Interview with Johann Evrard, CEO, Predictive Labs.

The foundation of a Meta-Market is a point about prices that is easy to miss. When 2 venues show different prices on the same event, it is natural to read it as disagreement: Polymarket believes one thing, Kalshi believes another. Often, that is not what is happening. The same event can trade at different prices for several structural reasons: different fees, different amounts of money sitting in the market, slightly different wording in how the bet is settled (how the winner is decided), a different referee deciding the outcome, sometimes even a different currency it pays out in. Here, the gap is mostly mechanical...

That is why anchoring on a single venue's price is a mistake. The number worth reading is the event's, pulled together from every venue that lists it. That combined view of one real-world event is what we call a Meta-Market.

A concrete case makes it clearer. 'How many times will the Fed cut interest rates in 2026?' runs on Polymarket and on Kalshi. Chances are the wording differs, the settlement rules differ, the currency differs, and the 2 prices sit a few points apart. But if you glance at both, you'll still think: that is obviously the same market. A Meta-Market is what lets a system reach that same conclusion on its own, and then show you one clean reading of the event instead of a dozen noisy ones.

From there, 2 things become possible, and they run along 2 axes.

The first is vertical: stacking every contract on the same event into a single Meta-Market. Once the same event is in one place, you can see when one venue prices it higher or lower than another. If that gap is wider than the fees and trading costs can explain, it is a real mispricing, and the only way to catch it is to look at the event as a whole rather than one venue at a time.

The second is horizontal: linking different events that move together. A Fed rate decision is tied to the markets on stock indices, the dollar, and government bonds, and the framework will show how a move in one of those central events should carry across to the ones connected to it. For a trader or a newsroom, the benefit is simple: a single combined number for each event, instead of half a dozen different prices on as many different venues that they would otherwise have to compare and merge by hand.

On whether this becomes a new kind of asset, it depends on which kind you mean. It is not a financial asset. Predictive Labs issues nothing and trades nothing: we run no venue, hold no positions, and never touch anyone's money. We are an analytics layer sitting above the venues, and there is no Meta-Market you can buy. It is closer to an information asset, a shared reference point, the way a stock index is one. The S&P 500 is a number people read to understand the market, because no single company is the market, so someone has to define the shared measure. Prediction markets are at that stage now. Many contracts point at the same event and none of them is the definitive one. The Meta-Market is meant to be that reference: the benchmark you read for an event.

The question bundles 2 problems that are worth pulling apart, because they have different causes and different fixes. One is about words: how a market is written, and who gets to decide what its wording means. The other is about power: who controls the vote that settles a disputed outcome.

Take the words first. Every market has a referee, a rule or a source or a group that decides the final outcome. In this industry that referee is usually called the oracle. The difficulty is that 2 markets can look like the exact same question and still have different referees who could honestly land on different verdicts for the same event.

The 2 disputes you raise are really 2 flavours of the same problem. The US-Iran case is about what a word means: nobody disputes what was signed, only whether it clears the bar of 'permanent'. The MicroStrategy case is about timing: whether a sale counts from the moment it happens or only once it is publicly provable. In both, the facts are not in question; the reading is. And a market written differently, or judged by a different referee, could land on a different answer to the very same news.

Our approach is to make that ambiguity visible. When 2 markets could settle differently under their referees, we keep them apart instead of merging them, and we show you they can resolve differently. That is the honest way to handle ambiguity: you put it in front of the user, clearly labelled.

On the power side, you have laid out the mechanism already, so I will not repeat it. The point worth adding is that this is a design problem, not a conspiracy, and one the people running these systems recognise. The reform efforts floated so far have stalled.

I want to be clear about something: we are not the final judge that settles these disputes, and we never intend to be. We do not resolve markets or overrule a referee. The moment an analytics company starts deciding outcomes, it stops being neutral and becomes another referee with its own stake in the result. What we can do is different: we aim to surface the risk of an unexpected resolution, rather than settle it. A credit rating is the closest analogy: it tells you in advance how likely a company is to fail to pay, so you can decide whether to take the risk. The aim is to warn you in advance, not to rule on the outcome.

This matters most for ordinary participants. A large, sophisticated trader can already read these risks. A retail user usually cannot, and is the one most likely to lose money to an ambiguity in the wording they never noticed. Making that risk legible is, for me, where the real public value sits.

The honest answer is that the line between a forecasting market (like the stock market or a commodities market) and a betting market (like an online casino) is blurred at best. The same exact contract can be a forecast for one person and a bet for another. A farmer who sells a contract on rainfall is hedging a real risk; the person beside him buying it for fun is gambling. The instrument is identical. What differs is who is using it and why.

There are, however, objective characteristics worth looking at, and they live in the structure around the contract and the way people use it.

A few are worth naming, because you can point at them. The cleanest is who sits on the other side. In a casino or a sportsbook you play against the house: the house sets the odds, builds in its margin, and makes money when you lose, so your counterparty wins on average by design. The major prediction market venues run the other way. You trade against other people, the way you do on a stock exchange, and the venue takes a small fee whether you win or lose without taking the other side of your bet. There is no house in the middle that wins when you lose. That is one of the oldest lines between investing and gambling, and by that measure these markets sit closer to investing.

The second test is what decides the payout. A contract that settles on a real, verifiable event, an interest-rate decision or an election result, behaves differently from a pure game of chance.

The third is whether the price carries information. A useful forecasting market produces a number you can use elsewhere, the way newsrooms and professional traders now read these prices as a real-time signal. A simple wager produces nothing anyone outside the bet would care about.

The most telling difference, though, is the last one. In a sports bet you put your money down and you win only if you predicted the result correctly. In these markets a position can gain or lose value simply because the price moved, and you can sell out before the event is even decided. That is how a stock or a bond trades. By that measure these markets share more features with financial markets than with betting, though whether that matters legally is exactly what regulators are weighing.

There is a live fight in the United States over exactly this. Part of the gaming industry wants sports-related contracts treated as betting and kept out of financial regulation; the other side argues they belong under the financial regulator. I am not going to tell a regulator where that line should fall, since whether a specific contract counts as gambling is a legal call for regulators and courts, not for an analytics company. And on our own position: we take no positions and hold no money, so that question sits with the venues, not with us.

On whether this can help regulators, the useful part is that the characteristics I just described are structural, so they can be checked rather than argued over. A framework that draws the line on structure (whether the contract settles on a public outcome, whether there is a house on the other side, whether the price carries information) is harder to game than one that draws it on subject matter (sports versus economics). Event-level data is what makes those structural tests measurable across venues. Where the line finally falls is for regulators; what we could add is an objective basis for drawing it.