Mid Price, Last Trade and Which One Is the Market
Three 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.
Open a market and you may see 62, or 60/64, or "last: 58". These are three different things and treating them interchangeably produces bad analysis.
The three numbers
Last trade is the price of the most recent transaction. It is a historical fact. Its usefulness depends entirely on when it happened.
Best bid and best offer are the highest resting buy order and the lowest resting sell order. They are live. Together they define what you could actually do right now.
Mid is the midpoint of bid and offer. It is not a price anybody is offering; it is the centre of the range in which a trade would happen.
Which to use, by purpose
To quote a probability: use the mid, and state the spread alongside it. "62, with a 60/64 market" is an honest sentence. "62%" alone is an overstatement of precision whenever the spread is wide.
To estimate what a trade costs you: use the side you would cross. Buying means paying the offer. Selling means hitting the bid. Your realised price is never the mid, and assuming it is will make every strategy look better than it is by half the spread each way.
To measure a move over time: use the mid at both ends. Comparing last trades introduces noise from which side each trade happened to hit; a sequence of buys and sells at an unchanged market will produce a "moving" last trade that has not moved at all.
To check whether a quote is alive: add Yes and No. If the pair is far from a dollar, something is stale. Yes and No shares covers the identity.
When the last trade lies
Three common cases.
Staleness. A market with no trades for two days shows a last trade from two days ago. The world has moved; the number has not. Any dataset built on last-trade prices will contain these ghosts, which is a real problem for backtesting.
Bid-ask bounce. In an active market with a two-cent spread, consecutive trades alternate between 61 and 63 as buyers and sellers arrive. The last-trade series oscillates by two cents continuously while the market is unchanged. Compute volatility from that series and you will overstate it substantially.
Size. A last trade of one contract at an outlying price is not the market. A last trade of 5,000 contracts is.
When the mid lies
The mid assumes the two sides are comparable, and sometimes they are not.
Asymmetric depth. Bid 60 for 3,000 contracts, offer 64 for 12. The mid says 62. The market, weighted by what is actually available, is much closer to 60. A size-weighted mid is more honest, and it is what this site uses when computing cross-venue differences, because an unweighted mid on a lopsided book overstates how tradeable the price is.
A one-sided book. Bid 40, no offer at all. The mid is undefined and any number you produce is an invention.
Very wide markets. Bid 30, offer 70. The mid of 50 tells you nothing except that nobody knows.
A practical rule
Report the mid. Trade against the side you will cross. Never compare a last trade on one venue with a mid on another — that is a unit error dressed up as analysis, and it manufactures gaps that do not exist.
This last point matters a great deal for cross-venue work. If Kalshi's last trade is 58 and Polymarket's mid is 64, the apparent six-point gap may be entirely explained by the Kalshi print being from yesterday. This site compares like with like and discounts quotes by their age; stale quotes and false gaps is the guide to that whole failure mode.
What "the price" means on this site
Every probability shown on these boards is derived from the live book where one is available, not from the last print, and every comparison between venues uses the same convention on both sides. Where a quote is old enough to be doubtful, the market is downweighted rather than shown as a headline gap.
That is a deliberately conservative choice. It means this site shows fewer dramatic disagreements than a naive last-trade comparison would, and the ones it does show are more likely to be real.
Next
Why prices near 1c and 99c behave differently is where all of these effects get strangest.
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 readThe Favourite-Longshot Bias, and Where It Shows UpAcross a century of betting and event markets, long shots are systematically overpriced and favourites slightly underpriced. Here is why, and where to expect it.
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 readDecimal, Fractional and American Odds, ConvertedFour ways of writing the same number, with the conversions in both directions and the reason converting everything to probability first is the only sane habit.
Risk and Costs · 3 min readWhen Not to TradeThe default action in a market with real costs is no action. A checklist of the conditions under which the correct response to an interesting number is nothing.