When the Outcomes Do Not Add to 100%
A field of outcomes should sum to one dollar. When it sums to more, you are seeing the spread. When it sums to less, something is missing or stale.
Binary markets are easy: Yes and No sum to a dollar. Fields with many outcomes should also sum to a dollar, because exactly one of them will happen. They rarely do on screen, and the direction of the discrepancy tells you what you are looking at.
The check
Convert every outcome to a probability and add them up. Call the total the book sum.
book sum = Σ price(outcome i)
If the field is exhaustive and mutually exclusive, the true sum is 1.00.
When the sum exceeds 100%
This is normal. Almost every multi-outcome field sums to more than a dollar at the offer, because you would have to cross a spread on every leg to buy them all.
A field of eight outcomes summing to 106% means it costs $1.06 to guarantee a dollar. That six cents is the cost of immediacy across eight markets. It is the multi-leg version of the vig, and it is the right number to compare against a sportsbook's margin on the same field. Decimal, fractional and American odds, converted covers why the comparison has to be done in probability space.
The size of the excess is a decent liquidity measure. Two per cent over on a ten-way field is a tight market. Fifteen per cent over is a market where the spread will dominate anything you think you know.
When the sum is below 100%
This is the interesting case, and there are four explanations. Work through them in order before getting excited.
A stale leg. One outcome has not traded and its quote is old. This is by far the most common answer. Check timestamps. Stale quotes and false gaps covers the general problem.
The field is not exhaustive. There is an unlisted "none of these" or "somebody else" outcome. A market on "which of these five people gets the role" is not exhaustive if a sixth person could get it. The missing probability lives in the case nobody listed.
The outcomes are not mutually exclusive. Two of the listed outcomes could both occur, in which case the true sum should exceed 100% and a sum below it is even stranger.
Genuine mispricing. It happens, usually briefly, usually in the seconds after news, and usually in sizes too small to matter. Buying the full field below a dollar is the textbook arbitrage, and it is textbook because it is rare and quickly taken.
The practical difficulty
Even when a field genuinely sums below a dollar, capturing it requires executing every leg at the displayed price simultaneously. In practice:
- Some legs are thin and you cannot get size.
- Prices move while you are filling the others.
- Fees apply per leg, and a field of ten legs pays ten fees.
- You need capital equal to the full sum committed until settlement.
By the time those are accounted for, a sum of 99% is not an opportunity; it is the market functioning normally with one slightly stale quote. A sum of 92% is either an error in your data or a field that is not exhaustive.
The related structure: nested markets
Watch for fields where outcomes overlap hierarchically — "wins the nomination" and "wins the election" for the same person, or "above 3%" and "above 4%" for the same statistic.
These have logical constraints that are stronger than the sum rule. Winning the election requires winning the nomination, so the election price can never exceed the nomination price. "Above 4%" can never exceed "above 3%".
Violations of these constraints are far more meaningful than a book sum being off, because they cannot be explained by an unlisted outcome. They are either stale data or a genuine inconsistency. They are also rarer than people expect, because they are the first thing anybody automated looks for.
Reading threshold ladders as a distribution
A set of markets on the same statistic at different thresholds is a distribution in disguise. If "above 3%" is 70% and "above 4%" is 40%, then the market implies a 30% chance the value lands between 3% and 4%.
Doing this across a full ladder gives you the market's implied distribution, which is dramatically more informative than any single binary price. It also gives you a strong consistency check: the implied probability of every band must be non-negative, and if one comes out negative you have found a genuine inconsistency or a stale quote.
This is the single most useful technique in this guide, and it is underused because it requires reading several markets at once rather than one. Trading the calendar is where it pays off most.
Next
Telling a significant move from noise closes out the pricing section with the measurement this site's rankings are built on.
Educational content only. Nothing here is financial advice, a solicitation, or an offer to trade. Prediction markets can move against you and a position can lose its entire value. Where this guide describes a difference between venues it describes a theoretical cross-market edge in observed prices, which is not a profit: fees, spreads, capital lockup, transfer time and differing settlement rules all come out of it first.
Keep reading
Prices and Probability · 4 min readWhy Prices Near 1c and 99c Behave DifferentlyThe extremes are their own world. Tick size, capital cost and fees mean a price of 2 cents is not simply a small probability, and should not be read as one.
Prices and Probability · 3 min readTelling a Significant Move From NoiseA five-point move means nothing on its own. Measured against a market's own normal variation and its volume, it either means something or it does not.
Prices and Probability · 3 min readMid Price, Last Trade and Which One Is the MarketThree numbers claim to be the price and they disagree. Which one to quote, which one to trade against, and how to tell when none of them is real.
Kalshi · 3 min readThings That Surprise People on KalshiThe settlement quirks, fee shapes and structural details that catch new users out, collected in one place so they catch you somewhere cheaper.