Millions of dollars trade on Polymarket temperature markets every month. Most of that money is bet using a number that won't exist by tomorrow morning.
The temperature you're watching on Weather Underground right now — the one driving your position, your confidence, your sizing — isn't the number Polymarket uses to settle. It never was.
Most Traders Have Never Read the Rules
Here's something that should scare you: the majority of people trading Polymarket temperature contracts have never actually read the resolution rules.
I don't mean they skimmed them. I mean they've never opened the rules tab. They see "highest temperature in Miami," assume their weather app tells them what that is, and trade accordingly.
Let's fix that right now. Here are the actual rules, copied verbatim from a Polymarket temperature market:
"This market will resolve to the temperature range that contains the highest temperature recorded at the Miami Intl Airport Station in degrees Fahrenheit on [date]."
"The resolution source for this market will be information from Wunderground, specifically the highest temperature recorded for all times on this day by the Forecast for the Miami Intl Airport Station once information is finalized, available here: https://www.wunderground.com/history/daily/us/fl/miami/KMIA."
Two things to notice:
- Resolution uses WU's historical record, not the live display. The phrase "once information is finalized" is doing all the heavy lifting.
- It's whole degrees only. The rules explicitly state: "The resolution source for this market measures temperatures to whole degrees Fahrenheit (eg, 21°F). Thus, this is the level of precision that will be used."
That word "finalized" is where the money hides. Because WU's live data and WU's finalized historical data are two completely different things.
This is exactly why MinuteTemp exists. We show you the data source that actually matches settlement — not the live WU number that vanishes overnight.
Two Pipelines, One Website
Here's the part that breaks most traders' mental model of how weather data works.
Weather Underground isn't one data feed. It's two completely separate pipelines displayed on the same website under the same branding. And they often disagree.
Pipeline 1: The Settlement Path
ASOS Sensor (Airport) → Hourly METAR Report → WU Historical Record → Polymarket Settlement
This is the path that matters. ASOS (Automated Surface Observing System) sensors sit at major airports — they're the official FAA/NWS weather instruments. They generate METAR reports, which are the global standard for aviation weather. These METARs get published every hour. At end of day, Weather Underground builds its historical record from these METAR observations.
This historical record is what Polymarket reads to settle your bet.
Pipeline 2: What You're Actually Watching
Multiple Sources (ASOS + Personal Weather Stations + other) → WU Live Display
WU's live "Current Conditions" page aggregates readings from many sources: the official ASOS sensor, yes, but also personal weather stations mounted on people's rooftops, third-party sensor networks, and other unofficial sources. These personal stations haven't been calibrated to ASOS standards. They're sitting on hot asphalt, next to HVAC units, in urban heat islands.
They run hot. Consistently. Often 2-5°F above the airport sensor.
The gap in plain English: The number WU shows you right now often comes from a rooftop sensor two miles from the airport. The number Polymarket settles on comes from a calibrated instrument on the tarmac. They are not the same.
MinuteTemp tracks the METAR source (Pipeline 1) and shows you the running daily high as it updates throughout the day. When WU live says 87°F and the actual METAR-sourced high is 84°F, MinuteTemp shows you both so you can see the divergence in real time.
Ghost Temps in the Wild
Here's what a ghost temp looks like in practice.
It's mid-afternoon. The sun is blazing. WU's live reading for KMIA shows 87°F. You look at the Polymarket "84°F-85°F" bracket — it's priced at 94 cents. Feels like free money. The airport is clearly above 84.
But here's what's actually happening:
- The 87°F reading is coming from a WU live feed
- The most recent METAR from the airport's actual instrument shows 84°F
- By end of day, WU will build its historical record from the METAR observations
- The 87°F ghost will simply... disappear. It will never appear in the historical table.
That 94-cent bracket? It will be the settlement if there isn't a METAR report with the 87°F. But imagine you're in the "86-87°F" bracket, priced at 80 cents because WU live is showing exactly that range. Settlement comes in at 84°F and you're in the wrong bracket entirely.
The market prices the ghost. Other traders see 87°F on WU, drive the "86-87" bracket up. If you know the METAR shows 84°F, you know that bracket is overpriced. That's the edge.
And it repeats. Day after day, market after market, city after city.
MinuteTemp was built for exactly this scenario. Instead of manually decoding METAR reports or hoping WU live matches settlement, you see the actual running high from the settlement-grade data source — updated throughout the day as new METARs come in.
The °F / °C Trap
Here's a wrinkle most traders miss entirely: not all Polymarket temperature markets use the same unit.
Miami settles in whole degrees Fahrenheit:
"measures temperatures to whole degrees Fahrenheit (eg, 21°F)"
London settles in whole degrees Celsius:
"measures temperatures to whole degrees Celsius (eg, 9°C)"
This matters more than you think. Here's why:
1°C = 1.8°F. So when London's market rounds to the nearest whole °C, each bracket effectively spans a 1.8°F range. Miami's brackets span just 1°F. The rounding impact is nearly double.
Consider this: A true temperature of 15.4°C rounds to 15°C. A true temperature of 15.5°C rounds to 16°C. That 0.1°C difference — barely perceptible in the real world — moves you into a completely different settlement bracket.
Now combine this with ghost temps. If a personal weather station near London City Airport is reading 1-2°C high, and the market brackets are single degrees, you're potentially watching a reading that's two brackets away from where settlement will land.
What Settles Is What Settles — Even When It's Wrong
There's one more rule buried in the resolution text that most traders overlook:
"Any revisions to temperatures recorded after data is finalized for this market's timeframe will not be considered for this market's resolution."
Read that again.
If Weather Underground posts an incorrect historical value — a data glitch, a processing error, a sensor malfunction that makes it into the finalized record — Polymarket still settles on it. There's no appeal process. No "well the real temperature was..." correction. The WU historical page is the oracle. Whatever it says after finalization is the truth, even if it's wrong.
This means the settlement source isn't really "temperature." It's "whatever number appears on a specific webpage after a specific time." That's a subtle but important distinction.
You can't predict WU glitches. But you can watch the METAR data that feeds the historical record, know what number should appear, and trade with confidence — or know when something looks off.
MinuteTemp shows you the raw METAR data in real time, so you can see exactly what readings are flowing into the pipeline before WU finalizes its historical page. If there's a discrepancy between what the METARs show and what WU eventually posts, you'll see it developing — not after the fact.
Your Edge: How to Actually Trade This
Enough theory. Here's how to put this into practice.
1. Stop watching WU live
Seriously. Close the tab. The "Current Conditions" page is not the settlement source. Every minute you spend watching it is building a position on bad data. Use MinuteTemp instead — it shows you the METAR-sourced running daily high, which is what WU's historical record will be built from.
2. Watch METAR data throughout the day
METARs are published every hour (typically at :53 past the hour UTC) for all major airport stations. The running maximum temperature from these reports is your best predictor of the final settlement value.
You can get raw METARs from aviationweather.gov, but the format looks like this:
METAR KMIA 171753Z 18010KT 10SM FEW045 SCT250 29/21 A3001 RMK AO2
SLP163 T02940211 10294 20228 58010
See that T02940211? That's the temperature encoded at 0.1°C precision. 0294 = 29.4°C = 84.9°F. WU will round this to 85°F in the historical record.
Unless you want to learn METAR decoding, MinuteTemp does this parsing for you automatically. We decode every METAR as it's published, extract the precise temperature, convert to the correct unit for each market, and show you the running daily high — updated hourly as new observations arrive.
3. Fade the ghost in the last 2 hours
Ghost temps are most dangerous — and most profitable to fade — in the last few hours of the trading day. That's when:
- The sun is hottest, amplifying the gap between urban personal stations and the airport ASOS
- Markets are most liquid (more traders watching, more volume)
- WU live divergence from METAR peaks
If WU live shows 87°F and MinuteTemp shows the METAR high is 84°F, the "86-87" bracket is likely overpriced. The "84-85" bracket is likely underpriced. That's your trade.
4. Check the market's unit before trading
Miami → °F. London → °C. Munich → °C. Always confirm which unit system the specific market uses. The bracket width and rounding behavior change everything.
5. International markets have the same problem
WU uses the same dual-pipeline architecture for every station worldwide. KMIA (Miami), EGLC (London City), EDDM (Munich), LFPG (Paris), SBGR (Sao Paulo) — same ghost temp risk, different stations, different timezones.
In fact, international markets can be even more prone to ghost temps because personal weather station density varies wildly by region. A station in central London might be 3°C above Heathrow's ASOS.
MinuteTemp covers all Polymarket temperature stations — domestic and international. Same METAR-sourced data, same real-time running high, same settlement-grade accuracy. Check your market's station now.
The Ghost Temp Checklist
Bookmark this. Check it before every trade.
Signs you're looking at a ghost temp:
- WU live is 2°F+ (or 1°C+) above the most recent METAR reading
- WU is showing an unusual spike that the hourly METAR didn't confirm
- WU "Current Conditions" references a nearby personal weather station, not the airport ICAO code
- The reading feels surprisingly high or low compared to the forecast
- Other traders in chat are citing a number that doesn't match your METAR source
How to verify settlement data:
- Go to the WU history page for the specific station (e.g.,
wunderground.com/history/daily/us/fl/miami/KMIA) - Look at the observation table — these are the METAR-sourced values that settle
- Compare against the METAR running high on MinuteTemp to see if they align
- If WU live and WU historical diverge, trust historical — that's what settles
The Uncomfortable Truth
Here's what bugs me most about this.
The ghost temp problem isn't a bug in Polymarket. The rules are clear — they tell you exactly what the settlement source is. The problem is that nobody reads the rules, and the most natural thing to do — check Weather Underground for the current temperature — gives you the wrong number.
The entire market is structured around a data source that 90% of participants are reading incorrectly. The live WU display and the historical settlement record share a URL, share a brand, and look like the same product. But they're fundamentally different data.
Every day, traders are pricing brackets based on ghost temps that will vanish by morning. If you're one of the few who knows the difference, you don't need to be smarter than the market. You just need to be reading the right number.
That's the entire edge. Not a model. Not an algorithm. Just reading the right data source.