The price is not the probability.
What a market percentage tells you, what it leaves out, and how to keep your own estimate separate.
Start with the contract, not the number
A prediction market might display a YES price of 42 cents. That is a price for a particular contract, under a particular set of rules, at a particular moment. If the contract pays one dollar when its condition is met and zero otherwise, 42 cents invites a familiar probability interpretation. The important word is interpretation. Nobody has measured the true probability and placed it on the screen.
The number emerges from orders, available capital, participants’ beliefs and the venue’s mechanics. It may incorporate useful information that you do not have. It may also reflect a thin order book, an old trade, different financing costs, or confusion about resolution. Treating the displayed percentage as an observation preserves its value without assigning it more authority than it deserves.
Keep two columns in your head
One column is what the market currently offers. The other is your own estimate of the event. A useful research process keeps them separate long enough to inspect the disagreement. If the market price is 42 cents and you estimate a 60% chance of YES, the gap is 18 percentage points. That gap is not evidence that you have discovered a mispricing. It is a statement about the difference between two inputs.
Ask where your 60% came from. Did you begin with a relevant reference class? Did you read the resolution rules before considering the headline? Could you explain the evidence that moved you from 50% to 60%? If the estimate is mostly an emotional response to a compelling story, adding decimal places does not make it more rigorous. A range of defensible estimates is often a better starting point than a single confident number.
Work through both possible results
Consider 100 hypothetical contracts bought at 42 cents each. Entry cost is $42. If YES resolves and each contract pays $1, the gross receipt is $100 and the net result before costs is $58. If NO resolves, the net result is negative $42. Those are the two modeled outcomes; neither is the probability-weighted average.
At your assumed 60% probability, the expected net result is 0.60 times $58 plus 0.40 times negative $42, which equals $18. Equivalently, multiply probability by the total possible payout and subtract entry cost. This arithmetic describes a repeated-experiment average under the assumed probability. It does not mean the particular position will earn $18, and it does not verify the 60% input.
Costs change the comparison
Suppose you add an entry fee equal to 2% of the $42 cost. That is $0.84, bringing total initial cost to $42.84. The modeled YES result becomes $57.16 and the NO result becomes negative $42.84. Expected net result at the same probability falls to $17.16. Break-even probability becomes 42.84%, rather than 42%.
Real contracts may use different fee formulas and may involve exit costs, slippage or financing. The Oddscope stress test deliberately uses a simple entry-only fee assumption that you can inspect. Its result is useful as a sensitivity exercise, not as a substitute for reading a venue’s actual schedule. If changing a plausible cost assumption reverses your conclusion, that is information worth recording.
A good disagreement has a stopping condition
Before becoming attached to a view, identify evidence that would reduce the gap. Perhaps the market definition includes an exception you missed. Perhaps the relevant deadline is earlier than the public announcement you expected. Perhaps a direct source contradicts the interpretation circulating in commentary. Write down what would make you lower or raise your estimate, and separate that evidence from price movement itself.
You can then use the market as a conversation partner rather than an authority or an opponent. Check its observation time, read its rules, state your estimate, and test a range of assumptions. The aim is not to make the numbers agree. It is to understand why they disagree, which uncertainty remains, and whether your conclusion survives a less convenient version of the story.