- 7-day aggregateWindow
- One requestCost
- Mid +
spread rangeReturns - Per UP /
DOWN sideBreakdown
Sometimes you do not want the path a market took, you want its shape. How busy was it, where did the probability sit, how wide did the quotes get. The /v1/markets/:id/summary endpoint answers all three in a single call by folding seven days of snapshots into a compact set of aggregates, so you can size up a market before deciding whether the raw history is even worth pulling.
A summary is a triage tool. It reads the same captured snapshots a time-series query would, but instead of handing you every row it returns the counts and the min / avg / max of the fields that matter, mid price and spread, across the trailing seven days, split by the UP and DOWN side. One request, one picture.
What the summary returns
Activity
The snapshot count and the first / last seen timestamps tell you how alive the market was over the window.
- Snapshot count
- First / last seen
- Coverage at a glance
Where probability sat
Average, minimum, and maximum mid price bound the range the market’s implied probability travelled across the week.
- Avg / min / max mid
- Implied-probability band
- 0-1 units
How tight it traded
Average, minimum, and maximum spread show whether quotes stayed tight or blew out, a fast read on liquidity quality.
- Avg / min / max spread
- Liquidity proxy
- Per-side split
Because each market carries two tokens, the breakdown is per side, UP and DOWN reported separately. That matters: a market can look balanced on one side and thin on the other, and an aggregate that blended them would hide it.
When the aggregate beats raw snapshots
Pulling raw snapshots is the right move when you care about the sequence, the path, the timing, the moment something moved. The summary is the right move when you care about the envelope. Those are different questions, and the cheaper one answers a surprising number of them.
- You want to know if a market is liquid enough to bother with before committing a paginated pull, average spread and snapshot count answer that in one call.
- You are ranking or filtering a list of markets and need one comparable number each, not thousands of rows apiece.
- You want the probability band a market traded in over the week without reconstructing the whole series.
- You are building a dashboard tile that shows a market’s recent envelope and refreshes cheaply.
Scanning many markets fast
The summary’s real power shows up at breadth. List the live markets, then call the summary once per market id, and you have a comparable seven-day envelope across the whole board, sortable by activity, by average spread, by how far the mid travelled, for the cost of one cheap request each rather than thousands of snapshot rows per market.
- 1List the markets you care about, a category, a timeframe, or the full live board.
- 2Resolve any human slug to its 0x conditionId so each summary call targets the right market.
- 3Call the summary once per market id and collect the per-side aggregates.
- 4Sort or threshold on whatever the screen needs, busiest, tightest, widest range, then pull raw snapshots only for the handful that survive the filter.
The fields behind the numbers
- snapshot countHow many captures in the window
- first /
last seenCoverage edges of the 7-day window - avg ·
min · max midThe implied-probability band - avg ·
min · max spreadQuote width, per UP / DOWN side
A summary, not the path
The aggregate tells you the range, not the route. A min and a max mid of 0.20 and 0.80 could be one slow drift or a dozen violent round trips, the summary cannot distinguish them. When the order of events matters, the summary is the screen, not the answer; pull the raw snapshots for the markets that pass it.
There are two other honest limits worth stating. The window is the trailing seven days, so it is a recent-shape read, not a full-history one, for deep history you query snapshots over your own range. And the spread aggregate is a quote-width statistic built from the book, a proxy for liquidity quality, not a record of what actually executed.
Ask for the path when you care how a market moved. Ask for the summary when you only need to know whether it is worth asking.
Try the summary endpoint
The docs cover the summary call and its per-side fields; the crypto page shows which markets you can summarise.
Frequently asked questions
What does the market summary endpoint return?
A seven-day aggregate for a single market: the snapshot count, the first and last seen timestamps, and the average, minimum, and maximum of both mid price and spread, broken out per UP and DOWN side. It is one request that describes the market’s recent shape rather than its full tick-by-tick history.
When should I use the summary instead of pulling snapshots?
Use the summary when you care about the envelope, how busy a market was, what probability band it traded in, how tight its quotes were, and especially when you are screening many markets and need one comparable number each. Pull raw snapshots when you care about the sequence: the path, the timing, or the exact moment something moved.
Can I summarise lots of markets at once?
You call the summary once per market id, but each call is cheap, so listing the live markets and summarising each gives you a comparable seven-day envelope across the whole board for far less than pulling thousands of snapshot rows per market. Then you only pull raw history for the handful that pass your filter.
What can’t the summary tell me?
It cannot tell you the order of events. A min and max mid of 0.20 and 0.80 could be one slow drift or many round trips, the aggregate cannot distinguish them. It also covers only the trailing seven days, and its spread figures are a quote-width proxy for liquidity, not a record of executed trades. For path, timing, or deep history, query the raw snapshots.