How to read a multi outcome market
A two outcome market hands you two prices. A multi outcome market splits the same question into brackets, where exactly one bracket pays a dollar and every other one settles at zero. The arithmetic looks familiar and the reading craft is not. This guide uses the July 2026 Federal Reserve decision market, closing 29 July, to cover what the prices mean, why they do not sum to 100, the coverage trap that can zero every share on the page, and why resolution rules matter more here.
Two outcomes versus several
In a binary market, YES and NO are welded together: one redeems at a dollar, the other at nothing, so holding one of each always pays a dollar and the pair costs about a dollar. A multi outcome market generalizes that. You get a set of shares, one per bracket, and exactly one of them redeems. Buying one share of every bracket produces that same guaranteed dollar, so the full set should cost about a dollar. Everything else follows from that sentence.
The July FOMC market, bracket by bracket
Here is the board for the July 2026 Federal Reserve decision, five mutually exclusive outcomes, exactly one of which must be the result.
| Outcome | Market price |
|---|---|
| No change | 79.8% |
| 25 bps increase | 19.4% |
| 50+ bps increase | 0.7% |
| 25 bps decrease | 0.4% |
| 50+ bps decrease | 0.1% |
Read it as a distribution, not a headline. No change is running at roughly four times the price of a quarter point rise, a cut of any size is priced as a rounding error, and the whole live argument sits on the boundary between the top two brackets.
Why these sum to 100.4 and not 100
Add the column and you get 100.4 percent. The obvious trade suggests itself: sell one share of every bracket, collect 100.4 cents, pay out 100 at resolution, keep the change. It fails, for a familiar reason. These are executable prices with the spread already inside them, so you would be hitting bids on every leg, and fees and locked capital take the rest. The overround is friction made visible, not free money. For clean probabilities, divide each price by 100.4, which puts the leader near 79.5 percent. That correction is cosmetic at the top and proportionally much larger on brackets priced near zero.
The trap: brackets that do not cover everything
The mirror image is the one that costs money. When the listed brackets sum to well under 100, that is not a discount. Either the book is thin and stale with no real quotes on some legs, or the bracket set does not cover every possible result and the missing probability belongs to an outcome you cannot buy. In that second case, if the unlisted outcome wins, every listed share settles at zero at once.
The check takes ten seconds. Add every bracket, then try to name a real world result that none of them would pay on. If you can name one, the gap is coverage risk, not opportunity. The five Fed brackets pass, because hold, up a quarter, up more, down a quarter, down more spans the whole number line. Sets built from named candidates, or from price ranges that stop at a round number, often fail. Some venues list an explicit other bracket, and its price is the market's own estimate of that risk. Where none exists, you carry it unpriced.
The shape says more than the leader
Two markets can print the same leading number and describe different worlds. One dominant bracket with a single live alternative, which is what the Fed board shows, is a settled consensus with one open question, so trading it is a bet on one boundary. Two close brackets at the top is a market genuinely undecided, where small news swings the lead back and forth. Probability spread thinly across many brackets means the leader holds a plurality, not a consensus. Brackets are ordered, so neighbors are substitutes, and news usually moves weight next door rather than to the far tail.
SmartX is an independent AI trading terminal for prediction markets. It tracks smart-money wallets by realized PnL and win rate, streams their live entries, and puts every venue in one terminal, so you can see which bracket the informed money is actually sitting in, for a flat 0.5% fee.
Open SmartX →Where the liquidity sits
Liquidity pools where the disagreement is, which here is the boundary between no change and a quarter point rise, and that is where the spread is tightest and the book deepest. The tails are thin structurally: a maker selling a share priced near zero risks most of a dollar to earn a fraction of a cent, so they quote wide and post very little size. Tail percentages are therefore the least reliable numbers on the page, a single tick can double one with no news behind it, and the figure you see may be days old. Check the live bid and ask before trusting a cheap bracket.
Read the resolution rules first
A binary market needs one definition, the one that says what counts as YES. A multi outcome market needs every boundary between adjacent brackets, because the structure only holds if they are mutually exclusive and jointly exhaustive. That is where disputes live: which announcement settles it, whether an edge case counts as the lower bracket or the higher one, what happens if the decision is delayed or taken outside the scheduled meeting, and what happens if the event does not occur at all. On a rate market, brackets written on the target range and brackets written on some other measure can split on identical news. Read the rules before the prices, not after you are holding.
A multi outcome board is a distribution rather than a coin flip, which is more information once you know how to read it. This guide is education, not financial advice.
Frequently asked questions
Why do the outcomes add up to more than 100 percent?
Because those are executable prices with the spread and fees already inside them. The five Fed brackets sum to 100.4 percent, and that extra 0.4 is transaction cost showing up in the quotes rather than a mispricing. Divide each price by the total to recover clean probabilities.
If the brackets add up to less than 100 percent, is that free money?
No, and it is the most expensive misread in this format. It usually means the book is thin with no real quotes on some legs, or the listed brackets do not cover every possible result. In the second case an unlisted outcome can win, and then every listed share settles at zero at once.
Should I normalize the prices before comparing them to my own forecast?
Yes. Divide each bracket by the sum of all brackets so the set totals 100. On the leader the adjustment is small, but on brackets priced near zero the overround is proportionally much larger, and skipping the step makes every longshot look likelier than the market actually thinks.
Is a bracket priced near zero a cheap lottery ticket?
Rarely at the price you see. Deep tail brackets carry the widest spreads and the least depth on the board, the quoted number may be a stale last trade, and one tick of movement can double it without any news. Check the live bid and ask before sizing anything.