The two tables
The first moves the temperature and leaves everything else as you set it; the second moves the altitude. The highlighted row is the one matching your conditions. In the altitude table, a green number means the altitude works in your favour.
By temperature
| Temperature | WBGT | Effect | Adjusted pace | Time |
|---|
By altitude
| Altitude | Air density | VO₂ max | Effect | Adjusted pace |
|---|
The altitude table isolates the altitude effect: it leaves the heat out, because in the mountains the two rarely coincide at their worst. In practice, climbing 1000 metres usually takes 6 to 7 degrees off the temperature.
How this is worked out
These are two separate problems that are almost always explained together and badly. Heat limits you because your body has to split cardiac output between the working muscles and the skin, and because when sweat does not evaporate your core temperature climbs until your nervous system makes you slow down. Altitude limits you because there is less oxygen in every litre of air and your VO₂ max drops.
They are worked out separately and then multiplied, not added. A 5 % loss from heat and a 5 % loss from altitude are not 10 %, they are 10.25 %.
WBGT, not the temperature
The temperature your phone shows says very little on its own. What actually predicts how much you will suffer is the WBGT, the wet bulb globe temperature, which is the standard in sports medicine and the one federations use to call races off.
This calculator estimates it as 0.7 × the wet bulb temperature + 0.3 × the dry bulb temperature, plus a solar radiation loading of 1.5 degrees in partial sun and 3 in full sun. The wet bulb comes from Stull's approximation, which is reliable between −20 and 50 degrees and between 5 and 99 % humidity.
The reference scale, to place yourself: below 10 is ideal racing weather. Between 18 and 23 you should already be moderating your target. Above 28 many organisations shorten or postpone events, and above 32 they cancel them.
Why humidity weighs more than degrees
Here is the point almost no calculator explains. Compare two days at the same 30 degrees:
- 30 °C at 25 % humidity, full sun: WBGT 24. A half marathon costs you around 5 % more.
- 30 °C at 70 % humidity, full sun: WBGT 30. The same half costs you over 10 %.
Twice the penalty at the same temperature. The reason is simple: your only serious cooling mechanism is sweat evaporation, and sweat does not evaporate into air that is already full of water. In high humidity you can be soaked and not be cooling at all. That is why the wet bulb carries 70 % of the index and the dry temperature only 30 %.
The practical consequence: 28 dry degrees in Madrid in June are far more manageable than 26 humid ones in Ecuador or south-east Asia, whatever the thermometer says.
Altitude: what you lose and what the thin air gives back
Below about 500 metres, nothing happens. Between 500 and 1500 the drop in VO₂ max is gentle, on the order of 3 % per 1000 metres. Above 1500 it accelerates to about 6.5 points per 1000 metres for someone arriving unacclimatised.
But that is not the whole story, because thin air also slows you down less. At 2000 metres the air is 18 % less dense, and in running that gives back a small part of what it took: air resistance is around 2 % of the cost of running at distance pace, so you recover tenths, not points.
The historical check is the 1968 Mexico City Games, at 2240 metres. The distance events were run 6 to 7 % below expectations — the 10 000 m was won in 29:27 when the world record stood at 27:39 — while the sprints flew: the 100 m, the 200 m and the long jump produced marks that took years to beat. In events under two minutes, the aerodynamic advantage beats the loss of oxygen. This calculator is built for endurance events, so do not use it for a 400.
On a bike, altitude changes sign
This is the most interesting case and the one almost nobody works out properly. In cycling, air resistance grows with the cube of speed and dominates everything else above about 30 km/h. So at altitude two things happen at once: you produce fewer watts, but the air slows you down much less. And the second one wins.
With 280 watts, 80 kilos and a CdA of 0.32, the model gives this:
| Terrain | At 1000 m | At 2000 m | At 3000 m |
|---|
| Flat | 3.2 % faster | 5.1 % faster | 6.2 % faster |
| False flat, 2 % | 1.8 % faster | 1.6 % faster | about level |
| 5 % climb | level | 3.2 % slower | 8.8 % slower |
| 8 % climb | 1.0 % slower | 5.3 % slower | 12.4 % slower |
The crossover sits at around 2 % gradient. Below it the air rules and altitude hands you free speed; above it gravity rules, and gravity knows nothing about air density, so all that is left is the lost watts.
This is not theory: it is why every serious hour record attempt is made at altitude. With a specialist's numbers — 400 watts, a CdA of 0.20 — the model gives 51.7 km/h at sea level and around 54.5 at the 1900 metres of Aguascalientes: almost six percentage points for free. The velodromes of Aguascalientes and Mexico City are not famous by accident.
If what you want is to fine tune the watts of a specific climb, that is in the climbing watts calculator, and you can estimate your CdA from a photo in the aerodynamics one.
Acclimatisation works, and the timelines differ
To heat you adapt fast. In 10 to 14 days of daily exposure of about an hour, the body increases plasma volume, starts sweating earlier and with less salt, and drops heart rate at the same intensity. That recovers a little under half of the lost performance, which is what the selector above applies. The adaptation is lost fast too: two or three weeks without heat and most of it is gone.
To altitude it is slower and never complete. The first responses — hyperventilation, higher heart rate — appear within hours. The one that matters, the increase in red blood cells, takes two to three weeks and only gets you halfway: however well you acclimatise, at 2500 metres you will not run like at sea level.
A practical scheduling note: if you race at altitude without being able to spend weeks there, the least bad option is usually to arrive on the day or the day before and race before the insomnia, dehydration and loss of appetite of the first 48 to 72 hours set in, which is when you feel at your worst.
What this model cannot see
It does not see wind, which on a bike can outweigh everything above put together. It does not see radiation reflected off the tarmac, which on a black road at midday is far more than the three degrees the sun selector adds. It does not see your history: someone raised in Cuenca at 2500 metres does not respond like someone arriving from Belgium.
Nor does it see the thing that decides many hot races, which is hydration and salt. A pace adjustment will not save you if you lose two litres an hour and replace nothing: that is planned separately, in the session plan calculator.
And a warning that is not about performance: above WBGT 28 the risk of heat stroke stops being theoretical, and above 32 racing is a bad idea however well you feel. The symptoms — confusion, no longer sweating, chills in the heat — appear late, and sometimes the person having them is the last to notice.
Frequently asked questions
How much slower do you run in the heat?
It depends above all on humidity and duration. At a WBGT of 15 to 20 a marathon costs 3 or 4 % more; between 20 and 25, between 5 and 7 %. A short event suffers considerably less than a long one, because heat accumulates: at 30 degrees with humidity a 10 km loses around 7 % and a half marathon more than 10 %.
What is the ideal temperature to race a marathon?
Between 5 and 12 degrees, dry and overcast. Below a WBGT of 10 the effect is essentially nil, which is why records are set in Berlin, London or Valencia in autumn and spring, and never at midday in summer.
Why is humid heat worse than dry heat?
Because the way you cool down is by evaporating sweat, and sweat does not evaporate into air already saturated with water. In high humidity you sweat as much or more, but that sweat drips onto the road without cooling you: you lose water and salt and get nothing back. That is why the WBGT index gives 70 % of the weight to the wet bulb and only 30 % to the dry temperature.
How much performance do you lose at altitude?
Below 500 metres, none. Up to 1500, around 3 % of VO₂ max per 1000 metres. Above that, about 6.5 points per 1000 metres if you arrive unacclimatised. In endurance race time that comes out slightly lower, because thin air also slows you down less: around 6 or 7 % at the 2240 metres of Mexico City, which is exactly what was seen at the 1968 Games.
Do you go faster on a bike at altitude?
On the flat yes, and by a good margin. You lose watts because there is less oxygen, but air resistance falls faster than your power does: at 2000 metres the air is 18 % less dense. The net result on the flat is around 5 % faster. As soon as the road tilts above 2 % it reverses, because gravity does not care about air density: on an 8 % climb at 2000 metres you are more than 5 % slower.
Why are hour records set at altitude?
For the same reason. It is a flat event, on a track, where nearly all the power goes into pushing air. With a specialist's numbers, the gain between sea level and the 1900 metres of Aguascalientes is around 5 or 6 %, far more than the power that is lost. The optimum sits somewhere between about 1800 and 2500 metres: higher up, the lost watts and the difficulty sleeping and recovering start eating the advantage.
How long does heat acclimatisation take?
10 to 14 days of daily exposure, with sessions of about an hour in conditions like the race. You recover a little under half of what the heat takes from you. It is lost just as fast: two or three weeks in cool weather and you are almost back to the start, so it is worth keeping a heat session going into race week.
When should I arrive if I am racing at altitude?
Either very early or very late. Real acclimatisation — the rise in red blood cells — takes two to three weeks, so if you can be there that long, be there. If you cannot, it is almost always better to arrive on the day or the day before and race before the insomnia, dehydration and loss of appetite of the first 48 to 72 hours set in, which is the worst possible window.
At what WBGT is a race cancelled?
It depends on the organisation, but the usual reference is a flag scale: above 28 an event is shortened, postponed or flagged as high risk, and above 32 it is cancelled. Those are WBGT figures, not what your phone says: a WBGT of 32 might be 33 degrees with moderate humidity and sun.
