Efficiency and Efficacy in Prediction Markets
Betting on Yourself Has the Highest Payoff
On February 10th, I wrote about the popularity of prediction markets like Kalshi, Polymarket and ForecastEx and did it in the context of lame attempts by Democrats to regulate them:
Bookmakers have been taking bets on almost anything for which there is demand, numbers rackets became state lotteries, the Las Vegas line on American football point spreads is what I recall most from growing up. In fact, Lloyds of London is featured in Chapter Five of The Moneyball Method as the catalyst for global trade efficiencies.
That catalyst was the merchants and shipowners at Edward Lloyd’s coffee house that had the best information about far-off places than the most powerful naval office in the history of the world – the English Admiralty. And the odds of the Democratic caucus being in command of the intelligence or integrity that compares to the English Admiralty is identical to Lloyd Christmas dating Mary Swanson in Dumb and Dumber: “I’d say more like one in a million.”
But the important point is that prediction markets deliver more useful information about future outcomes than the priests and pundits of public policy. One reason is that everything the “experts” know and say has been discounted by market forces. Another reason is that every participant has some form of skin in the game.
And according to Polymarket, a decentralized platform using Blockchain technology, their success rate is 97% of resolved outcomes within four hours of determination, 94.4% within one week, and 92.1% one month out. Meaning that in the aggregate of all Polymarket betting arenas, the winning “Yes” or “No” bets were correct 97% of the time with only four hours to resolution and 92% were correct with one month remaining. Obviously, there is time value to everything, but in this context, to be correct means that a simple majority favored the outcome.
So what? Some of these may have become painfully obvious and a simple majority is not an overwhelming endorsement. As my previous article opined: “Predictions markets are as old as dirt, or at least markets for dirt. But high-speed internet and their creative entrepreneurial types are relentless in catering to the whims of people hungry for cheap meaning in their lives.”
Liquidity and Arbitrage
To further legitimize the predictive and fairness aspects of these playgrounds, Polymarket also publishes their Brier score, which does a better job of measuring the prediction accuracy of its betting pools. But to me, there is something far more interesting, it lies beneath the surface, and that is the opportunity and market for arbitrage transactions.
The first betting market I chose to inspect offered “Yes” and “No” choices for several different end dates, but they all added up to 99 cents for a one-dollar win. That may not seem like a lot, but anyone who takes both sides of the bet simultaneously will earn 1%, risk free when the betting pool is resolved. That is arbitrage: coupled transactions to profit from price disparities within marketplaces or among different markets for identical goods.
If it is an illiquid market, meaning too few participants, the spread may be greater than 1 percent, thus making the arbitrage opportunity greater. But speculators will need to find buyers and sellers at the same time for a profitable trade - and Polymarket offers that opportunity as well through market sponsorship. And if successful, liquidity is then added to that market, the spread shrinks for future market players, and all of this increases market efficiency for everyone.
Iran War Case Study
Back to the current level of information about uncertain outcomes in liquid Polymarkets, consider the American and Israeli mission to end the war initiated by the Islamic State of Iran in 1979. As of this writing (0845 ET on 14 March 2026), the betting arena titled “Iran x Israel/US conflict ends by,” the betting line is a 72% chance of military action by both sides ending on or before June 30, 2026.
The rules begin, “This market will resolve to “Yes” if there is a continuous 14-day period without any qualifying military action between Iran, and Israel and the United States that begins at any time between market creation and the specified end date (ET). Otherwise, this market will resolve to “No”.” In this case, to bet “Yes” will cost you 73 cents per share and to bet “No” will cost you 29 cents.
To a prediction markets novice like me, this raises red flags. What is “qualifying military action?” Particularly in a war that one side has been waging for 47 years. In this case, the market maker for this Polymarket defines that as: “any use of force by Iran, or Israel and the United States against the other’s soil, or official embassies or consulates, that is either officially acknowledged by the acting government or confirmed through a clear consensus of credible reporting.” This raises even more red flags.
Subjective Market Dangers
Official acknowledgement from a crime syndicate like the Islamic State means nothing. On top of that, they have proxies doing their bidding. And not much better is “clear consensus of credible reporting.” To help with that, this market maker clarifies, “includes, but is not limited to, airstrikes, naval attacks, and ground incursions. Cyberattacks, sanctions, and diplomatic actions do not count.” But who establishes consensus and what news agency is credible?
For better or worse, the referee used by Polymarket for disputed outcomes is the UMA Optimistic Oracle (universal market access). Briefly, this is a decentralized resolution and dispute process that requires a $750 bond to be posted by the “Proposer” of the resolution, and if there is a dispute, a $750 bond posted by the “Disputer.” Ultimately, the question is adjudicated by the vote of the qualified token holders. That is because there is no measurable profit incentive for the primary market participants - the governments of the United States, Israel and Iran.
I do not have enough knowledge or experience to opine on the efficacy of the UMA Oracle resolution and dispute process, but I do believe that devoting your time, energy, and money to events over which you have no skin in the game can be destructive. Time, energy, and money are the stuff of living your own life with pride.
To borrow your self-worth from the success or failures of others can only lead to more destructive behavior. To get politicians involved who have chosen careers for regulating the success or failures of others can only compound the problem.
Market Elegance
At the same time, the prediction markets can be a marvelous tool for teaching students about the efficiency of the price mechanism and elegance of free markets. On this, Dimensional Funds is a prime authority and offers their perspective:
Sports betting markets may be a useful parallel. When an undefeated team squares off against a winless team, few expect the latter to emerge victorious. The only way to induce gamblers to bet on the weaker squad is to lower the “price.” This is accomplished by a point spread indicating essentially how much the underdog can lose by and still be considered a winner for betting purposes.
Las Vegas line point spreads for NFL football have been remarkably accurate and reliable. So are prediction markets. But betting on the outcomes of strangers is not objective and it is not investing.