How Temperature Affects Baseball: What 10,422 Games Actually Show

A baseball flying through air collides constantly with air molecules, and every collision slows it down slightly. Warmer air is less dense — the molecules are spread further apart — so a hot-night fly ball loses less speed and lands further from the bat. University of Illinois physicist Alan Nathan's trajectory work puts it at roughly 0.33 feet of extra distance per 1°F on a typical home-run fly ball, which works out to about 1% more home runs per degree (Nathan, 2023).

A 2023 Dartmouth study went further, attributing more than 500 home runs since 2010 to climate-driven warming across 100,000 games (Callahan et al., Bulletin of the American Meteorological Society). Cold cuts the other way twice: denser air adds drag, and a cold ball is also less bouncy off the bat, leaving it at lower exit speed.

Those are physics estimates. What follows is our own check against real MLB games — every open-air game since 2021 with a game-time temperature reading at the park itself.

The gradient is monotonic — every step warmer means more runs and more home runs. Sub-50°F games scored 0.99 fewer runs than 70–84°F games (t = −4.42) and produced 0.77 fewer home runs (t = −10.13). On the hot end, 85°F+ games beat mild games by 0.73 runs (t = 4.99) and 0.35 home runs (t = 6.21). All four bands have hundreds of games, so none of these is a small-sample artifact.

Two honest caveats. First, these are raw averages: hot games cluster in July and August afternoons at certain parks, and we have not adjusted for team strength, lineups or pitching. Second, the sample includes Coors Field, where low air density from elevation amplifies everything — part of why even the coldest-band numbers are carried up by Colorado's appearances.

There is no league temperature floor — games get played, and pitchers just grip harder. Since 2021, 58 open-air games started below 40°F and 380 below 50°F, nearly all in April. The extremes: 27.1°F at Coors Field on April 21, 2021 (8 runs — elevation beats cold), and 27.4°F at Progressive Field on April 7, 2026, which managed only 3 runs and 1 home run, exactly what the drag-plus-dead-ball story predicts.

Cold also changes the equipment: a cold ball is less elastic and leaves the bat slower, and cold composite and wooden bats lose trampoline effect. That is why early-season scoring sag matters more than any single frozen game.

Averaged across every open-air game since 2021 (parks with at least 150 games), Truist Park is the hottest venue at 78.5°F per game and Oracle Park the coldest at 62.1°F — a 16-degree gap that is a big part of why Atlanta out-homers San Francisco so reliably. Home-run leaders and laggers don't map onto temperature one-for-one, because dimensions and fences matter just as much:

Dodger Stadium is the sharpest example of heat without fences to blame: 72.6°F average and 2.83 home runs per game, the highest rate in the sample. Oracle Park sits at the other corner — coldest park, fewest home runs (1.85) — which is exactly where the marine layer story comes from. Busch Stadium is the interesting outlier: second-hottest park, near the bottom in home runs at 1.93, because deep fences can offset hot air.

Largely, yes. Cold games finished +0.29 runs over the closing total, cool games +0.35, mild games +0.49 and hot games +0.63. If the market ignored temperature, hot games should drift massively over the number — instead the band differences are statistically indistinguishable (hot vs mild: t = 0.97; cold vs mild: t = −0.89). Totals set for a 90°F afternoon in Texas are already higher than the same matchup on a 45°F night in Cleveland.

The hot band's +0.63 is the largest residual in the table, which hints heat may be very slightly underpriced — but with t under 1 and dozens of splits tested across this site, that is directional only. We publish nulls as nulls: there is no reliable temperature edge against the closing line here.

• April baseball: does the early-season scoring dip repeat? — a same-park, same-season check
• Coors Field, altitude and the humidor — the extreme version of thin air
• Marine layer baseball — the cold-night coastal version
• Yankee Stadium vs Citi Field — same sky, different ballpark
• The windiest MLB stadiums, ranked
• Wrigley Field wind: blowing out vs in
• MLB temperature and scoring study — the underlying research
• MLB weather today

All open-air MLB games (open-roof stadiums) in our MLB game conditions data, 2021–2026: 10,422 games, all with a game-time temperature reading taken at each park's own coordinates. Games in domes or with retractable roofs closed are excluded. Temperature bands: below 50°F (380 games), 50–69°F (3,431), 70–84°F (5,347), 85°F+ (1,264). Closing totals from our closing-line table (10,389 games with a line). Differences between bands tested with Welch's t-test; |t| under 1.96 is reported as not significant. Raw averages, not adjusted for team strength, schedule or humidity; the 2026 season is partial. Physics references: Nathan (2023), baseball.physics.illinois.edu; Callahan, Dominy, DeSilva & Mankin (2023), Bulletin of the American Meteorological Society. Data queried 12 October 2026. By the OverUnderWeather Research Desk.