Does Rain Actually Slow You Down in a Marathon?
We checked 1,686 marathon race days for a rain penalty. Cold rain costs a little, warm rain costs nothing, and neither is the real risk.
The forecast flips to a rain icon two weeks out and the questions start: which shoes, whether to bring a throwaway layer, whether to just brace for a bad day. Rain is one of the two conditions runners worry about most before a race, right behind heat.
We went back to the same dataset behind our RunScore rebuild, 6.3 million finish times across 1,686 marathon editions from 2000 through 2026, and asked a narrower question: how much does rain, by itself, actually cost a marathoner? The answer is less than most runners assume, and less than our own score used to assume.
Why rain looked like a bigger deal than it is
Runners' complaints about rain are mostly complaints about being cold, and cold rain and warm rain aren't the same problem. Cold rain adds weight to shoes and clothing, chills skin, and can tip a marginal day into a genuinely dangerous one. Warm rain does some of that too, but it also cools a runner down on a day that would otherwise run hot, and that cooling appears to roughly cancel the friction and weight cost. That's why our model treats the two separately, and why we split the data the same way: rain on race days below 55°F counted as "cold," everything else as "warm."
What 1,686 race-years actually show
Of those 1,686 race-years, 177 had measurable rain during the hours runners were actually out on the course, a figure we track separately from a day's total rainfall (more on that below). One hundred were cold, 77 were warm. Only 25 saw more than a quarter inch. It isn't a large wet-weather sample, and the results reflect that: the cold-rain estimate is real enough to keep charging for, not sharp enough to be confident in.
Across the full sample, cold rain works out to about 1.1% slower per inch, with a 95% range running from roughly -2.3% to +4.5%. Statistically, we can't rule out that cold rain does nothing at all. Warm rain came back negative, about -1.6% per inch (range roughly -4.0% to +0.9%), meaning if there's any effect, it leans toward wet warm days running a touch faster, not slower. Neither number is one we'd bet a field's finish times on individually. But the direction, cold rain costs something and warm rain doesn't, held up as we added more races, which is more than a raw magnitude alone can offer.
The number that didn't survive more data
An earlier pass at this, done before the RunScore rebuild, used just 13 races and found a cold-rain penalty of 8.3% per inch, several times higher than either what the full sample now estimates or what the shipped score charges. That was a small, spring-heavy sample, and part of the effect turned out to be a results bug: some old files stored gun time instead of chip time, which inflates results in a big wave start. Rebuilding the same 13 races on the fixed pipeline cut the estimate to 4.6%. Adding 79 more races, and about 1,409 more race-years, cut it again, to about 1%, with a range wide enough to include zero.
Cold rain and warm rain are opposite stories: one adds a small, uncertain cost, and the other appears to cancel itself out. Both are minor next to heat or a hard headwind.
Does the forecast's rain chance matter?
We also tested a different question: once you know how much rain actually fell, does the forecast's percent chance of rain add anything on top? We checked two versions of "chance," the stored forecast probability and the share of race-window hours that were actually wet. Both came back essentially at zero, around -0.4% per 100 percentage points, with actual rain controlled for. A 90% chance of rain that produces a drizzle costs what the drizzle costs. The forecast percentage itself isn't doing any of the work.
The race everyone points to, and the one nobody does
Boston 2018 is the race that comes up whenever this topic does: temperatures in the high 30s and low 40s, driving rain, gusts to 32 mph, and more than 2,500 runners treated in medical tents. Forecasters were calling for up to two inches of rain across the Boston area that day (WBUR, Boston Magazine). Our own number for that day is smaller, and measures something different: 0.36 inches of rain during the hours the field was actually racing, not the full day's storm. Cold enough to land in our cold-rain bucket, but nowhere near the wettest race-window we have on record.
That belongs to a race nobody brings up: Catalina Island Marathon 2003, at 1.09 inches while runners were on course, the most of any of the 1,686 race-years in our sample. Marquette Marathon 2014 (0.71 inches) and Blue Ridge Marathon 2011 (0.63 inches) are next. All three are small races, which is exactly why we don't lean on any single one to say what rain does generally. It's the same caution that kept an early 13-race sample from being the final word on cold rain.
What we ship, and why it's cautious
RunScore charges 3% per inch of rain at 50°F and below, tapering to 1% per inch at 60°F and up. Measured against the data, that's on the high side for cold rain, and not zero for warm rain, when the data can't rule out zero being the right number there. We ship it that way on purpose. While the estimate stays this noisy, we'd rather the score run a little heavy on rain than a little light.
So does rain matter?
Probably less than you think, and probably not zero. Cold rain looks like it costs something, a few percent per inch at most, though we can't be fully sure it costs anything. Warm rain, by the numbers, doesn't cost you a thing. That isn't the same as saying rain is free. It's saying that on the one thing we can measure, how much slower the field ran, rain barely moves the needle next to heat or a hard headwind.
What rain actually threatens isn't your finish time. The real danger in a cold, wet race is a dropping core body temperature late in the run, when a tiring body is producing less heat and moving slower through hours of wet, windy exposure. Boston 2018's median finisher lost about 3% to the conditions. The medical tents that day were full anyway. A pace-cost number, ours included, has nothing to say about that risk, which is why we treat it separately: our guides on what to wear by temperature and race-day gear, and our piece on running a cold marathon, cover the hypothermia warning signs and what to do about them. If it's cold and wet on race morning, dress for the cold. The pace will probably take care of itself.
More guides
- Why We Rebuilt RunScore From 6.3 Million Marathon Finishes — the full calibration this post pulls from
- How to Run a Cold Marathon: The Condition Runners Underrate — why cold is fast and how to avoid hypothermia
- What to Wear on Race Day: The Complete Guide by Temperature — layering for a wet, cold morning
- Race Day Gear Guide — gear picks by temperature, including rain-specific calls
- Find your race — the live RunScore and rain forecast for your next start line