Course · Module 1 · 15 minutes
Prices are probabilities
A prediction market contract pays out a fixed amount if something happens and nothing if it doesn't. On the big venues that amount is one dollar. So when a Yes contract trades at 63 cents, the people trading it are saying, between them, that the event has about a 63% chance. That's the whole idea, and I've built everything else in this course on top of it.
Reading one price
Say you buy Yes at 40¢. If the event happens you get 100¢ back, so you make 60¢. If it doesn't, you lose your 40¢. For that to break even over many trades you need to be right 40% of the time, because 0.4 × 60 equals 0.6 × 40. So the price is also your break-even probability before costs, which is the first thing I check on any trade.
When Yes and No add up to more than 100
On a real order book you'll often see something like Yes at 63¢ and No at 39¢. Those add up to 102¢. The extra 2¢ is the spread, and it's what you pay the people providing liquidity. To get the market's own estimate, divide one price by the total: 63 ÷ 102 is about 61.8%. I sometimes use the midpoint between the best bid and the best offer on the Yes side instead, which comes to much the same thing on a liquid market.
| Yes price | No price | Total | Market's estimate |
|---|---|---|---|
| 63¢ | 39¢ | 102¢ | 61.8% |
| 50¢ | 52¢ | 102¢ | 49.0% |
| 20¢ | 81¢ | 101¢ | 19.8% |
Your own number
A price only becomes a trade when you have your own probability to compare it with. If you think the event is 50% likely and Yes costs 40¢, each contract is worth 50¢ to you on average and costs 40¢, so the expected value is 10¢ a contract. If you think it's 35%, the same contract is worth 35¢ to you and you'd be paying 5¢ too much. The hard part is making your 50% better than the market's 40%, and that's what I spend the rest of the course on.
Check a price
Try these
Have a go before you open my answers. None of them needs more than the arithmetic above, and the point is to get quick at it.
First, Yes is showing 72¢ and No is showing 30¢, so work out what the market thinks and how much spread you're paying.My answer
They add up to 102¢, so you're paying 2¢ of spread, and the market's own estimate is 72 divided by 102, which comes out at about 70.6%.
Second, say you rate an esports favourite at about 65% and Yes costs 58¢. Work out what 100 contracts should make you before fees.My answer
Each contract is worth roughly 65¢ to you and you're paying 58¢ for it, so you're 7¢ ahead on each one, or about £7 (or $7) across the hundred on average.
The last one is the one people get wrong, so explain to yourself how a trade with positive expected value can still lose you money.My answer
Expected value is what you'd make on average over a lot of trades like it. This particular contract still settles at 0 or 100, so you can have the probability right and lose anyway, which is exactly why module 4 is about sizing so that a run of those doesn't sink you.
Module 2 picks this up with fees and spreads and the break-even you really face once they come off. In the meantime I'd try a few tickets on the practice board, since doing it a dozen times is what makes it stick.