Jason Pipkin (@jipkin) header
Jason Pipkin (@jipkin) profile picture

science, but mainly politics

Joined February 2011
Photo by Jason Pipkin (@jipkin)Photo by Jason Pipkin (@jipkin)Photo by Jason Pipkin (@jipkin)Photo by Jason Pipkin (@jipkin)Photo by Jason Pipkin (@jipkin)Photo by Jason Pipkin (@jipkin)Photo by Jason Pipkin (@jipkin)Photo by Jason Pipkin (@jipkin)Photo by Jason Pipkin (@jipkin)Photo by Jason Pipkin (@jipkin)
Jason Pipkin retweeted
9
26
10
300
86,128
Jason Pipkin retweeted
On Insider Trading. Some say insider information can make prediction markets more accurate. But the same argument can be made for stock markets, where insider trading is banned. Insider trading erodes trust. When people believe a market is unfair, they stop trading. Liquidity dries up, volume collapses, and the market dies. Also, allowing it could incentivize bad actors to leak information they shouldn’t. So Kalshi bans insider trading. How do we police it? Just like stock markets. We spent years building a market surveillance system similar to those used by the NYSE and Nasdaq: 1. Detect: Our surveillance system (called Poirot, after the French detective) flags suspicious patterns in real-time by running trades through pattern recognition models. Insider trades stand out because they often are weird and are bigger in size: people don’t usually commit fraud for $25. 2. Investigate: We have a market regulation team that conducts investigations. They review KYC data, funding sources, prior trading history, and trade rationale. They contact traders when needed. 3. Enforce: If we find wrongdoing, penalties range from warnings and fines, to referrals to the CFTC (and sometimes DOJ) for civil/criminal prosecution. Transparency helps. All activity on Kalshi is public. We also report it to the CFTC daily. Because of that, our traders help by sending us whistleblower tips. To those traders, I know some of you are frustrated that we don't tell you what happens after you send a tip. I get the frustration. Legally, we need to keep our investigations private – but please know your help is super valuable. In the past year, we ran over 200 investigations and froze relevant accounts. Of these, over a dozen have become active cases and several have been referred to law enforcement. Today we go further: 1. Wharton Forensics Lab: We’re bringing Wharton Forensic Analytics Lab Director Daniel Taylor, one of the premier forensics experts in the country, to advise us. He will help us investigate intricate "cousin of the spouse of the dude at Tesla" insider cases. 2. Brian Nelson: We’re adding Brian, former Under Secretary of the Treasury for Terrorism and Financial Intelligence, to counsel us on financial compliance and market integrity. 3. Surveillance Audit Committee: We’re forming an independent committee of market integrity experts like Lisa Pinheiro to provide quarterly reports to Brian and publish relevant statistics. 4. System upgrades: We're partnering with Solidus Labs. Their institutional-grade behavior monitoring and pattern recognition tools will be added to our systems. 5. Robert DeNault as Head of Enforcement: Bobby is a force of nature and a self-professed financial crime nerd. He was a white-collar criminal attorney and spent law school studying presidential investigations and the Panama Papers. He’s also a great guy - we love Bobby. All industries have bad actors and no system is perfect, Kalshi's included. But we are committed to improving daily. Lots of work ahead!
115
82
76
735
1,008,665
Just the beginning. Thank you from the bottom of my heart to the incredible Kalshi family, you made this happen ♥️
Kalshi raised $1B at an $11B valuation. A decade ago, only a few thousand people knew what a prediction market was. Eighteen months ago, most prediction markets were banned - until we overcame the government to set them free. Over the past seven years, our community has opened up an entirely new category. Today, Kalshi is trusted, used, and loved by millions of people. It’s a part of everyday culture, and it’s driving one of the most important shifts in consumer behavior in recent history. The time has finally come for prediction markets to achieve their full potential and we are intent on making that happen. To all the believers and the early adopters: thank you.
169
83
15
2,372
271,044
Kalshi raised $1B at an $11B valuation. A decade ago, only a few thousand people knew what a prediction market was. Eighteen months ago, most prediction markets were banned - until we overcame the government to set them free. Over the past seven years, our community has opened up an entirely new category. Today, Kalshi is trusted, used, and loved by millions of people. It’s a part of everyday culture, and it’s driving one of the most important shifts in consumer behavior in recent history. The time has finally come for prediction markets to achieve their full potential and we are intent on making that happen. To all the believers and the early adopters: thank you.
585
383
409
4,069
8,630,327
Jason Pipkin retweeted
We're excited to power @RobinhoodApp's prediction markets hub! Together we'll bring event trading to 25 million users - starting with this year's college basketball tournament!
59
82
61
724
1,190,903
After successfully beating the federal government in a historic court battle and legalizing prediction markets in America, we had one month to go mainstream… one month to run one of the boldest marketing campaigns of the decade. Given the time crunch, we decided to go big and launch a massive ad campaign on cable TV. We partnered with Stink Studios and MJZ — two of the best creative and production studios — to make a killer commercial. Beautifully produced, highly cinematic, top tier cast, A-List director… Hollywood-level production. We got it done in record time and were ready to go… but: Every TV outlet blocked our ad and censored it from airing for reasons unknown to us. We ended up not running our election commercial, which was extremely disappointing. I was on the verge of losing my mind: we had spent a ton on this commercial, and the board was pissed. It turned out to be a blessing in disguise. While cable has since agreed to air it, our other channels were so successful that we felt it was best to go direct via X. No more middle-men. We deliver the message straight to you. I hope you enjoy our ad for the Oscars.
37
85
31
713
714,224
Gonna be a wild ride, and hopefully a long one
Kalshi just legalized trading on elections in the U.S. For the first time in 100 years, Americans will have access to legal election markets at scale. Historic moment for financial markets.
1
16
1,972
Unfortunately, the CFTC has stayed the contract, and opened it up for public comment again... We are not going to spend a ton of time repeating an advocacy campaign for the same matter: the public was clear the first time. But we encourage you to submit a comment if you want.
1
3
1
12
11,867
New piece I wrote on what I think would be a really bankroll-efficient strategy to farm the new volume rebate program on @Kalshi: medium.com/kalshi/how-to-the…
2
8
4
37
9,976
Also, news for those that don't know: I now work for Kalshi Trading (the trading affiliate of Kalshi Inc that exists separately from Kalshi Exchange.) So not a disinterested party, as explained in the piece.
1
14
1,637
1st of March: we've launched volume incentives! Get paid up to **1.5%** of your volume depending on your tier. Yes, I did say 1.5%
2
9
2
22
17,341
White blood cells showing up to help the body fight infection and other disease nitter.cf/funniestvidslol/status… – at Queens, NY
849
66,832
4,183
421,159
14,715,833
One of the more understated risks from AI is that it comes up with medicore answers that people mistake for being brilliant.
4
1
23
if your local meteorologist tells you that there's a 40% chance of rain, and then it does rain, do we just throw our hands up and say there's no way to know whether or not the meteorologist's model worked?
Replying to @whstancil
On the other hand, was it a 55% chance of a good night? A 45% chance? Even an 80% chance? Impossible to gauge. The model guessed, but we have no way to know if it was right or not.
1
11
538, for instance, has looked at their own work using a calibration metric - do events the model thinks are 40% to happen happen 40% of the time? [Note as well that their election models produce a wealth of predictions for many many races, and not just win probabilities].
1
4
there are legit critiques to be made of their model (it produced some wonky results in individual house races), but most of these are just aimed at the particulars of its design. The underlying philosophy is sound.
1
6
who among us didn't have CO-03 as the tipping point district for house control
Democrats have had a generally great last 24 hours of news in unresolved House races, with the exception of Rep. Lauren Boebert's (R) race in #CO03. Not inconceivable she could end up barely putting Rs in the majority.
2
14
Underappreciated unlikely winner last night: @PredictIt's servers, which never (fully) crashed. In fact, trading was buttery smooth from like 9:30pm on despite a big flip underway. Gonna miss the site :(
1
1
19
There were a lot of garbage polls, but the averages did a pretty reasonable job, and as a result the betting markets (which assumed polls were biased to democrats) got effectively pantsed by 538's model. (Unlike in 2018 where things were a draw or 2020 where markets won)
Except this time the polls correctly showed Democrats having a relatitvely good midterm and came very close the correct margin in most key races and a lot of journalists didn't take them at face value and went with the vibes instead.
3
9