Comment by bombcar
18 hours ago
>Our final in 2008 consisted of two parts: predicting the electoral outcome of the Presidential election of each state where each state represented one percentage of our grade, and then a wager from 1-50 percentage points on whether the stock market would rise or fall the day after the election.
Explain this more? Let's assume you're Nate Silver and predict the 50 state outcome perfectly - you have a 50% in the class, so failing? Then the only way to "win" is to wager 50 points on the stock market (doesn't matter which way it goes). Wagering less makes no sense, because you start at 50 and so going "up" 25 to 75% protects nothing as the downside is still way below failing.
It sounds like a game theory question - you should be able to get 40 points on the states easy enough even if you get the toss-up ones wrong, and then gamble the full total on the stock market (which in general should go up, the market loves certainty and hates uncertainty).
It was exactly a game theory question, and a perfect exercise in real world betting markets. You’ll never have the most information and you’ll never be the biggest fish.
I learned the lesson that day, and I’d argue that even Obama with 365 electoral votes and control of the legislature learned it soon afterwards. Being a naïve hopeful Obama supporter, I bet 50 points on up and lost my ass.
Nate Silver came into the national spotlight after his analysis that year. There were other polling prediction models out of Princeton, but I heavily relied on Nate Silver and fivethirtyeight. I remember predicting every state correctly except North Carolina.
Interestingly in the context of this post, the University of Iowa has been hosting a market for real monetary binary options on US political outcomes for 30 years now. [1] It’s probably some small stakes fun for Midwest market makers looking for some action during off season corn futures.
Other things we learned: - The players club at Harrah’s marked the beginning of the rewards points programs available at nearly every single seller of goods today. - Casinos, in cracking down on card sharp teams playing blackjack with a mathematical edge and who had been 86’d but often returned in disguise, developed software to identify people from security camera footage by their stride. This was in 2008. - Bet the pass line, and stack the odds behind your number. It’s the best odds in the casino and nobody likes the guy betting Don’t.
[1] https://iem.uiowa.edu/iem/