According to Financial Times, around $580 million worth of oil futures (Brent + WTI) were traded in a very tight time window, roughly 15 minutes before US President Donald Trump posted on Truth Social about “productive conversations” with Iran. What caught the eyes of many were:
- Unusually high volume
- Timing of trade
- High directional conviction
- Unscheduled news break
- Low premarket liquidity
During war conditions, large directional bets happen repeatedly. Only the successful ones are noticed ex post. In this case, is there a reasonable case to leave aside suspicions and see if there is room for predictive models to explain the bet on oil ahead of the announcement?
The answer is “yes”.
It is common for highly informed macro traders or hedge funds to track geopolitical nuances, military signalling, shipping insurance curves, Middle East freight rates, use sentiment analysis and make sense of language patterns in official statements and Trump’s communication patterns, etc. Trump is known for reversal announcements after escalation, raising reasonable odds of de-escalation after the Hormuz threats peaked. But, is this how financial markets operate?
How predictions work in stock markets
Suppose, Nifty spot is trading at 22000, with an implied volatility (IV) of At-the-Money (ATM) option at 14% and put delta at 30%. This set of information can be converted into a few probability events like:
- Nifty has a 68% probability of moving +/- 4% in one month
- Nifty has a 30% probability of closing below strike
Stock market predictions based on option greeks as illustrated at the beginning of this article is “model based” and not the absolute truth. Fat tails also exist, meaning, crashes happen more often than normal models predict. These models produce risk neutral probabilities accounting for risk and interest rates. In other words, they answer the question: “What risks are traders paying money to protect against?”.
What then are prediction markets?
Prediction markets let people trade probabilities, using real money, on clearly defined future events, and market prices converge to a crowd estimated likelihood. Let us briefly look at how these markets work.
Let us take for example how Polymarket, a prediction market operator, creates a market for an event with rising oil prices as the theme. With “Will Brent Crude touch $200/bbl by end of April?” it brings about a clearly worded question, a deadline and a resolution source for settlement.
Two outcome tokens exist: YES, if 200 is reached, and NO, if 200 is not reached by end of April. Each token pays $1 if correct and $0 if wrong. The operator, Polymarket itself does NOT trade. It just runs the arena. Suppose the market currently shows:
- YES = $0.62
- NO = $0.38
This means that YES carries a 62% probability, and NO is at 38% probability. This is not opinion polling. This is money weighted belief.
Like stock markets, there are casual traders as well as informed traders who buy either YES or NO. There are also traders who think market has overpriced YES or NO and thus take contrarian positions thereby pushing prices back to equilibrium. Traders can also exit early, before April end, if the YES or NO fetch higher price than their buying price. On 30th of April, let us assume that Brent does touch 200 dollars a barrel. In such a scenario, YES token pays $1, NO token pays $0 and anyone holding YES makes 1- purchase price.
Stock markets Vs prediction markets
Option based probabilities are disciplined, but not pure forecasts. They predict movement, not bias. Prediction markets on the other hand works by turning beliefs into prices, and prices into probabilities, using real financial incentives.
In prediction markets, trade decisions are based on event outcomes, whereas in stock exchanges, trade decisions are based on cash flows. The price in the former is a probability, while that in the latter is valuation. While prediction markets are zero sum (minus fees), Equities are positive sum (economic growth). Longterm capital prefers wealth creation, not wagering.
In India, there are no regulated prediction markets like Polymarket or Kalshi. Previously, “opinion trading” platforms operated in a grey zone and have largely shut down after regulatory tightening in 2025.
However, we can only stay isolated thus far, as interconnection of global financial markets ensure that large events’ ripple effect is felt almost everywhere. That being the case, it is imperative that, we understand how prediction markets can influence events.
The predicament
The real predicament of predictions is not about its accuracy, as much it is about the influence on real world decision making. Steven Spielberg’s film, Minority Report is built around “PreCrime”, a system that uses predictive models to foresee crimes before they happen, and people are arrested and punished before committing the crime. Can the idea of prediction itself shape behaviour? When markets place odds on events like war or political upheaval, will it not legitimise speculation on human suffering, incentivise manipulation, and potentially encourage actors to bring about the very outcomes being bet on?
Philosophically and politically, this debate sits at the heart of the “predicament of predictions”. If a system predicts events accurately enough, and capital flows attach to those predictions, prediction risks turning into prejudgment and pre incentivization.
Where does prediction end and inducement begin? Socrates may have already given the answer: “Man’s wisdom should rise above his capacity to harm”.
