Opinion review

Opinion is an on-chain binary market on BNB Chain with the widest category set of any venue added to this site in September: macro, sports, esports, crypto, politics, business, finance, culture, technology and pre-token markets, plus hourly up-or-down markets on three assets and a dedicated World Cup section.

The books are populated rather than nominal. When we read the app on 8 September 2026 it carried live two-sided prices across macro, sports, esports and hourly crypto markets simultaneously — Sabalenka, Tiafoe and Invictus Gaming all quoted, alongside hourly BTC, ETH and BNB. By reported volume it sits third in the category behind Kalshi and Polymarket.

It also names its operator in its terms of use, which most on-chain venues do not. That is worth crediting even though no jurisdiction of incorporation, licence or regulator is stated anywhere we could find.

The access rules are unusually broad and they are the first thing to check. Opinion is closed to the United States, the United Kingdom and all of the European Union. That is a larger exclusion than almost any venue we cover, and it is worth reading before anything else on this page, because for most English-speaking readers it settles the question.

Where it is reachable, the venue is one of the more substantial options outside the top two. Our own availability rows record the restriction with the date we read it; if the venue’s own terms and our page disagree, theirs is current and ours is stale, and we would rather be told than be wrong quietly.

Opinion documents its dispute process more concretely than most: a window opens after resolution, “typically 2-hour long except some asset-price related markets”, and filing a dispute requires a stake of “max(1% of market open interest, 1,000 $OPN)”. A rejected dispute forfeits “your entire stake”, and “Each resolution allows only one dispute”. Those are real numbers and real consequences, which is more than several larger venues publish.

The primary resolver, though, is “Opinion AI”, described in the documentation as “the primary oracle mechanism” — with no account anywhere of its inputs, its model, or the sources it reads. A market is settled in the first instance by a system the venue will not describe.

The review step has the same problem in a different form. It is written without an actor throughout: “Once a dispute is filed, a review process will take place”, and “The dispute automatically succeeds if the review process times out”. Who performs that review is never identified, and whether the team can resolve or override a market itself is not addressed at all.

The stake formula deserves a second look before you rely on it. One per cent of open interest is a sliding scale, so on a large market the cost of challenging a settlement rises with the amount at stake — which is defensible as an anti-spam measure and also means the markets where a wrong settlement costs most are the markets where challenging it is most expensive.

The floor of 1,000 $OPN is a token-denominated cost, so its real value moves with the token. And because only one dispute per resolution is permitted, the first challenger effectively spends the market’s single appeal: if they file badly, nobody else can file well.

Opinion also reserves a stated discretion to adjust the stake percentage. None of this is hidden — it is all in the venue’s own documentation, which is why the rating is what it is rather than lower. It is the unnamed resolver and the actorless review that hold it down, not the parts that are written plainly.