The first six hours: liquidity in a weather market's life
A temperature market opens about fifty six hours before it closes. Volume figures make these look like the busiest thing on the board, and in aggregate they are. What the daily total hides is that a market spends its first hours essentially untradeable, and that the moment it becomes tradeable is different for every city. We sampled live order books across a full market lifecycle to find out when the ladders actually firm up.
What opening looks like
An hour after a batch of temperature markets opened, we pulled the top of book on the leading rung of each. The median spread across twelve cities was fourteen cents.
The distribution was the interesting part. Three cities sat at three cents or tighter and were perfectly tradeable. Four sat between four and sixteen. The remaining five ranged from twenty four to thirty five cents, with resting size on the leading rung in the single digits of shares.
A thirty five cent spread on a rung priced around forty cents is not a market. It is two people who have not met yet. Any order you send crosses most of the way to the other side of the distribution, and the price history that later gets recorded will not reflect what you paid.
The tightening is fast, then it stops
Sampling the same markets again over the following hours, most of them collapsed quickly. Two cities went from thirty and thirty five cents down to two cents inside about ninety minutes. Another went from sixteen to one. By the time the markets were roughly a day old, the median across all cities was one cent, and it stayed near there through the rest of the lifecycle.
But some never tightened. Two cities were still sitting at twenty four to thirty cents hours later, unchanged, while their neighbours on the same ladder had long since firmed up. Those are not markets that are waiting for information. They are markets nobody is quoting.
City matters more than clock
The natural theory is that spreads follow the local trading day: a city's market tightens when the people who care about that city wake up. We tested it and it does not hold.
At one sampling instant, two European cities on the same ladder and the same local hour showed four cents and thirty five cents. Two Asian cities at the same local early afternoon showed one cent and thirty cents. Local time explained very little.
What separates them appears to be whether a market maker has taken an interest in that particular city, which is a persistent property rather than a time-of-day effect. Some cities are quoted from the first hour. Some are barely quoted at all.
The practical implication is that any rule of the form "wait N hours after open" will systematically miss the cities that firm up late, and will overpay in the cities that never firm up at all. The workable rule is a spread threshold, checked repeatedly, per city.
Why this breaks naive analysis
Most published analysis of these markets uses price history, because price history is free and complete. Price history records trades. It does not record what was resting on either side at that moment.
If you build a study that says entering N hours before close historically returned some percentage, you have implicitly assumed you transacted at the recorded price. In the tight part of the lifecycle that is roughly fair. In the first hours it is fiction, and the first hours are exactly where the apparent returns tend to look largest, because thin markets are also mispriced markets.
Subtracting a realistic cost changes conclusions. Applying a measured one to two cent cost across a full sweep of entry times cut returns by several percentage points everywhere, and cut them most in the early window where the spread is widest. Some entry times that looked attractive on paper stop being attractive once you pay to get in.
Depth, not just spread
Spread tells you the cost of crossing. Depth tells you whether you can cross at all.
In the mature part of the lifecycle, the leading rung typically carried well over a hundred shares on each side, which comfortably absorbs retail-sized orders. In the first hour we saw leading rungs with five or six shares resting. At that depth a modest order walks up several price levels and the effective cost is far worse than the quoted spread suggests.
If you are sizing anything beyond a few dollars, the number to look at is resting size on the rung you want, not the headline spread.
What this means in practice
Three things follow from the measurements.
Do not trade the first hour. Whatever edge you think you have will be smaller than the spread you pay to express it.
Use a threshold, not a timer. Check the spread on the rung you want and act when it is acceptable, not when the clock says so. The cities differ too much for a schedule to work.
Accept that you will miss some cities. A market that never tightens is a market you should not trade, and skipping it costs you nothing except a trade you would have regretted.
Frequently asked
Is the wide spread an opportunity? Occasionally, for a patient limit order, but the same thinness that creates the spread means your resting order only fills when someone is motivated to hit it, which is not usually a good sign for you.
Do the big-volume cities always have tight spreads? Not reliably at open. Daily volume is dominated by the mature part of the lifecycle, so a city can carry large totals and still be untradeable in its first hours.
How close to expiry do spreads stay tight? They stay tight, but prices rise as uncertainty resolves. Near close the leading rung is frequently correct and correspondingly expensive, so tight spreads do not by themselves make late entry attractive.
Where to check depth
Top-of-book alone is not enough to size an order. SmartX shows the Polymarket temperature ladders with resting size per rung, which is the number that actually determines what you pay.