket, and this would be a Kelly-correct investment.
The discussion was specifically about negative-EV lotteries since those are the ones that require convex utility.
No, 5$. The reason is that you sell/auction these lottery tickets. The price that the customer pays has to be subtracted from this payout (and this way the final payout very likely will become negative). But the price that the customer will pay in this experiment is not known beforehand, so the initial payout has to be positive.
Lotteries are rational when you consider the entertainment value. Most lottery players dont expect to win, but as long as theres the possibility it allows them to daydream about what they would do if they did.
Indeed. Negative-sum games (like lotteries) are a form of consumption. From the player's perspective they should be compared to other forms of consumption like buying a beer or a rollercoaster ride. You do it for temporary pleasure, not as an investment.
Most players are fully aware of this, just as most beer-drinkers drink in moderation. (See Brenner, Brenner & Brown for more data.)
The prospect of becoming rich isn’t as enticing when you’re already rich. What would a rich person who wins the lottery do? Quit their job I guess but mostly quality of life unchanged. They already have most of what they want. Not the case for poor people.
Neither of those are markets, and the fact that dumb money exists in markets doesn’t change the fact that they are rational on balance because smart money will eat the dumb money fairly quickly.
The number of people who have told me "because betting markets have Trump at X% in 2024, that's the odds that he wins" in the past six months is staggering. I think some distribution of the idea that betting markets aren't actually pure likelihood estimators is a good thing.
I can't quite recall since it was a while ago but I built a fairly large USDC position in it over 1 month, as much as the liquidity allowed me to. My main concern was counter party risk since in crypto lots of projects fail and didn't want a ton of capital sitting on the platform.
Is there any possibility of unwinding the distortion caused by site fees, time-value issues, and other market imperfections to recover the true odds implied by the betting markets, or is it just too complicated for that to be feasible?
Yes, but instead of getting a single number you'll get a range. Like it might be that if the probability of Biden winning the election is anywhere between 49% to 55% then it's not worth betting for him or against him. The worse the imperfections are the wider the interval gets.
Political campaigns spend billions to make their candidate win. We know that one of the most important things in order to get a candidate to win, is the perception of strength.
There's the theoretical risk of complacency, that people don't turn up to vote for your candidate because they think he will win anyway, but in practice that seems to be more than made up for with discouragement on the other side - especially on the donor's side. Voters may be willing to turn up to cast a vote of defiance, but self-interested donors will not throw money at a candidate who won't get in a position to reward them for it.
So if people are paying attention to prediction markets at all, why wouldn't a campaign pour money into it to create the perception that their candidate is doing better than they are? Betting on a statistically impossible walkover win isn't a problem if your goal is to create the impression of fanatic support anyway.
There's not that much money in these markets, so it shouldn't take all that much money to move them, in the scale of a US presidental election.
Long shot bets are almost always overpriced on PredictIt.
PredictIt specifically encourages these long shots to be over weighted thanks to their $850 risk limit in any given market. If one long-shot bettor places a max buy order for YES at 10c per share, the order book has 8500 buys at that price point. I can come along, see that what they're betting on is absolutely nuts and max out NO... but at 90c per share I can only fill 944 of those orders before I'm maxed out at my $850 risk limit. I might be supremely confident and WANT to bet $7,650 on this but I simply am not allowed to. So it takes eight more people like me to come along and plunk that money down before we're moving the needle on the order book.
In the 2020 election they somewhat mitigated this by making a dozen markets that were all basically the same question: Is Trump going to succeed at overturning this? They had markets for each state, plus combo markets they added like "Will Trump win any of AZ, GA, PA", plus the electoral college markets, the margin-of-victory markets, etc. But it was still way more effort to go in with a big chunk of change and manage spreading your bets out across all those different markets to max your $850 caps as efficiently as possible.
This is in addition to the other factors, that it is simply more FUN for people to gamble on long odds because it's like buying lottery tickets (who wouldn't want a 10:1 return?). And that, when you factor in fees, opportunity cost (I could have this money in the S&P 500), and counterparty risk (how long will my money be tied up if PredictIt gets shut down), it's really not very attractive to put money in shares in the $0.90-0.99 range unless A. you already have the cash sitting in your PredictIt balance and B. the end date of the market is known and approaching rapidly.
The one and only time I traded on PredictIt, around the beginning of the pandemic, I made an easy couple hundred bucks on this market: https://www.predictit.org/markets/detail/4683
The majority of traders seemingly looked at the headline and failed to understand the contract terms, which only required a "quarterly annualized" GDP growth of 5%, during at least one quarter of 2020. So a severe economic downturn in Q2, followed by even a modest recovery in Q3, would have made a "yes" outcome all but guaranteed.
The comments became pretty hilarious once someone pointed this out.
> The majority of traders seemingly looked at the headline and failed to understand the contract term
Nice way to blame the masses. Seems to me more like the terms contradicted the headline, thus making this a mechanism for exploiting people who assume they're making an honest bet. Seems like fraud, honestly.
Seems like fraud to spell out the terms of the bet in plain language and then enforce those terms?
The annual rate of increase in real gross domestic product (GDP) in one or more quarters prior to or including the fourth quarter of 2020, as rounded to the nearest tenth of a percent, per the advance, second or third estimate, as published subsequent to the launch of this market for any such quarter by the Bureau of Economic Analysis (BEA), shall be 5.0% or higher.
Are you seriously claiming the average person's interpretation of "Will the US economy hit 5.0%+ GDP growth by year-end 2020?" would match the "spelled out terms"?
I'm seriously claiming that if you don't read the actual terms of a legal agreement, you shouldn't wager on the outcome of that agreement. The headline statement leaves a ton of ambiguity, so any reasonable person would read the short paragraph that describes in full detail how the outcome will be measured and then bet accordingly. Caveat Emptor and all that.
No. If this was in an advertisement I'm pretty damn sure it would violate truth-in-advertising laws. Clarifications are one thing, contradictions are another. Most "reasonable" people reading it come away with an understanding of it that contradicts the fine print. Which is exactly what happened, and hence why so many people lost money. You don't get to put contradictory terms in a contract then blame it on "caveat emptor".
With any understanding of how GDP is measured / reported, there's no contradiction at all.. I'm honestly confused what you're finding so misleading about this particular bet.
Is the difference between "Will the US economy hit 5.0%+ GDP growth by year-end 2020?" and "Will the US economy hit 5.0%+ annualized GDP growth in any quarter in 2020?" really that hard to understand?
Any sane person will tell you "Will the US economy hit 5.0%+ GDP growth by year-end 2020?" means GDP at the end of 2020 will be at least 5% higher than it was at beginning of 2020 (unless there's an explicit starting point mentioned). Aside from the question of the data source, it's pretty crystal clear what it means, and that it lacks any notion of quarters or annualization. And it's very obviously different from "Will the US economy hit 5.0%+ annualized GDP growth in any quarter in 2020?", which is what the fine print is measuring.
If it was really as clear and obvious as you claim, then please explain while the exchange felt the need to add a very visible disclaimer explaining the terms?
"It" as in the actual title they used? It was damn clear, it didn't need any clarification aside from the data source and the starting point of the measurement.
"It" as in their intended title? It obviously didn't match what they wrote, hence the "clarification". Not particularly different from how politicians frequently "clarify" that by X they actually mean not-X...
Riddle me this: imagine your convince your spouse to pour all your joint savings into a $1M home because its value will hit 5%+ growth by the year's end. And then at the end of the year it turns out your home is now worth $500k, after briefly dipping down to $400k in Q3. On a scale of 1 to 10, exactly how amused do you think your spouse would be to find out you actually meant "annualized quarterly growth rounded to the nearest tenth of a percent"?
Huh, I swear Delta offered me something that would give me like 50 bucks back if it was a rainy day but I see it nowhere on the internet so maybe it was a dream!
I worked at a prediction market startup for a while.
Even completely rational prediction markets generally have a structural problem with events with likelihoods close to 1.0 or 0.0. This leads to cases where conspiracy type events have much higher likelihoods than they should, leading people to assume the market has lost all rationality when actually it hasn't.
For example, imagine if there's an event with a predicted likelihood of 1 in 50, but it should actually be 1 in 1000.
Assuming the market uses real money, then on paper someone should be able to make a roughly 2% return by betting that the event won't occur. However, if this market isn't closing for a while (for example for an election at the end of the year), then no-one is incentivised to take the bet and correct the market, because they'll get a much better risk free return just by putting their money in a term deposit and collecting the interesting. And so the market remains un-corrected. Basically prediction markets have an accuracy upper bound determined by the current risk free return.
For markets using "play" money a slightly different effect is at play. When it's not real money involved, generally the people are incentivised to try and top a leaderboard of some sort. In this case, a really effective strategy for people who currently aren't near the top of the leaderboard is to take long shot bets. If they loose, who cares they weren't gonna win anyways, but if the the bet plays off they have a lot to gain. Basically there's an asymmetry in returns that once again messes with probabilities.
> Assuming the market uses real money, then on paper someone should be able to make a roughly 2% return by betting that the event won't occur. However, if this market isn't closing for a while (for example for an election at the end of the year), then no-one is incentivised to take the bet and correct the market, because they'll get a much better risk free return just by putting their money in a term deposit and collecting the interesting. And so the market remains un-corrected. Basically prediction markets have an accuracy upper bound determined by the current risk free return.
What if the deposited USD automatically earned interest at the current risk free return (think something like USDC on Aave in crypto land)?
Yeah as far as I’m aware that’s the solution. Though then the market can’t use the deposit to make money for themselves, and needs to rely on other mechanisms like trading fees etc. So really it comes down to how much the market cares about the accuracy of long tail events.
It's more than the risk free return. It's also fees. Whatever the fee percentage is, all in, to get a dollar out of the platform and back into cash on a winning bet, plus the risk free rate over the time period, that's your upper bound. Given the comments above about fees it doesn't surprise me at all to see 5-10% mispricing.
> Even completely rational prediction markets generally have a structural problem with events with likelihoods close to 1.0 or 0.0. This leads to cases where conspiracy type events have much higher likelihoods than they should, leading people to assume the market has lost all rationality when actually it hasn't.
Is there academic research on this by any chance? Recently I've been wondering if there is a similar issue with VC funded of startups, where people overestimate the odds of a building 10B or 100B business (especially relative to a 1B or 100M business), causing them to pursue investments and strategies with lower expected value.
> Is there academic research on this by any chance
Read "Secrets of Sand Hill Road".
It's written by Scott Kupor - one of the Managing Partner at A16Z - and explains in depth how the VC model works.
It's also required reading if you want to interview at a VC fund, and highly recommended by other founders to learn how to better strategize when raising funds.
Secrets of Sand Hill Road, Monetizing Innovation, and Working Backwards are the only books you need to successfully understand how to raise and build a VC funded startup. The pricing strategy guide by I think Sequoia is also useful.
It seems like there is a lot of resistance to these kinds of markets. Both in terms of bureaucratic restrictions, but also in terms of sentiment, including here on HN.
It's not that people disagree with the market clearing price, and are motivated to take a position. It's that they don't want other people to converge on a prediction in this way, and want to ban it, or keep the stakes so small that the market doesn't attract real expertise.
It's like people are more into beliefs as decoration, rather than as something to compare and test against other people's beliefs.
Another reason to keep the stakes small is to avoid problems like the assassination market[0]. I'd rather have podunk, inaccurate prediction markets like PredictIt than one which amounts to putting massive bounties on public figures. Or incentivizing public figures to do their job badly like make a surprise change in their vote on an important bill, just to win a longshot bet.
PredictIt even used to have markets on, like, which SCOTUS justice would be the next to leave. People got paid when RBG died. That literally would've been this, if it weren't for the fact that the small investment limit kept the returns low. Realistically I don't think anybody was going to be committing assassinations just to make a thousand bucks.
The same things that keep prediction markets inaccurate keep them from causing bigger problems.
> It's that they don't want other people to converge on a prediction in this way, and want to ban it, or keep the stakes so small that the market doesn't attract real expertise.
Do you have an example? I haven't seen this sentiment anywhere, including HN.
I haven't actually seen much resistance to the existence of prediction markets. HN leans heavily toward freedom and letting consenting adults do what they want with their money on the internet.
The majority of negative sentiment I see is against the obsession with prediction markets in certain circles (rationalist community, for example). Parts of the rationalist community speak about prediction markets as a solution for everything.
> Do you have an example? I haven't seen this sentiment anywhere, including HN.
Do a Cmd+F for "stupid" on this very post. It may not be the prevailing sentiment, but it certainly occurs here. Maybe just a very vocal minority.
Predict it has something like an $800 maximum. That's simply too small to get a PhD to quit his day job and start trading on prediction markets for a living. Until the market capitalization can attract experts, it will not reflect the opinions of experts. This seems to be a bureaucratic restriction. Why aren't the big exchanges creating order books for these markets?
I find this interesting but I'm not sure I understand and I'd like to. Are you suggesting that there is a psychological aversion to prediction markets when the outcomes in question are as personal as politics? Would a similar analogy be, say, reluctance to the idea of finding one's romantic partner through a formalized process like algorithmic matchmaking?
What I have observed is: people calling these markets or their participants stupid. Then when they are told they can make money off of these supposedly stupid market participants, another made up reason is given why they won't put their money where their mouth is.
If the algorithms for finding romantic partners are similar to netflix recommendations, then I think the aversion to them is totally warranted; I don't think that's the same thing. If a closed source algorithm makes bad recommendations, I can dismiss it as useless. If I don't like the price of a prediction market, I am forced to reconcile what I know about the structure of the market, with my personal beliefs.
Anytime these markets come up in conversation or are used in a thought experiment, it reveals how few of an individual's outwardly professed beliefs they themselves take seriously. It's almost like most of the time someone says they believe something they are actually trying to manipulate anyone listening, rather than solicit criticism to refine the belief.
I'm not surprised that PredictIt performed so poorly. They take 10% of your profit, so you need a large margin of expected value to justify getting in and correcting a mispriced question.
To my knowledge, these markets are deliberately restricted. If you allowed unlimited betting, then experts would come in and eat these uninformed bettor's lunches. Except if you allowed that, then you face an ethical problem. You'd be providing a way for people to profit from influencing outcomes (e.g. questions that could resolve based on someone's death).
The answer is stupidly simple but only touched on briefly in the article. It is the fees and the limits bet sizes combined.
Let’s take a simple example, which is Donald Trump‘s current odds of winning the Republican nomination. Predictit currently has this at $0.92. Realistically I think the odds are higher. Not 100% but higher than $0.92. The problem is that the maximum I can bet is $500, so that’s 543 “shares” that I can buy. And I’ll make $0.08 on each share if he wins, so that’s $43. Then when I withdraw my money I will get a 5% fee, so my initial $500 turns in to $516. Oh, and then I have to wait until primaries, and for it to actually pay out. And there IS a non zero chance trump loses. Very few individuals will care about turning $500 into $515, so this bet makes sense. And since you have a cap on bet size, no institution will care either.
> Unless we make sure that market structures are providing the right incentives in the right way, we shouldn’t be surprised if prediction markets continue to struggle.
This conclusion is incredibly vague. What are "the right incentives in the right way"? Is there anything concrete that the author proposes here?
If you have some deep subject knowledge to make a prediction, your bet will be drowned by others. Sure, you can make a buck from your prediction, but you won't get instantly rich, and you won't be able to make these kinds of bets consistently.
People with subject knowledge can make money, but no mechanism guarantees that the majority of people using the market are subject matter experts. If anything, most people using the prediction market are going to be people paying to make their chosen outcome look more preferable. They aren't trying to make a financial payoff on the bet, they're trying to make an emotional payoff and are willing to lose money doing so.
This is the best case scenario for prediction markets I think, as it means the odds more accurately reflect the outcome likelihood (e.g. people aren't betting against their favourite in an hedge against emotions).
At least then they would break even in an efficient market. But it isn't an efficient market due to fees and opportunity cost... So you're right, it is irrational to participate.
Hypothetically someone who was confident they had better knowledge would be willing to invest more in their prediction. Those who were right about that would have more money to invest on the next iteration, which would amplify this effect over time.
Realistically, to my knowledge, this isn't really possible on current platforms since they have built in limitations and fees.
Otherwise, the profit-share demolished gains as a de-facto trading fee. Plus you could only make money on a few hundred at a time. I did some of the cheap bets that were mispriced but also some more obvious ones. A few multiples on the money, but it was too unwieldy to play it. The politics ones are easy. People like to bet on the outcome they'd like to occur rather than the one they think will occur.
I understand, though. I find it hard to bet against Arsenal. Just seems like a miserable way to live.
An important aspect of this is that it's not legal for most people to run a prediction market. PredictIt is allowed to operate because it is run by a university, and even that is now under question. The result is that there is little competition in this space, which is why PredictIt has huge fees and limits. As others have pointed out, that's a big part of why these markets deviate from reality.
I think the better title here is Prediction markets have a regulation problem. Out of the prediction markets you mentioned only one is regulated: Kalshi and only one is unable to list election markets: Kalshi. With their current lawsuit against the CFTC I am hopeful regulated election markets will be a reality very soon
That was extremely common with print magazines when those still existed in large numbers. Maybe still is common but it's been at least a decade since I last purchased one.
It would be nice to have a site like electionbettingodds.com but with intervals rather than point preditions.
Given the betting odds it's possible to calculate upper and lower bounds on the probabilities. Ideally you would take into account the fees and the opportunity cost of having your money locked in the bet. The bounds would be the range in which it wouldn't be possible to make an expected profit by betting either way.
The advantage would be that instead of an absurd point estimate like "Michelle Obama has a 6.90% of winning the election" you'd get something like "the probability of Michelle Obama winning the election is between 7.0% and 0.01%". Which makes it clear that while the upper bound can't be driven to 0, the lower bound is perfectly sensible.
> instead of an absurd point estimate like "Michelle Obama has a 6.90% of winning the election"
A probability distribution is defined over a set of outcomes. If “Michelle Obama winning” is one independent, categorical outcome, then there is one probability associated with that outcome.
> you'd get something like "the probability of Michelle Obama winning the election is between 7.0% and 0.01%"
Stacking another layer of probability on top of probability distributions still results in only one final distribution for prediction purposes.
I think maybe what you’re trying to capture is how your “confidence” about a prediction ties into the evaluation of the reward or penalty. This would just be modulated by adjusting how wide/uniform your probability distribution is and the amount that you bet, but perhaps the websites could add some tools to help people visualize this better.
Yes, I agree that in principle every person should have a probability for each event. But what does the market tell us about what our own belief should be? If you don't have any nonpublic information and you trust the market to be efficient then your assumption should be that you yourself cannot make an expected profit in the market. So the useful information from the market can be given in terms of the bounds within which you can't bet profitably. You yourself should update your probabilities to be within those bounds.
Again, the assumption of rationality, along with the lack of appreciation for the diminishing value of money from, e.g. whales is the silly part here.
While I'd concede that "they're stupid" is probably the primary reason someone might go for the Michelle Obama bet, I can think of AT LEAST one other -- e.g, I'm insanely rich Tyler Perry, and I just like the idea of that being in the air?
I thought they were going to talk about how when a prediction market gets too large, it can actually cause candidates to throw an election. Like assassination markets can affect the external world too
Another thing I don't think the article touched upon - it makes sense to use these markets as a hedge for things you're hoping won't happen, especially as prediction markets become less of a niche thing. For example, it would make sense to put down a bunch of money on Trump winning in 2024. Not because you'd be happy if he won, but rather so that if he does win your earnings can mitigate damage elsewhere in your life. Of course you might not want that contract to be denominated in USD...
A predictions market is one of those things that sounds fine on paper and works terribly in practice. The crucial failure is the inability to separate the incentives created by the prediction from the event being predicted. The tail ends up wagging the dog.
So despite the "advocates claim that by providing a marketplace for bets on uncertain events, prediction markets give predictors a financial incentive to be correct", what happens is that prediction markets give predictors a financial incentive to make it be correct, which leads to all sorts of shenanigans in practice. Look at the stock market, which largely is a predictions market, and look at how much regulatory machinery is needed to dissuade obvious scams like insider trading and pump & dumps. And even then, the stock market is highly irrational precisely because of the financial incentive; look at how many years it's taken the GME fiasco to shake out and how many true believers are not merely still holding the bag, but buying up as many bags as they can.
The ultimate example that I always trot out is that a predictions market on the topic of "will $PROMINENT_POLITICIAN be dead by $DATE" is just a distributed assassination contract (to be fair, it's also a distributed bodyguard contract...). Welcome to the free market, I hope your friends are richer than your enemies!
Haralabob Voulgaris is a pretty well known NBA (basketball) gambler. He has talked about how his biggest lifetime score ever was betting Biden after it was clear Biden was going to win but before the race was called.
He talked about it on Bill Simmons Podcast and was just incredulous at how people could be that stupid and how much money had made betting on something that was a stone cold lock.
The example the author gives is the 2020 election. Without rehashing anything or endorsing anyone’s viewpoint, there were lots of issues being raised, lots of court cases in flight, and possible procedural issues in the states and in the Congress related to electors and certification.
The election wasn’t “over” until certification happened (late) on January 6.
Why should bettors be prohibited from continuing to bet?
And worse, if bets were resolved too early, then chaos would have ensued in the betting market if the electors were not certified and the House selected Donald Trump.
Many ppl here may be too young to remember 2000 where recounts kept going on and on in Florida and the Supreme Court intervened.
Again, I’m not advocating any position, I’m just pointing out that the election isn’t over just because $NEWS_SOURCE declares a winner.
Elections should be perceived to be conducted in a fair and impartial manner as well as actually be conducted fairly and impartially. Therefore, betting on elections should be banned because an incentive is created to rig the outcome.
Elections are about power; there could hardly be a bigger motivator to rig the outcome even in the absence of any external factor like betting markets.
Please, please, please keep advocating for this. I made hundreds of dollars in 2020, all of it after the election was called, buying Biden shares on states that had already been called at like 80 cents on the dollar, all of it from suckers who thought Rudy Giuliani’s court cases were legitimate. I’d love to do it again if Biden wins in 2024.
> But instead, the majority of the bettors against Biden bet on the astonishing outcome of “Trump wins by 280+ electoral votes.” That, it should be clear, was impossible — it would require results to be overturned not just in states with tight races but also in Democratic bellwethers like California.
That was not impossible at all once you take into account Trump's VP certification plan, which many people understood very early on. A longshot, yes, but certainly not impossible.
If anything, the prediction markets priced in an insurrection better than any pundit I recall watching. To flip it around and call prediction markets "irrational" in this situation is misguided and ignorant of political history.
Happy to see that the article gets there eventually, but it's buried pretty deep: People betting on Donald Trump to win the election he had just lost weren't actually betting money. They were spending it.
They are saying "I will spend $,000 just to show that I support Trump to the point where I [pretend to] believe him when he says he won the election [that he clearly just lost]". It's MAGA culture's way of - wait for it - virtue signaling.
And the prediction market is just vacuuming up most of that money for themselves. You can bet against it for a nominal profit, but you're paying the vig to do it, and the vig is a big chunk when the odds are sitting at 9:1.
I remember in 2020 election markets, from ~mid 2019 onward you could take whether hillary clinton would run in 2020. It was a funny market until she started missing filing deadlines and something weird happened - the line became MORE favorable towards hillary running. It went on like that for quite a long time, well past the point where anyone with a room temp IQ would figure out she had zero chance of running because she wasnt running.
Even up to about a week before the actual 2020 election, she was still trading at almost 10 cents on the dollar for “yes she is running.” Just wild. I don’t think it’s a problem for these markets, just the nature of the beast being that people easily believe conspiracy theories and propaganda and if you’re a gambler there’s a lot of money to make.
I think for now they are still pretty good at assessing something like "Are the odds that Trump wins Wisconsin in 2024 closer to 2/3 or 1/2". The post-election stuff was annoying (I had several thousand dollars on the site) but the market limits + uncertainty over the platform made is understandable.
It would be nice if there was a platform that didn't have limits and had interest rates closer to savings interest rates.
There was a non zero chance that the GOP would block the result and keep Trump in power. That is why you could still bet on that outcome after the voting results were finalized.
I'm going to talk about political gambling but I'm not going to talk about PredictIt.
PredicitIt is irrelevant to the conversation. It's high fees an miniscule limits completely distorts what is "rational" or "reasonable" to expect the market to do.
Betfair on the other hand is a reasonable thing to talk about. 2% fee on net winnings on a market and zero deposit or withdrawl fees makes it a pretty efficeint market for political betting.
The Betfair political markets are dumb as a bag of rocks and there simply isn't enough smart money to suck up the dumb money. Political markets are incredibly thin, the average India Premier League cricket match gets way, way more action than all but the very largest political markets. The 2020 Presidential election was fairly unique in that the majority of the money was wagered after the election happened and after it was clear Biden had even. Idiot Trump supporters buying into conspiracy theories flooded the market with money.
Political betting is very easy _but_ compared to sports gambling there is no where near enough events to make a living doing it. It's all very well having a huge edge (and betting on a even that has already completed is the largest edge you could ever have) but if you only get 1 betting opportunity a year that's not a way to make a living. It is a great way to have some fun though.
Even in 2016, two days after the election, Clinton was @1.05 to win the popular vote. Right now on Betfair, despite almost £6 million having been traded on the market Michelle Obama is the 2nd favourite for the Democratic nomination @9.2 (that's a better than 10% chance). (For comparison about market size currently the Comilla Victorians vs Rangpur Riders match has over 4.5 million wagered).
This isn't an awful breakdown, it's just a bit pearl-clutching about several problems that are well understood.
First, let's clear up some confusion.
Markets are not accurate ways to determine accurate predictions. Somebody, somewhere is getting confused with "Wisdom of the Crowds" thinking.
In the "Wisdom of the Crowds" - take, for example, everyone at the State fair taking a guess at how heavy a bull is - everyone has a single guess (and only one guess), and on aggregate you'll probably (that word is doing a lot of work here), get a more accurate answer than any one answer probably (again, lot of work there), will. You're reducing a distribution of guesses into a single number, and the distribution of those guesses is likely uniform, and likely going to find an average (and a median), quite close to the true value.
Markets are not that. They are not efficient, they aren't accurate predictions. They are about weight of money. One moron who is hellbent on a World view with $10m in his account goes up against 1000 people with $10 each in their account, is going to move the number in their favour.
We see this irrationality every day in sports trading, stocks, commodities, options, FX, crypto, you name it.
In fact irrationality is so common in markets because of the weight of money there is a famous saying (often misattributed to Keynes):
"The market can remain irrational longer than you can remain solvent."
Time and time again its come true. All you need to do is look into the LTCM fiasco where Nobel Prize winners in economics (literally the creators of the Black-Scholes model), took a rational view of the market, bet on it big, and almost took down Wall Street in its entirety. They weren't wrong, they were just on the wrong side of the weight of money.
I spend a lot of time trading sports (gambling is legal and tax free where I live, and well regulated betting exchanges exist), and thinking a lot about value. What does it mean to say a price about a selection in a market is wrong when it comes to some men running a field after a ball? There's a lot of people interested in this problem, and there is some evidence that they're not awful at it (Vegas lines on NFL games for example, adjusted to remove the house edge, converted into probabilities, aren't far away from observed outcomes), but every day awful prices are put up all over the place and still taken as if value by morons.
Is it any surprise that newly-minted crypto-millionaires hold a lot of money in a market and can throw the price off? Not really.
What you should do about it? Bet against. Take the action. Understand Kelly criterion, do the math, and trade that. You will bring the market back to reasonable pricing through applying your own weight of money and make a profit while doing so.
Why aren't others doing so? Who knows. Liquidity, being busy with something else, not wanting to, there's a multitude of reasons. But when you spot value - somebody offering you money on the table betting for things that are literally impossible - I suggest you stop complaining and fill your boots. Morons are born every minute, but it's not often enough a lucky millionaire one is prepared to hand you their money on a plate.
If media pundits are looking at these things as predictors of something, well you know there is one less media pundit to trust on this topic as they clearly have got confused. The next time they spout something that sounds statistical or mathematical or based in science, question if it actually is. Or if they, like many others before them, have confused markets for efficient proxies to uncover genuine wisdom, through the use of weight of money as the best means to extract secrets from the gods.
I remember Nate Silver speaking of the golden age of online poker, when amateurs were the dominant audience, and a decent (non-pro) strategy player could make a living. Word got out, and just like that, it was over.
This is something a number of people do on PredictIt, since that one does have a conservative lean. It takes a lot more than the individual limit to fully price correct, though.
I personally haven't maxed out on it (I think the limit is $800), but I've got a few hundred bet that Biden will be the Dem nominee / that Newsome won't be.
Not sure why we don't have more people picking up the free money, but given that I haven't picked up a much as I possibly can, I'm not one to criticize.
sure, and thanks for agreeing with my premise! but liberals are not really into stupid gambling games, so this "market" is inherently stocked with a limited supply of mindshare
As somebody who made a decent chunk of change (enough to get a 1099 from PredictIt in 2020 and 2021) off the situation... long shot bets are almost always overpriced on PredictIt.
PredictIt specifically encourages these long shots to be over weighted thanks to their $850 risk limit in any given market. If one long-shot bettor places a max buy order for YES at 10c per share, the order book has 8500 buys at that price point. I can come along, see that what they're betting on is absolutely nuts and max out NO... but at 90c per share I can only fill 944 of those orders before I'm maxed out at my $850 risk limit. I might be supremely confident and WANT to bet $7,650 on this but I simply am not allowed to. So it takes eight more people like me to come along and plunk that money down before we're moving the needle on the order book.
In the 2020 election they somewhat mitigated this by making a dozen markets that were all basically the same question: Is Trump going to succeed at overturning this? They had markets for each state, plus combo markets they added like "Will Trump win any of AZ, GA, PA", plus the electoral college markets, the margin-of-victory markets, etc. But it was still way more effort to go in with a big chunk of change and manage spreading your bets out across all those different markets to max your $850 caps as efficiently as possible.
This is in addition to the other factors, that it is simply more FUN for people to gamble on long odds because it's like buying lottery tickets (who wouldn't want a 10:1 return?). And that, when you factor in fees, opportunity cost (I could have this money in the S&P 500), and counterparty risk (how long will my money be tied up if PredictIt gets shut down), it's really not very attractive to put money in shares in the $0.90-0.99 range unless A. you already have the cash sitting in your PredictIt balance and B. the end date of the market is known and approaching rapidly.
No conspiracy thinking required. Biden has a non-zero chance of dying before the election and Harris is very unpopular. It's not unthinkable that should Biden not be able to run, someone like Obama would step in at the last minute.
Would I pay 20:1 odds for that scenario? Hell no. But I might buy at 100:1 or greater.
I'm sorry, what? The line of presidential succession has been long established, it's in the constitution, and Kamala Harris is next in line. And then a lot of people who aren't Obama.
Lines of succession are a separate thing from winners of elections. If Biden were to die before election day, that doesn't mean Harris would be on the ticket as the presidential nominee. I'd wager against that for sure.
So who would be tapped to fill in for the recently-deceased candidate? Could be Harris, or Obama, or a dozen other hopefuls.
That's a different issue. Presidental succession says who replaces Biden if he dies in office, but the question here is who would stand as the Democratic candidate for the 2024 election if he died.
Harris would be likely, but maybe not a 100% certainty.
In the minds of conservatives and some leftists, the concept that Kamala being unpopular is so ingrained that they cannot believe that she would be the nominee no matter what the circumstances. In reality she's not much more unpopular than her President and would likely see a massive approval bump if the worst happens to the current POTUS.
Even without that, she's the odds-on favorite to be the nominee in 2028 by incumbency alone. Basically anyone that follows politics understands this, but the silly conspiracies are based on motivated reasoning.
As many mentioned here, the fees are high. That contributes to market inefficiency for sure.
Arbitrage opportunities pop up there all the time, but ones that are profitable after the fees less so. They do happen though.
He is neglecting how long contracts take to settle. There’s a lag of two months between when the election and when president is sworn in. It’s not unreasonable to sell your winning position at 98% just to be able to bet on more things.
It was the first time in 20 years of hoping for such an opportunity that I managed to get to bet on something that had already happened. That’s every gambler’s dream and I’m unaware of many instances of it happening without breaking laws or trickery. But even then, you bet 93 cents to win 7 and then pay 5% on cashout so it wasn’t exactly making people rich.