How it works

Every TV station in Kansas City says its forecast is the one to trust. This site writes down what each one forecast, waits, and checks.

What's collected

Every 30 minutes, the current forecast from each source (see Sources): the four stations' weather teams and websites, the National Weather Service, and the main computer models. When a forecast changes, the new version is kept with the page or picture it came from, so every number here can be traced back. All forecasts are for Kansas City International Airport (KCI), where the official measurements are taken.

A station's weather team forecast is what its meteorologists put out themselves: KSHB 41's daily forecast article, KMBC 9's daily forecast article, KCTV5's on-air 7-day graphic, and FOX4's on-air 7-day graphic (its "Long Ranger", shown at the end of its forecast videos); the graphics are read by text recognition. A station's website forecast is the multi-day forecast on its site and app, supplied by a weather data company: at some stations the team sets it, at others it doesn't match what the team says on the air.

What it's checked against

• High: the official high for the day, from the National Weather Service's daily climate report for KCI.
• Low: the low on the morning of that day (a forecast for "Wednesday night: low 61" is Thursday morning's low). It's checked against the official low when that came overnight; on days when it came late in the evening (a cold front arriving), against the lowest temperature from 7 p.m. to 9 a.m. instead, since that's the low the forecast was for.
• Rain: 0.01 inch or more at KCI during the day.

How it's scored

Each forecaster is judged on the forecast it had out at 9 a.m. each day, for 1 to 7 days ahead, so everyone is judged at the same moment. "1 day out" is tomorrow's forecast.

• Miss: how many degrees the forecast high or low was off, averaged. 2.0° means it was off by two degrees on average.
• Within 3°: the share of forecasts that were 3 degrees or closer.
• Rain right: a chance of 50% or more counts as calling for rain. Right means it called rain and it rained, or didn't and it didn't.
• Brier score: a measure of how good the rain chances are, not just the calls. It's the average of (chance − what happened)², where what happened is 1 for rain and 0 for none. 0 is perfect; always saying 50% scores 0.25.

The home page ranks forecasters by their average miss for highs and lows together, 1 to 2 days out, over the last 30 days. A forecaster needs 5 scored days to be ranked.

Weather teams and websites: why both?

Each station puts out two different forecasts, and they often disagree:

• The weather team's forecast: what the station's meteorologists forecast themselves, on the air and in their daily forecast articles.
• The website's forecast: the multi-day forecast on the station's website and app, supplied by a weather data company. Some stations' meteorologists set it themselves; at others it's the company's automatic forecast.

On the day this site started (October 14, 2026), KSHB 41's weather team forecast 65° for the day, 61° that night and 69° the next day; its website said 66°, 62° and 68°. Scoring them together would blame or credit the team for a computer forecast it didn't make, or credit the station for its team when a viewer was really looking at the app. So they're scored separately, which also shows whether a station's meteorologists beat its own website.

Not every station publishes both in a form that can be read:

The websites aren't all alike. On the first day, KCTV5's website forecast matched its team's on-air graphic exactly for all seven days (it adds days 8 to 10 automatically), while KSHB 41's, KMBC 9's and FOX4's websites differed from their teams'.

FOX4's team forecast comes from its "FOX4 Forecast" videos: the last few seconds of each new one, where the Long Ranger graphic is shown, are read frame by frame, and a number is kept only when the frames agree. The video itself isn't kept; one frame is, as the record. A video that doesn't end with the graphic is skipped, and FOX4's previous forecast stays current.

Because KCTV5's website has matched its graphic, the two are compared every time they're read: a difference means either the graphic was misread or the station's two forecasts have parted, and it's noted either way.

Why stations, and not individual meteorologists?

A fair question, and we looked into it. Each station is scored as a whole, for four reasons:

• Most forecasts carry no name. KCTV5's 7-day graphic and FOX4's on-air forecast don't say who made them, and the station websites' forecasts are automatic. Only some daily articles have a byline, so any ranking of people would leave most of them out.
• A name on a forecast isn't necessarily who made it. At most stations the forecast is a team product: the chief meteorologist often sets it, and whoever is on shift presents it or writes it up.
• Too few days per person. Each meteorologist works only some shifts, so it would take a long time before one person's record meant much.
• Not everyone gets the same weather. Whoever works the stormy days and the sudden cold fronts would look worse through no fault of their own.

The station's forecast is what viewers actually get, whoever is on the air, so that's what's measured.

Fine print

• When a source can't be read (a site is down, or a page changes), that day is simply missing for it; no forecast is filled in or guessed.
• The official day for weather records runs midnight to midnight standard time, so during daylight time it's 1 a.m. to 1 a.m.
• Every forecast, score and day's weather can be downloaded from the Data page, to check the numbers.
• This site isn't affiliated with any station or with the National Weather Service. Forecasts are quoted only to compare them.