A price is a probability, until it isn't.

A contract that pays $1 if an event happens is worth its probability. If it traded at 40¢ while the event happened 60% of the time, buying it would make money on average, and people would buy it until the price stopped being 40¢. Prices on a liquid market are therefore pushed toward the crowd's collective estimate by the simple mechanism of it being profitable to push them there.

What that gives you is an estimate, not a measurement. It reflects the beliefs of the people who chose to trade, weighted by how much they staked. That is a genuinely useful thing — it aggregates information no single participant holds — and it is not the same as knowing what will happen.

The price is where a trade breaks even before costs. Once the fee and the crossed spread are included, the probability you need is strictly higher than the number on the screen — about 51.8% on a 50¢ contract taken at the offer, and higher again on a wide book. Every contract on the exchange is therefore priced slightly against a taker, in the same way and for the same reason a casino's odds are.

This is not a criticism of the exchange; the fee is public and modest. It matters because the gap is usually the same size as the edge people believe they have. The break-even guide works it out in cents at each price.

Across a century of racetrack data and a good deal of sports betting since, one pattern recurs: longshots are overbet and favourites are underbet. Outcomes priced at very low probabilities happen less often than their prices imply, and heavy favourites happen slightly more often than theirs. The usual explanations are behavioural — small stakes buying a large payoff are attractive out of proportion to their value — and the effect has been observed in prediction markets as well as bookmaker-set ones.

Two things follow, and only two. Cheap contracts are, on this evidence, worse value than their price suggests; and the bias is a documented tendency across large samples, not a rule that holds on any particular market you are looking at. It is a reason to be sceptical of a 3¢ lottery ticket, not a strategy.

Buy contracts at 95¢ and you should expect to win about 95% of the time. That is what the price means. A trader who does exactly that and reports a 95% win rate has demonstrated nothing whatsoever — they have reproduced the price. After fees they have lost money.

This is why every win rate published on this site renders beside the average entry price that produced it, and why the gap between the two is the only figure that carries information. The table is the same arithmetic at several prices: the third column is the edge, and it is the only column worth reading.

A 40% win rate at 30¢ beats a 95% win rate at 95¢, decisively, and it looks far worse in a screenshot. Anyone selling a track record without the entry prices beside it is selling the screenshot.

A price is only as informative as the money behind it. A market quoted 20 bid, 45 ask with no volume is not telling you the probability is 32.5%; it is telling you nobody has bothered to find out. Wide spreads, stale quotes and small resting size all degrade the reading, and they degrade it most on exactly the obscure markets where a newcomer is most likely to believe they have found an inefficiency.