It's Back !!!
JWC Wagering-Based 2024 Presidential Election Model

After 12 years of increasingly nasty & generally appalling American politics, it's time to fire up a simple, yet remarkably accurate model for using "free market" wagers to predict the winner of the 2024 Presidential contest.

To find out about the general assumptions & simple math of the model (which I devised in 2004), click this link.  Since 2004, it has always been correct:  See how it performed in 2006, 2008, 2010 & 2012.

What is different:  The earlier models used data from the British betting website TradeSports, but TradeSports political wagering was shut down by the U.S. government after 2012.  However, several websites have recently begun taking political wagers.  I'm using daily data from Polymarket (an American financial exchange founded in 2020) since it publishes wagering contracts at the state level.  Several other websites also publish aggregate wagering probabilities for the two candidates, but don't provide state-by-state granularity; these are summarized on the Real Clear Polling website.  

There are some caveats concerning Polymarket's data & the wagering model.  First, its political wagering market is closed to people in the United States, apparently due to legal constraints.  On the other hand, in addition to personal gain (or loss), that fact may cause such a market to possibly be more objective, assuming its participants have actually done state-by-state research.  In addition, as in financial markets, a few large bettors could perturb & / or manipulate the market itself, perhaps achieving gain indirectly or counter-intuitively.  In fact, some interesting articles have recently appeared in the Wall Street Journal about the predictive value of such markets & how "whale" bettors might affect them  ( see:  ref01  ref02 ).  Note that the wagering model does not account for minor factors such as lack of contracts for D.C. (the model allocates all 3 D.C. votes to Harris) or the fact that some states (such as Maine & Nebraska) are not "winner takes all".

And of course, lots of things can happen in late October:  Even higher levels of candidate hyperbole / lies, massive ad spending, scandals, hoaxes, terrorists & who-knows-what-else may sway the sentiment that informs wagering.  The robustness of wagering models is sure to be tested very soon.

In any case, forget about those highly subjective and unstable Opinion Polls & take a look at the current 2024 Presidential Wagering Model Prediction shown below (or download the pdf):