Comment Text:
Gaming
I view election prediction markets as based on skill - not gaming. This is based on my personal experience (and that of many others!) of doing research, building models, and consistently making money. Of course it is possible that I was just “lucky”, but this repeated pattern of having an edge fits with a general pattern in my life where I have consistently made accurate and profitable predictions in multiple markets.
My background is that I'm a self-taught web developer with a BA in Computer Science and Economics and MA in Sociology.
In 1993, I traded my first election prediction market for the Canadian Federal Election - the University of British Columbia Electronic Stock Market (based on the University of Iowa one). That was the only year that I lost money on prediction markets - and I minimized my losses to $14.
In 1996, I traded the US presidential primary on the University of Iowa Electronic Stock Market and made several dollars. I also started work on an arbitrage bot (and made 5 cents on manual arbitrage). This is the earliest historical example that I know of anyone developing a bot for a prediction market.
In 2012, I bought my first (and only) house within 1%-2% of the bottom of the Philadelphia housing market.
In 2016, I built a multi-variate regression model at the county level including demographics, past voting, FB likes and more for the Democratic Presidential Primary featuring Clinton vs Sanders. After weekly iterations and improvements, my model ultimately had a lower mean error than the 538 model. I also had a real-time model that was often a couple minutes faster than the NYT “The Needle” (and very similar to its predictions). I turned $500 into $5000 on PredictIt. In the general presidential election, I had no bets on the outcome until I saw that Trump was going to win and made $3000 on PredictIt on election night.
In 2017, I built an adhoc model to predict the Philadelphia District Attorney Democratic primary - mostly for fun as there was no prediction market on the race. It accurately predicted that Larry Krasner would outperform the polling average and win.
In 2017-2022, I consistently made money on PredictIt by doing a lot of research and using my knowledge and experience from following international politics for the past 35 years. For international elections, I will do a deep dive into the current and historical conditions of a country. I will analyze the candidates, polling, and historical election results - as well as rely upon several of my overarching hypotheses regarding how other traders misprice contracts. The fact that I'm able to make money after paying a 10% fee on my profits is evidence that my methods work.
In the summer of 2022, I created an adhoc model including a combination of qualitative and quantitative factors for US political primaries and wrote 60 pages of notes. Notably I was able to predict a lot of winners in races that lacked polls. For instance, in the NY 19 special election, I predicted that the Democrat who was behind by 6% in the polls was a good deal at 30 cents to win (and he won by 2.4%). In the Democratic NY 12 primary that featured two incumbents, I had Nadler as a good deal at 30-35 cents vs Maloney while the latest old polls had them tied (Nadler ended up winning by a landslide). Overall I had approximately a 20% edge on my bets on PredictIt and made several thousand dollars after fees.
In 2011 to present, I consistently made above-average predictions regarding the pricing of cryptocurrency based on my research into market fundamentals, psychology, and the similarities with the stock bubbles in Japan, Taiwan, and US communications/tech. For instance, in 2011 I shorted Bitcoin down to $2.20.
I have never gambled at a casino, bought a lottery ticket, played slots, or even participated in a charity raffle. I do not gamble on games of chance. I am a prediction market trader and I buy or sell shares on prediction markets based on my analysis of their expected value.
I bet on candidates based on expected value. I am a socialist, but I have made money betting on Trump and other conservative candidates (ex. I had a large bet on Keiko Fujimori, the daughter of the famous Alberto Fujimori Peruvian far-right dictator - because I got it at 30 cents and she came within 0.2% of winning). I am able to almost completely remove my political biases from my prediction market betting and this is one of the reasons I make money.
Public Interest
Accurate election prediction markets promote election integrity. They are essential to protecting our democracy against an onslaught from people who want to discourage voting, intimidate election officials, illegally gerrymander districts, and steal elections.
We are facing a choice as to how we best inform the public about elections. Do we simply accept the views of partisan political pundits and politicians who are pushing their agenda and whose priorities are to make money and help their party win at any cost? Or do test these biases with science? I think we should go with the imperfect science of polls, data models, and prediction markets as our best scientific approach to predicting election outcomes, informing the public, and protecting our Democracy.
Of course the science is imperfect. But prediction markets are useful as they encourage people to build data models (like myself) and are able to react in real-time to events where it could take polls days or longer to identify trends.
If the CFTC permits election prediction markets to grow, we will see a flourishing of data models. And this is likely to lead to improved polling and forecasting in the future.
...
In many election races there is a lack of public polling. Either the polls are kept private by the campaigns that paid for them, or often times the race is so minor that there just isn't that much polling. In these cases, it is in the public interest to have the best estimate of the state of the race and prediction markets can provide this (Ex. see my model for the 2017 Philadelphia District Attorney primary).
...
Election prediction markets encourage the public to engage and educate itself about the political process. This most directly affects the prediction market users, but also their friends and acquaintances are drawn into conversations about politics that would otherwise not have occurred. In addition, the media can use election prediction markets to educate the public about elections.
16. Should campaign committees, political action committees, candidates for the House and
Senate, and other entities involved in political fundraising and expenditures or likely to
hold non-public information, or subject to Federal Election Commission oversight, be
prohibited from participating in the contracts?
Yes. This is a clear conflict of interest.
