Stellaris

Learn

What a prediction-market trust score is, and what it can't tell you

A prediction-market price tells you what traders expect. A trust score tells you how much weight that price deserves. Here is how ours works, and where it stops being useful.

Last reviewed 3 November 2026 · Method v1

The short version

A prediction-market price is a number from 0 to 100 that you can read as a probability: a contract priced at 70 cents implies about a 70% chance. What the price doesn't say is how much money stands behind it, or whether a handful of accounts are setting it.

A trust score fills that gap. It is a number from 0 to 100 for how much weight a price deserves, shown on our pages as one of five levels. It is not a forecast and it is not advice. It says how solid the number is, not whether the number is right.

Why two identical prices can mean different things

Take two markets that both show 94%. In one, millions of contracts are waiting to trade within a few cents of the price, so one trader can't move it much. In the other, a few thousand contracts are waiting, and one order can push the price several points.

Our midterms brief showed the pattern. At 21:00 UTC on 2 November 2026, Polymarket's market on Democrats winning the House had about 16.6 million contracts within 5 cents of its price. A "blue mega-tsunami" market had about 9,000. The first is hard to push around. The second moves on small trades and is better read as colour than as a signal.

What goes into the score

Each market is checked on six things, scored from 0 to 100 and combined with fixed weights:

  • Spread: how far apart the best buy and sell prices are. A wide gap means the price is a guess between two views.
  • Depth: how many contracts are waiting within 5 cents of the price. More depth means one trader moves the price less.
  • Activity: how recently the market last traded. A price that hasn't traded for days may be stale.
  • Concentration: what share of recent trading came from the few busiest accounts. High concentration means a few traders could be setting the price.
  • Wash trading: what share of recent trading looks like the same money going round in circles, such as one account on both sides of a trade.
  • Turnover: how much trading there is compared with the money held in the market. Very high turnover can mean churn rather than new conviction.

Depth and spread carry the most weight, because they say most directly how easily the price can be moved. The weights are fixed, and every score carries the version of the method that produced it, so an old number can always be traced to the rules in force at the time.

The five trust levels

The 0 to 100 score is shown as five levels, so a reader can take it in at a glance:

  • 5, Strong: 80 and above.
  • 4, Good: 65 to 79.
  • 3, Fair: 50 to 64.
  • 2, Thin: 35 to 49.
  • 1, Weak: under 35.

What a trust score can't tell you

  • Whether the price is right. A deep, well-traded market can still be wrong, and a thin one can still be right.
  • Anything about the future. The score describes how the price was formed, not what will happen.
  • Manipulation spread across many accounts. Our wash-trading check looks for simple patterns, such as the same account on both sides or two accounts trading back and forth. It will miss schemes spread across many accounts, and it can flag a legitimate market maker. That is why it carries only a small share of the score.
  • What the contract's rules say. Two venues can word the same question differently, with different dates or definitions. A gap between venues isn't always a disagreement. Read the rules before comparing.
  • Everything on every venue. Some checks need data a venue doesn't publish. Kalshi's public data has no account information, for example, so concentration and wash trading can't be measured there, and its score rests on fewer checks. The score records which checks it used.

The method is new. Version 1 has not yet been tested against real outcomes at scale. We will publish how it performs, including where it was wrong, and corrections are added as new, dated entries rather than edits.

How to use a trust score

  • Look at depth first. If little money is behind a price, treat it as colour, however confident the percentage looks.
  • Check the time. Prices move, so every number should carry when it was taken.
  • Compare venues only after reading the rules. A gap can come from different wording rather than a different view.
  • Treat low scores as a reason for caution, not as a trading signal. Nothing here is advice.

Sources

This is analysis of public market data, not investment advice or a recommendation to trade. Stellaris is not a trading venue, broker or adviser and does not trade. Prices are those shown by the venues at the times stated and can change. Trust levels come from our published method and have known limits.

See the latest briefRequest access