One of the most chaotic and corrupted prediction market resolutions happened earlier this week on a $7m market on Polymarket, with the aid of an UMA dispute.
Multiple bulwarks to prevent something like this from happening totally failed.
Let's call it a “Home Alone Attack.”
The attackers waited for the adults to leave, and moved in.
But in this version of the story, the idiot Wet Bandits bypassed all the booby traps and robbed the house. The profit was ~$200k for Harry & Marv, plus additional money for the rest of the crew.
Long thread⌛
Multiple bulwarks to prevent something like this from happening totally failed.
Let's call it a “Home Alone Attack.”
The attackers waited for the adults to leave, and moved in.
But in this version of the story, the idiot Wet Bandits bypassed all the booby traps and robbed the house. The profit was ~$200k for Harry & Marv, plus additional money for the rest of the crew.
Long thread⌛
The market in question is about a mineral deal between the US and Ukraine. Market pictured below.
To state the obvious, there is not yet a mineral deal, and the market suddenly expiring for yes on a random March weekend is a very perplexing outcome.
Prior to the market being expired, the deal had previously reached a high of 97% odds in late February, only for the mineral deal to fall apart in the Oval Office when JD Vance famously insisted that Zelensky wasn't thanking the US profusely enough.
Recently, it traded below 10% as April approached.
But the holders of Yes were forming a plan to try to push it to Yes regardless of whether a deal materialized in the final week.
To state the obvious, there is not yet a mineral deal, and the market suddenly expiring for yes on a random March weekend is a very perplexing outcome.
Prior to the market being expired, the deal had previously reached a high of 97% odds in late February, only for the mineral deal to fall apart in the Oval Office when JD Vance famously insisted that Zelensky wasn't thanking the US profusely enough.
Recently, it traded below 10% as April approached.
But the holders of Yes were forming a plan to try to push it to Yes regardless of whether a deal materialized in the final week.
On Saturday evening, March 22nd, the largest Yes holder pulled up to the market, trading around 15%, and proposed Yes as the outcome -- at a seemingly random time.
But it was not random. It was done intentionally, and planned in advance.
The user, 50 Pence, in coordination with another Yes holder, timed it to when Polymarket employees LEFT THE OFFICE and the site was unmonitored. This would make it difficult for Polymarket to clarify the market in the dispute.
They also simultaneously timed it to just before UMA's dispute process rolls over their day (just before 8pm), greatly minimizing the reaction time for users to post arguments either for or against.
With the parents gone and UMA's process tilted in their favor, they proposed a resolution to attempt to rob the house.
They then set about spamming a very spurrious argument into the UMA evidence thread on Discord (many of the Yes posts were made within a few seconds of the thread going up).
The argument? The market, trading at 15%, has been Yes since February.
But it was not random. It was done intentionally, and planned in advance.
The user, 50 Pence, in coordination with another Yes holder, timed it to when Polymarket employees LEFT THE OFFICE and the site was unmonitored. This would make it difficult for Polymarket to clarify the market in the dispute.
They also simultaneously timed it to just before UMA's dispute process rolls over their day (just before 8pm), greatly minimizing the reaction time for users to post arguments either for or against.
With the parents gone and UMA's process tilted in their favor, they proposed a resolution to attempt to rob the house.
They then set about spamming a very spurrious argument into the UMA evidence thread on Discord (many of the Yes posts were made within a few seconds of the thread going up).
The argument? The market, trading at 15%, has been Yes since February.
The argument from Yes holders was - a bit unbelievably - that the market has been Yes since before the Vance dress down in the Oval!
They cited, as their primary piece of evidence, a Fox Business interview from February 27th where Treasury Secretary Bessent said that Zelensky was about to sign a deal that he (Bessent) had negotiated with Ukraine.
[To be clear, the market does NOT require signatures, only an agreed upon deal]
This WOULD be a very compelling argument to expire the market on Feb 27th or on Feb 28th. The argument being that - despite the collapse of the deal and the lack of a formal signature - the countries had first agreed upon it. You can then argue it was moot that it subsequently fell apart.
But to make that argument nearly 30 days later in late March, with 0 new information or sources? It's totally bonkers.
And an immediate and fatal problem arises for that argument: UMA had ALREADY DECIDED that the "agreement" on the 27th didn't count! The market for a deal by February 28th, with the exact same rules, already expired to No via UMA!
The Yes voters were asking UMA to vote against their own oracle with 0 new pieces of information.
As you can imagine, MOST users were arguing that this should not count, but their posts were buried near the end of the thread.
They cited, as their primary piece of evidence, a Fox Business interview from February 27th where Treasury Secretary Bessent said that Zelensky was about to sign a deal that he (Bessent) had negotiated with Ukraine.
[To be clear, the market does NOT require signatures, only an agreed upon deal]
This WOULD be a very compelling argument to expire the market on Feb 27th or on Feb 28th. The argument being that - despite the collapse of the deal and the lack of a formal signature - the countries had first agreed upon it. You can then argue it was moot that it subsequently fell apart.
But to make that argument nearly 30 days later in late March, with 0 new information or sources? It's totally bonkers.
And an immediate and fatal problem arises for that argument: UMA had ALREADY DECIDED that the "agreement" on the 27th didn't count! The market for a deal by February 28th, with the exact same rules, already expired to No via UMA!
The Yes voters were asking UMA to vote against their own oracle with 0 new pieces of information.
As you can imagine, MOST users were arguing that this should not count, but their posts were buried near the end of the thread.
These days, most disputes on Polymarket are decided unilaterally by Polymarket's clarifications. A dispute is made, Polymarket clarifies "XYZ happened, and these are the rules, and based on our interpretation of the rules, it should be Yes/No/Too Early."
UMA has never voted against a Polymarket clarification (they got close a couple times). UMA voters don't put in a lot of due diligence or work, so if Polymarket says it should be Yes/No/Too Early, they don't need to do any work at all. Easy. Polymarket decides most contentious questions by fiat, for better or worse.
Why is this done by Polymarket? Theoretically, Polymarket wrote the rules and knows their intent so the rules writers are the best placed to determine what they intended with the market (these clarifications can cause a lot of consternation, but that's not the focus of this thread).
50 Pence reasoned that Polymarket would decree that the February 27th "deal" was not enough for the market to expire a month later, and he wanted to avoid them weighing in. He waited for the offices to be vacant, and snapped into action.
UMA has never voted against a Polymarket clarification (they got close a couple times). UMA voters don't put in a lot of due diligence or work, so if Polymarket says it should be Yes/No/Too Early, they don't need to do any work at all. Easy. Polymarket decides most contentious questions by fiat, for better or worse.
Why is this done by Polymarket? Theoretically, Polymarket wrote the rules and knows their intent so the rules writers are the best placed to determine what they intended with the market (these clarifications can cause a lot of consternation, but that's not the focus of this thread).
50 Pence reasoned that Polymarket would decree that the February 27th "deal" was not enough for the market to expire a month later, and he wanted to avoid them weighing in. He waited for the offices to be vacant, and snapped into action.
But in a plot twist, on Sunday night, 23 hours and 58 minutes into the UMA voting, with 2 minutes remaining before voting ended, Polymarket issued a clarification to the market. They said that Yes should not prevail. That voters should vote "Too Early" and let the rest of the month play out to see if an actual deal is reached.
The price immediately crashed into the mid-single digits for Yes, with the thinking that UMA will now either definitely vote "Too Early" or Polymarket will otherwise prevent Yes from winning.
The statement was unequivocal: "It is too early to resolve the market" and the nature of this statement could (and did) lead to some users thinking it was a guaranteed outcome.
But given that it was made 2 minutes before voting ended, it is ~impossible for voters to switch their vote, and so some Polymarket users began loading up on cheap yes lottos hoping for UMA voters to just ignore Polymarket or not notice or not care.
The price immediately crashed into the mid-single digits for Yes, with the thinking that UMA will now either definitely vote "Too Early" or Polymarket will otherwise prevent Yes from winning.
The statement was unequivocal: "It is too early to resolve the market" and the nature of this statement could (and did) lead to some users thinking it was a guaranteed outcome.
But given that it was made 2 minutes before voting ended, it is ~impossible for voters to switch their vote, and so some Polymarket users began loading up on cheap yes lottos hoping for UMA voters to just ignore Polymarket or not notice or not care.
Here is where things go from very stupid to inexplicably stupid.
UMA voting is a two-step process. You cast your vote in one 24-hour period, and reveal your vote in the second 24-hour period.
At this juncture, UMA voters have the choice between revealing their votes in this market, or not revealing their votes.
Either UMA or Polymarket could have very easily insisted that a multi-million dollar market on a high-profile topic not be resolved yet, and have a re-vote. The risk of revealing is that UMA would overturn a Polymarket clarification for the first time (and in a very embarrassing manner).
A very simple solution. But no one at Polymarket or UMA asked UMA voters to not reveal their votes.
Instead of making a rational choice, Polymarket made a reckless one and decided to just let it play out. With 2 minutes left in the reveal period, and the vote now just barely headed to Yes with 54% of the vote, Polymarket makes another clarification that essentially says "Nevermind, just kidding with that clarification from yesterday."
The market then goes to nearly 99c, and expires a few minutes later for Yes. Fortunes are made and lost in the blink of a clarification. Users who assumed that Polymarket wouldn't let the market expire to Yes were liquidated.
UMA voting is a two-step process. You cast your vote in one 24-hour period, and reveal your vote in the second 24-hour period.
At this juncture, UMA voters have the choice between revealing their votes in this market, or not revealing their votes.
Either UMA or Polymarket could have very easily insisted that a multi-million dollar market on a high-profile topic not be resolved yet, and have a re-vote. The risk of revealing is that UMA would overturn a Polymarket clarification for the first time (and in a very embarrassing manner).
A very simple solution. But no one at Polymarket or UMA asked UMA voters to not reveal their votes.
Instead of making a rational choice, Polymarket made a reckless one and decided to just let it play out. With 2 minutes left in the reveal period, and the vote now just barely headed to Yes with 54% of the vote, Polymarket makes another clarification that essentially says "Nevermind, just kidding with that clarification from yesterday."
The market then goes to nearly 99c, and expires a few minutes later for Yes. Fortunes are made and lost in the blink of a clarification. Users who assumed that Polymarket wouldn't let the market expire to Yes were liquidated.
The problem then became compounded, as UMA was now overruling themselves and declaring the February 27th events as an oracle-verified agreement.
This led to a domino effect where other markets now had to be expired as well, including a contentious $1.5m market on whether Ukraine would agree to pay back US aid.
That also went to Yes without Ukraine agreeing to paying back US aid. Fruit of a rotten tree.
This led to a domino effect where other markets now had to be expired as well, including a contentious $1.5m market on whether Ukraine would agree to pay back US aid.
That also went to Yes without Ukraine agreeing to paying back US aid. Fruit of a rotten tree.
And how about the largest profiter in this market?
It's our old friend, and the most prolific scam artist in prediction market history, aenews2. He took home $55,000 in profit.
aenews, if you don't know or have forgotten, has gotten caught multiple times undertaking sophisticated scams in previous years. I've written about just a small subset of these in previous threads, including him pretending to be in a dockworkers union for months. Or pretending to see the movie House of Gucci 3,000 time in order to rig a review market. He is, after all the well-documented scamming, currently one of the most trusted members of the UMA "community."
Since we last wrote about him scamming the UMA system for personal gain, he has now added more titles and gained more trust within UMA.
He is at the pinnacle of trusted users on UMA, both a "Verifier" and a "Truthor." I'm surprised UMA hasn't hired him yet.lo
His argument for Yes, as stupid as it was, was posted to the evidence thread within 15 seconds to make sure it was among the first to be read. A voter reading the thread would see that a highly trusted user was telling them to vote Yes and possibly/probably stopped reading the thread at that point.
His partner, 50-Pence, for his part leading this campaign, almost tripled that and made around $140,000 in profit. It was a paper gain -- he went from losing $190k in the market to down $50k.
It's our old friend, and the most prolific scam artist in prediction market history, aenews2. He took home $55,000 in profit.
aenews, if you don't know or have forgotten, has gotten caught multiple times undertaking sophisticated scams in previous years. I've written about just a small subset of these in previous threads, including him pretending to be in a dockworkers union for months. Or pretending to see the movie House of Gucci 3,000 time in order to rig a review market. He is, after all the well-documented scamming, currently one of the most trusted members of the UMA "community."
Since we last wrote about him scamming the UMA system for personal gain, he has now added more titles and gained more trust within UMA.
He is at the pinnacle of trusted users on UMA, both a "Verifier" and a "Truthor." I'm surprised UMA hasn't hired him yet.lo
His argument for Yes, as stupid as it was, was posted to the evidence thread within 15 seconds to make sure it was among the first to be read. A voter reading the thread would see that a highly trusted user was telling them to vote Yes and possibly/probably stopped reading the thread at that point.
His partner, 50-Pence, for his part leading this campaign, almost tripled that and made around $140,000 in profit. It was a paper gain -- he went from losing $190k in the market to down $50k.
In reaction to the market resolving against reality, Polymarket made a system-wide announcement that the outcome was unfortunate, but that it "wasn't a market failure" despite it being a very large market failure.
Even forgetting the users who lost money from correctly predicting the outcome, and forgetting users who lost money from their botched clarification, they now have a market from February 28th that resolved "No" for a deal, and a market from late March that resolved "Yes" for a deal purely based on information that was widely known on February 27th.
Both UMA and Polymarket now have simultaneously declared that a deal was AND was not agreed to on February 27th.
UMA's spokesman Benni called it "a series of unfortunate events." UMA otherwise hasn't released any information on their checked-out voters being duped and catastrophically misresolving a market. Questions about why aenews continues to be trusted go unanswered. A post where I said "Yikes, you just voted against your own oracle" was deleted by a moderator. Double yikes.
But don't worry! We are assured that there are secret war rooms where they are definitely working on a solution. They've got their best people on it. Well, scratch that, their best people are deleting messages on discord, but their SECOND best people are on it.
Even forgetting the users who lost money from correctly predicting the outcome, and forgetting users who lost money from their botched clarification, they now have a market from February 28th that resolved "No" for a deal, and a market from late March that resolved "Yes" for a deal purely based on information that was widely known on February 27th.
Both UMA and Polymarket now have simultaneously declared that a deal was AND was not agreed to on February 27th.
UMA's spokesman Benni called it "a series of unfortunate events." UMA otherwise hasn't released any information on their checked-out voters being duped and catastrophically misresolving a market. Questions about why aenews continues to be trusted go unanswered. A post where I said "Yikes, you just voted against your own oracle" was deleted by a moderator. Double yikes.
But don't worry! We are assured that there are secret war rooms where they are definitely working on a solution. They've got their best people on it. Well, scratch that, their best people are deleting messages on discord, but their SECOND best people are on it.
My solution? Well there's a lot of things that need fixed here, but the main solution is pretty boring, and it's the same as it was a year ago, two years ago, three years ago, four years ago.
And actually the same solution that I've been posting about since before Polymarket existed and one of the reasons I even joined Twitter in the first place: stop empowering repeated scam artists to rip your users off in prediction markets.
Prediction markets are peer to peer information markets. If you can find compelling ways to manipulate information that trick even a single person, you can profit.
For instance, PredictIt did nothing when people set up scam businesses in their comment section pumping and dumping newbies with fake information, fake tweets, etc. Stuff that is blatantly illegal to do in the stock market, in fact prosecuted on a regular basis, was an everyday occurrence on PredictIt. PI didn't even pretend to lift a finger to stop it.
But whereas scams on PI would net you maybe a few hundred bucks, the supercharged nature of Polymarket turns that into the tens of thousands. And worst of all from a business perspective: almost all of those losses will be from your newest and most unsophisticated users, the exact user you most want to attract and protect in order to grow.
The Home Alone Attack succeeded here with an absolutely braindead argument that had no business getting close to succeeding. How? Just by virtue of watching the house and noticing when the parents left.
They then (just barely) manipulated UMA to vote against their oracle, a manipulation that UMA had facilitated them to do by labeling them as trusted. And Polymarket sat by and watched it happen.
Prediction markets are a nascent and growing space. They're also lawless. It's the Wild West minus the murder. Users can essentially commit what would be market manipulation/fraud with no fear of any consequence. In fact, not only no fear, but both monetary and sometimes reputational rewards. I've noticed the newest class of users venerates ripping other people off. Not a good place for prediction markets to be headed.
There will obviously always be people trying to make money through unethical or fraudulent means. But helping them do it is really beyond the pale. And not doing anything to fix it at any point in time is a ridiculous dereliction. Nobody at Polymarket or UMA lost money here. Their customers lost the money. Two bumbling fools, with the assistance of UMA and the carelessness of Polymarket, made out like bandits.
And actually the same solution that I've been posting about since before Polymarket existed and one of the reasons I even joined Twitter in the first place: stop empowering repeated scam artists to rip your users off in prediction markets.
Prediction markets are peer to peer information markets. If you can find compelling ways to manipulate information that trick even a single person, you can profit.
For instance, PredictIt did nothing when people set up scam businesses in their comment section pumping and dumping newbies with fake information, fake tweets, etc. Stuff that is blatantly illegal to do in the stock market, in fact prosecuted on a regular basis, was an everyday occurrence on PredictIt. PI didn't even pretend to lift a finger to stop it.
But whereas scams on PI would net you maybe a few hundred bucks, the supercharged nature of Polymarket turns that into the tens of thousands. And worst of all from a business perspective: almost all of those losses will be from your newest and most unsophisticated users, the exact user you most want to attract and protect in order to grow.
The Home Alone Attack succeeded here with an absolutely braindead argument that had no business getting close to succeeding. How? Just by virtue of watching the house and noticing when the parents left.
They then (just barely) manipulated UMA to vote against their oracle, a manipulation that UMA had facilitated them to do by labeling them as trusted. And Polymarket sat by and watched it happen.
Prediction markets are a nascent and growing space. They're also lawless. It's the Wild West minus the murder. Users can essentially commit what would be market manipulation/fraud with no fear of any consequence. In fact, not only no fear, but both monetary and sometimes reputational rewards. I've noticed the newest class of users venerates ripping other people off. Not a good place for prediction markets to be headed.
There will obviously always be people trying to make money through unethical or fraudulent means. But helping them do it is really beyond the pale. And not doing anything to fix it at any point in time is a ridiculous dereliction. Nobody at Polymarket or UMA lost money here. Their customers lost the money. Two bumbling fools, with the assistance of UMA and the carelessness of Polymarket, made out like bandits.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
