The word average covers two different calculations, and in market data they often disagree. The mean adds everything and divides by the count. The median is the middle value when the numbers are lined up in order. Which one you are shown changes what you believe.
When they agree
If values cluster evenly around a center, the mean and the median land close together. Daily percentage moves of a large index on ordinary days are roughly like that. In that case either one is a fair summary and the choice does not matter much.
When they part
Many market quantities are lopsided. A few values are enormous and most are small. Market caps are like this. So are pool depths, trade sizes and spreads on a thin list. In a lopsided set the mean gets pulled toward the giants. The median stays with the typical member.
Take the spreads on Robinhood’s stock tokens on Sunday 4 October 2026, with the stock market shut. The median spread across the list was 4%. About a third of the tokens had spreads over 10%, and a few were far wider. A mean of that list would be dragged up by the extreme quotes and would describe almost no actual token. The median says what the middle token looked like. A wide quote is not a price has the detail.
One outlier can own a mean
The mean has no defense against a single bad value. One mistyped print, one quote with a stray side, one token priced at many times its value, and the average of the whole list moves. The median barely notices, because one value going from large to absurd does not change which value sits in the middle.
That makes the median the safer default for anything built from raw feeds, where a bad row is a matter of time.
When the mean is the right one
The mean is right when the total matters. Profit and loss is the standard case. A trader’s result is the sum of every trade, so the average outcome per trade has to be a mean. A median trade can be a small gain while the mean is a loss, if a few large losers outweigh many small winners. Expectancy is a mean for exactly this reason.
Funding averaged over time is also a mean, because what a holder pays is the sum of every hourly charge.
What to check on any dashboard
Three questions. Which average is this. What was it taken over. Is the set lopsided. A mean of a lopsided set should come with the median beside it, or with the largest values named. An average over a short record should say how short, which is the rule in a percent change needs a base and a window.
The two averages together also tell you about shape. A mean far above the median says a few large values dominate. That is worth knowing before reading any single row.
Where NoVo shows it
On the Crypto Market Map, the Stocks On-Chain tab marks a token price that is not the live midpoint with an approximately sign, so a wide quote cannot quietly distort the figures built on it.