MLB wind direction and scoring

Across 11,581 outdoor games at 30 parks from 2021 to 2026, direction matters and speed alone does not. Wind blowing out is worth +0.16 home runs a game against calm conditions, while wind blowing in costs -0.24 home runs and -0.34 runs. Suppression is the stronger half of the story: wind blowing in at 10–15 mph gives up -0.80 runs, and above 15 mph it gives up -1.01.

Out means the wind is travelling towards the outfield, in means it is coming back from the outfield, and across is everything in between — each judged against that park's own field bearing, not a compass point. The smaller figure in each cell is the season-adjusted gap against the calm control, with its t statistic. Anything under roughly 2 is noise.

This is where the effect earns its keep. Helping wind adds home runs at ordinary speeds, and above 15 mph that benefit disappears — the 361 games there are statistically indistinguishable from calm (t = -0.29 home runs, t = -0.75 runs). That is no benefit above roughly 15 mph, not evidence that strong out-wind actively hurts scoring. Hurting wind just keeps hurting, and accelerates.

Run the same games without the orientation adjustment and the signal mostly disappears into a vague, mild suppression — helping and hurting winds cancelling one another out. This is the gap we flagged in our own NFL wind study, and in baseball it is decisive: the direction split reveals a boost and a suppression that the speed-only view cannot see.

Repeating the direction comparison on dry games above 60°F, against a calm control filtered the same way, to check that wind is not just carrying rain and cold along with it. The home-run effect survives on both sides; the run-level suppression softens and slips below the significance guide.

Only parks with at least 40 games of wind blowing out and 40 of wind blowing in are shown, which is why heavily one-directional parks and the retractable-roof venues are absent. The figures are the season-adjusted gap between out-wind and in-wind games at that park.

These are single-park samples in the low hundreds, so treat them as texture rather than as prices. They are useful mainly for showing that the direction effect is not one famous ballpark carrying the league.

League scoring drifts year to year, so every figure above is a game measured against its own season's average rather than against the six-year pool.

• Coors Field orientation. We updated the park's field bearing from 35° to 0° (north) before publishing, because two independent published sources — Andrew Clem's stadium statistics and The Shadium — both list 0°, while an initial manual imagery read was ambiguous. This changed which games were classified as wind blowing in versus out at Coors.
• Minute Maid Park sample. Only 11 roof-open games were recorded there across six seasons, far below our 40-game minimum. It is included in the league-wide totals but excluded from the park-level breakdown, and no park-specific claim is made about it.
• Roof status. Every retractable-roof game in the sample has a recorded roof state; none were ambiguous. Fully enclosed parks are excluded entirely rather than treated as a calm control.
• Park-level table filtering. The park-by-park table only shows venues with at least 40 games of wind blowing out and 40 of wind blowing in, which is why several parks — including most retractable-roof venues and minor-league or international sites — do not appear.

• Home runs are the clean signal. Direction moves home runs decisively in both directions and moves total runs far less. If you want the mechanism, it is carry, not offence in general.
• Suppression beats assistance. Wind blowing in is worth more than wind blowing out, and it scales with speed; 359 games at 10–15 mph give up -0.80 runs.
• The helping wind's benefit disappears above 15 mph. Out wind above 15 mph shows no home-run edge over calm — statistically indistinguishable from the control, not a reversal. The claim is no benefit at that speed, not that strong out-wind punishes scoring.
• Crosswind sits near the noise threshold. Across-the-field wind is directionally negative in both runs and home runs, but with t statistics around 1.96 it is not clearly distinguishable from zero — do not treat it as a reliable small effect.
• Speed alone is close to useless. Without the park orientation, the whole thing flattens out. That is the practical takeaway for anyone reading a raw forecast number.

This study compares games with each other, not with the betting line. The follow-up test against the closing total found wind blowing out at 10+ mph gives no Over edge. See wind blowing out vs the betting line and the park-by-park wind sensitivity study.

• Sample: 11,581 outdoor MLB games, 2021–2026, at 30 parks with a resolved field bearing, a final score and a home-run count. Fully enclosed parks are excluded outright; retractable-roof parks contribute only their roof-open games. Neutral, minor-league and international-series sites are dropped.
• Weather: hourly archive reanalysis at the park's own coordinates, read across the game window from first pitch rather than off a single snapshot.
• Classification: wind is converted to the direction it is travelling, compared against the park's home-plate-to-centre-field bearing, and labelled out within 45° of that bearing, in beyond 135°, and across in between. Games under 5 mph are the calm control regardless of direction.
• Every figure is season-adjusted — each game against its own season's average — and the dry-and-mild table repeats the comparison against a control filtered the same way.
• Orientations were verified against published stadium orientation figures and cross-checked against aerial imagery before the analysis was run, not after. Coors Field was corrected to 0° (north) using Andrew Clem's stadium table and The Shadium, which both list 0°; an earlier stored value of 35° and an ambiguous manual imagery read were inconsistent with those published figures. Rogers Centre and Marlins Park had closed roofs in the available imagery and rest on the published figure.
• Minimum sample is 40 games per band, a threshold already in use across our earlier studies. Bands below it are flagged thin and greyed out.
• Not controlled: team quality, pitcher matchups, gusts as distinct from sustained wind, humidity and air density beyond temperature, altitude beyond the park-level view, and extra-inning games. In addition, the 361 high-wind-out games likely cluster at a handful of exposed parks, so some of the fade in that band may reflect which parks see that wind pattern most often rather than the physics at any given park.
• This study reports runs and home runs, not Over/Under results; those are tested against the closing total in the follow-up betting-line study.

Data generated 2026-10-28. Numbers refresh as the archive tops up; corrections are dated and kept visible.

• Open-Meteo ERA5 historical reanalysis — archive-api.open-meteo.com (hourly game-window weather at each venue’s own coordinates)
• MLB official game feed — mlb.com/scores (per-game conditions and retractable-roof state)
• ESPN MLB scores and odds — espn.com/mlb (schedules, final scores and closing totals)
• Published ballpark orientation figures (Andrew Clem's stadium tables and The Shadium), cross-checked against aerial imagery.
• This study has been corrected since first publication. Every dated change is listed on our corrections log.