The Wages of Election Modeling: Harry Crane vs. Super Model
Putting forecasters to the test by simulating the gains (or losses) you'd make trading their recommendations on Polymarket every day
This is our 4th forecaster match-up. See the original post in the series for more context on what’s going on here.
Previous match-ups:
The Rules
Forecasters start with $10,000.
Once a day, the forecasters get to buy or sell/short Trump shares, depending on whether their own odds are higher or lower than the Polymarket price.
The size of each trade is driven by the size of their bankroll and how certain they are of the mispricing - the bigger the mismatch, the bigger the trade.
We mark the holdings to market each day, so unrealized gains are reflected in daily account value.
A total of 48 trades are made, from September 19 through November 5. Any realized gains booked along the way are theirs too keep. After the election, all Trump shares pay $1 and Harris shares become worthless.
Harry Crane vs. Super Model
TL;DR Super Model’s own multi-market-derived forecast did loads better than any of the forecasters we’ve scored thus far, but that’s a low bar. That’s evidenced by the fact that we still would’ve lost money (-15%) using it to trade on Polymarket, since their Trump price was consistently (if only slightly) higher than ours, meaning we would’ve consistently been making small pro-Harris bets.
Crane’s forecast, on the other hand, blows everyone out of the water, returning +85% over the 7 weeks of simluated trading (the only profitable forecast we’ve encountered so far). If you haven’t seen it, check out Crane’s final forecast, which had Trump at 66.5% to win and explains how his methodology differs from legacy forecasters.
Below, we dive into the daily performance metrics - how the respective forecasters’ odds differed from market prices each day, how big a daily bet they place on Trump or Harris as a result, and how their portfolio value grew/shrank over time.
Previously…
