Difference between revisions of "Impact of Heat on Marathon Performance"

From Fellrnr.com, Running tips
Jump to: navigation, search
User:Fellrnr (User talk:Fellrnr | contribs)
User:Fellrnr (User talk:Fellrnr | contribs)
Line 3: Line 3:
 
The study showed how finish time changes with temperature, and the results are shown in the following graph.  
 
The study showed how finish time changes with temperature, and the results are shown in the following graph.  
  
http://jfsavage.smugmug.com/photos/817939120_uZxSb-X3.jpg
+
http://jfsavage.smugmug.com/photos/817939120_uZxSb-O.jpg
  
 
This graph works well for runners finishing the race at 3 hour pace or faster, but that excludes the majority of marathon runners. How can we extend this to slower runners?  
 
This graph works well for runners finishing the race at 3 hour pace or faster, but that excludes the majority of marathon runners. How can we extend this to slower runners?  
 
==The Answer(s), Part 2==
 
==The Answer(s), Part 2==
I’ve taken two approaches to extending the graph, as you can see below. The first approach is to assume that the decline in performance is linear with a zero origin, as shown by the red line. The second approach is to assume the decline is linear from the two noted end points, as shown by the other colored lines. This will give us two projections of performance decline and I’ve created tables at the end of this page based on these assumptions.  
+
The approach I've taken is to assume the decline is linear from the end points of each curve, as shown by the colored lines. This will give us a projection of performance decline and I’ve created a table at the end of this page based on this assumption.
 +
 
 +
http://jfsavage.smugmug.com/photos/818611486_iAHcm-O.jpg
 +
 
 
==Your Mileage May Vary==
 
==Your Mileage May Vary==
 
There are a lot of flaws in this approach, which you should be aware of.  
 
There are a lot of flaws in this approach, which you should be aware of.  
Line 15: Line 18:
 
* Extrapolating information is a flawed approach, as there is no way of knowing what the actual data would look like.  
 
* Extrapolating information is a flawed approach, as there is no way of knowing what the actual data would look like.  
 
==Usability of the data==
 
==Usability of the data==
Given all these problems, is this approach usable or useful? I think there is some use, as long as you understand the limitations. If the charts below say you can run a 4:48 marathon based on your cool weather performance and the projected temperature, this does not mean you can run a 4:48 in practice. What it does provide is a sense of just how much you will need to slow down on a hot day, and how important weather is in your results. Hopefully this information will cause you to reset your goals and promote a flexible approach on race day.  
+
Given all these problems, is this approach usable or useful? I believe there is value as long as you understand the limitations of the approach. If the charts below say you can run a 4:48 marathon based on your cool weather performance and the projected temperature, this does not mean you can run a 4:48 in practice. What it does provide is a sense of just how much you will need to slow down on a hot day, and how important weather is in your results. Hopefully this information will allow you to reset your goals and promote a flexible approach on race day.  
 
==Marathon Selection==
 
==Marathon Selection==
 
One key note from this analysis is that weather will play a huge role in your results. If you are seeking a race to achieve a specific time goal, such as Boston Qualification, choosing a race that has a high probability of cool or cold weather becomes important. You are probably better off with a hillier course than warm weather for instance.  
 
One key note from this analysis is that weather will play a huge role in your results. If you are seeking a race to achieve a specific time goal, such as Boston Qualification, choosing a race that has a high probability of cool or cold weather becomes important. You are probably better off with a hillier course than warm weather for instance.  
==Zero Origin Table==
+
==Projected Performance==
{| {{table}}
 
| align="center" style="background:#f0f0f0;"|'''40f'''
 
| align="center" style="background:#f0f0f0;"|'''50f'''
 
| align="center" style="background:#f0f0f0;"|'''60f'''
 
| align="center" style="background:#f0f0f0;"|'''70f'''
 
| align="center" style="background:#f0f0f0;"|'''80f'''
 
|-
 
| 3:00:00||3:05:24 (3%)||3:10:48 (6%)||3:16:12 (9%)||3:21:36 (12%)
 
|-
 
| 3:05:00||3:10:42 (3%)||3:16:25 (6%)||3:22:07 (9%)||3:27:49 (12%)
 
|-
 
| 3:10:00||3:16:01 (3%)||3:22:02 (6%)||3:28:03 (10%)||3:34:04 (13%)
 
|-
 
| 3:15:00||3:21:20 (3%)||3:27:40 (7%)||3:34:01 (10%)||3:40:21 (13%)
 
|-
 
| 3:20:00||3:26:40 (3%)||3:33:20 (7%)||3:40:00 (10%)||3:46:40 (13%)
 
|-
 
| 3:25:00||3:32:00 (3%)||3:39:00 (7%)||3:46:01 (10%)||3:53:01 (14%)
 
|-
 
| 3:30:00||3:37:21 (4%)||3:44:42 (7%)||3:52:03 (11%)||3:59:24 (14%)
 
|-
 
| 3:35:00||3:42:42 (4%)||3:50:25 (7%)||3:58:07 (11%)||4:05:49 (14%)
 
|-
 
| 3:40:00||3:48:04 (4%)||3:56:08 (7%)||4:04:12 (11%)||4:12:16 (15%)
 
|-
 
| 3:45:00||3:53:26 (4%)||4:01:52 (8%)||4:10:19 (11%)||4:18:45 (15%)
 
|-
 
| 3:50:00||3:58:49 (4%)||4:07:38 (8%)||4:16:27 (12%)||4:25:16 (15%)
 
|-
 
| 3:55:00||4:04:12 (4%)||4:13:24 (8%)||4:22:37 (12%)||4:31:49 (16%)
 
|-
 
| 4:00:00||4:09:36 (4%)||4:19:12 (8%)||4:28:48 (12%)||4:38:24 (16%)
 
|-
 
| 4:05:00||4:15:00 (4%)||4:25:00 (8%)||4:35:01 (12%)||4:45:01 (16%)
 
|-
 
| 4:10:00||4:20:25 (4%)||4:30:50 (8%)||4:41:15 (13%)||4:51:40 (17%)
 
|-
 
| 4:15:00||4:25:50 (4%)||4:36:40 (9%)||4:47:31 (13%)||4:58:21 (17%)
 
|-
 
| 4:20:00||4:31:16 (4%)||4:42:32 (9%)||4:53:48 (13%)||5:05:04 (17%)
 
|-
 
| 4:25:00||4:36:42 (4%)||4:48:24 (9%)||5:00:07 (13%)||5:11:49 (18%)
 
|-
 
| 4:30:00||4:42:09 (5%)||4:54:18 (9%)||5:06:27 (14%)||5:18:36 (18%)
 
|-
 
| 4:35:00||4:47:36 (5%)||5:00:12 (9%)||5:12:49 (14%)||5:25:25 (18%)
 
|-
 
| 4:40:00||4:53:04 (5%)||5:06:08 (9%)||5:19:12 (14%)||5:32:16 (19%)
 
|-
 
| 4:45:00||4:58:32 (5%)||5:12:04 (10%)||5:25:37 (14%)||5:39:09 (19%)
 
|-
 
| 4:50:00||5:04:01 (5%)||5:18:02 (10%)||5:32:03 (15%)||5:46:04 (19%)
 
|-
 
| 4:55:00||5:09:30 (5%)||5:24:00 (10%)||5:38:31 (15%)||5:53:01 (20%)
 
|-
 
| 5:00:00||5:15:00 (5%)||5:30:00 (10%)||5:45:00 (15%)||6:00:00 (20%)
 
|-
 
| 5:05:00||5:20:30 (5%)||5:36:00 (10%)||5:51:31 (15%)||6:07:01 (20%)
 
|-
 
| 5:10:00||5:26:01 (5%)||5:42:02 (10%)||5:58:03 (16%)||6:14:04 (21%)
 
|-
 
| 5:15:00||5:31:32 (5%)||5:48:04 (11%)||6:04:37 (16%)||6:21:09 (21%)
 
|-
 
| 5:20:00||5:37:04 (5%)||5:54:08 (11%)||6:11:12 (16%)||6:28:16 (21%)
 
|-
 
| 5:25:00||5:42:36 (5%)||6:00:12 (11%)||6:17:49 (16%)||6:35:25 (22%)
 
|-
 
| 5:30:00||5:48:09 (6%)||6:06:18 (11%)||6:24:27 (17%)||6:42:36 (22%)
 
|-
 
| 5:35:00||5:53:42 (6%)||6:12:24 (11%)||6:31:07 (17%)||6:49:49 (22%)
 
|-
 
| 5:40:00||5:59:16 (6%)||6:18:32 (11%)||6:37:48 (17%)||6:57:04 (23%)
 
|-
 
| 5:45:00||6:04:50 (6%)||6:24:40 (12%)||6:44:31 (17%)||7:04:21 (23%)
 
|-
 
| 5:50:00||6:10:25 (6%)||6:30:50 (12%)||6:51:15 (18%)||7:11:40 (23%)
 
|-
 
| 5:55:00||6:16:00 (6%)||6:37:00 (12%)||6:58:01 (18%)||7:19:01 (24%)
 
|-
 
| 6:00:00||6:21:36 (6%)||6:43:12 (12%)||7:04:48 (18%)||7:26:24 (24%)
 
|}
 
 
 
==110 Offset Table==
 
 
{| {{table}}
 
{| {{table}}
 
| align="center" style="background:#f0f0f0;"|'''40f'''
 
| align="center" style="background:#f0f0f0;"|'''40f'''

Revision as of 16:21, 24 March 2010

It’s harder to run in hotter weather, something that is easy to forget when toeing the line at a marathon. If you’ve set goals based on cool weather training or performance without taking the heat into account, you will go out too fast and your overall time will suffer. But how much does your performance change with temperature? I found a study[1] that gives some answers.

1 The Answer, Part 1

The study showed how finish time changes with temperature, and the results are shown in the following graph.

817939120_uZxSb-O.jpg

This graph works well for runners finishing the race at 3 hour pace or faster, but that excludes the majority of marathon runners. How can we extend this to slower runners?

2 The Answer(s), Part 2

The approach I've taken is to assume the decline is linear from the end points of each curve, as shown by the colored lines. This will give us a projection of performance decline and I’ve created a table at the end of this page based on this assumption.

818611486_iAHcm-O.jpg

3 Your Mileage May Vary

There are a lot of flaws in this approach, which you should be aware of.

  • This study uses only sub-3 hour marathon runners, and their performance characteristics could be quite different to slower runners.
  • There is a wide individual variation in performance decline due to temperature. The level of heat adaptation the runner had, the body size and shape, etc all play into this.
  • The lines in the graphs above are clearly not linear, so making this type of assumption is overly simplistic and probably excessively conservative.
  • Extrapolating information is a flawed approach, as there is no way of knowing what the actual data would look like.

4 Usability of the data

Given all these problems, is this approach usable or useful? I believe there is value as long as you understand the limitations of the approach. If the charts below say you can run a 4:48 marathon based on your cool weather performance and the projected temperature, this does not mean you can run a 4:48 in practice. What it does provide is a sense of just how much you will need to slow down on a hot day, and how important weather is in your results. Hopefully this information will allow you to reset your goals and promote a flexible approach on race day.

5 Marathon Selection

One key note from this analysis is that weather will play a huge role in your results. If you are seeking a race to achieve a specific time goal, such as Boston Qualification, choosing a race that has a high probability of cool or cold weather becomes important. You are probably better off with a hillier course than warm weather for instance.

6 Projected Performance

40f 50f 60f 70f 80f
3:00:00 3:05:24 (3%) 3:10:48 (6%) 3:16:12 (9%) 3:21:36 (12%)
3:05:00 3:10:42 (3%) 3:16:25 (6%) 3:22:07 (10%) 3:27:49 (13%)
3:10:00 3:16:01 (3%) 3:22:02 (7%) 3:28:03 (10%) 3:34:04 (14%)
3:15:00 3:21:20 (4%) 3:27:40 (7%) 3:34:01 (11%) 3:40:21 (15%)
3:20:00 3:26:40 (4%) 3:33:20 (8%) 3:40:00 (12%) 3:46:40 (15%)
3:25:00 3:32:00 (4%) 3:39:00 (8%) 3:46:01 (12%) 3:53:01 (16%)
3:30:00 3:37:21 (4%) 3:44:42 (9%) 3:52:03 (13%) 3:59:24 (17%)
3:35:00 3:42:42 (5%) 3:50:25 (9%) 3:58:07 (14%) 4:05:49 (18%)
3:40:00 3:48:04 (5%) 3:56:08 (9%) 4:04:12 (14%) 4:12:16 (19%)
3:45:00 3:53:26 (5%) 4:01:52 (10%) 4:10:19 (15%) 4:18:45 (20%)
3:50:00 3:58:49 (5%) 4:07:38 (10%) 4:16:27 (15%) 4:25:16 (21%)
3:55:00 4:04:12 (5%) 4:13:24 (11%) 4:22:37 (16%) 4:31:49 (21%)
4:00:00 4:09:36 (6%) 4:19:12 (11%) 4:28:48 (17%) 4:38:24 (22%)
4:05:00 4:15:00 (6%) 4:25:00 (12%) 4:35:01 (17%) 4:45:01 (23%)
4:10:00 4:20:25 (6%) 4:30:50 (12%) 4:41:15 (18%) 4:51:40 (24%)
4:15:00 4:25:50 (6%) 4:36:40 (12%) 4:47:31 (19%) 4:58:21 (25%)
4:20:00 4:31:16 (6%) 4:42:32 (13%) 4:53:48 (19%) 5:05:04 (26%)
4:25:00 4:36:42 (7%) 4:48:24 (13%) 5:00:07 (20%) 5:11:49 (27%)
4:30:00 4:42:09 (7%) 4:54:18 (14%) 5:06:27 (21%) 5:18:36 (27%)
4:35:00 4:47:36 (7%) 5:00:12 (14%) 5:12:49 (21%) 5:25:25 (28%)
4:40:00 4:53:04 (7%) 5:06:08 (15%) 5:19:12 (22%) 5:32:16 (29%)
4:45:00 4:58:32 (8%) 5:12:04 (15%) 5:25:37 (23%) 5:39:09 (30%)
4:50:00 5:04:01 (8%) 5:18:02 (15%) 5:32:03 (23%) 5:46:04 (31%)
4:55:00 5:09:30 (8%) 5:24:00 (16%) 5:38:31 (24%) 5:53:01 (32%)
5:00:00 5:15:00 (8%) 5:30:00 (16%) 5:45:00 (24%) 6:00:00 (33%)
5:05:00 5:20:30 (8%) 5:36:00 (17%) 5:51:31 (25%) 6:07:01 (33%)
5:10:00 5:26:01 (9%) 5:42:02 (17%) 5:58:03 (26%) 6:14:04 (34%)
5:15:00 5:31:32 (9%) 5:48:04 (18%) 6:04:37 (26%) 6:21:09 (35%)
5:20:00 5:37:04 (9%) 5:54:08 (18%) 6:11:12 (27%) 6:28:16 (36%)
5:25:00 5:42:36 (9%) 6:00:12 (18%) 6:17:49 (28%) 6:35:25 (37%)
5:30:00 5:48:09 (9%) 6:06:18 (19%) 6:24:27 (28%) 6:42:36 (38%)
5:35:00 5:53:42 (10%) 6:12:24 (19%) 6:31:07 (29%) 6:49:49 (39%)
5:40:00 5:59:16 (10%) 6:18:32 (20%) 6:37:48 (30%) 6:57:04 (39%)
5:45:00 6:04:50 (10%) 6:24:40 (20%) 6:44:31 (30%) 7:04:21 (40%)
5:50:00 6:10:25 (10%) 6:30:50 (21%) 6:51:15 (31%) 7:11:40 (41%)
5:55:00 6:16:00 (11%) 6:37:00 (21%) 6:58:01 (32%) 7:19:01 (42%)
6:00:00 6:21:36 (11%) 6:43:12 (21%) 7:04:48 (32%) 7:26:24 (43%)

7 The Details

The study looked at 140 marathon results from 6 races (Boston, New York, Twin Cities, Grandma's, Richmond, Hartford, and Vancouver). Only the first 300 finish times were used as races only started recording all finishers in the 1990s. This gives a pool of 42,000 finish times. The races were divided up into four groups The actual temperature measurement used is Wet Bulb Globe Temperature (WBGT) which is similar to ‘heat index’ given on weather forecasts, but WBGT includes the heat of the sun, which ‘heat index’ does not[2].

8 See Also

9 References

  1. http://www.medscape.com/viewarticle/555022 Impact of Weather on Marathon Running Performance (free access, but signup required)
  2. http://en.wikipedia.org/wiki/Wet_Bulb_Globe_Temperature Wet Bulb Globe Temperature