Heat and speed – a statistical look at the Virgin Money London Marathon

VM London Marathon 2018

Now that more of the London Marathon stats have had time to settle, I have had a chance to look at the stats from the race and see how the heat may or may not have affected the performances of the 40,273 finishers.

If you look at the stats below you will see the stats for the four hottest years on record for the race , 1996, 2007, 2011 and 2018 (final temp is yet TBC but is showing as 24.1). As I mentioned in my pre race article, it seems that in the past a warm race doesn’t massively affect the percentage of finishers.

London Marathon Temperatures 1981 to 2018

YearTemp degC 0900Temp degC 1000Temp degC 1100Temp degC 1200
22 April 2018TBCTBCTBC24.1
22 April 200716.318.620.521.7
21 April 199617.620.320.121
17 April 201114.115.918.319.9
20 April 19868.110.215.617
13 April 20039.511.51516.4
17 May 198810.613.214.616.4
26 April 200914.114.916.116.2
25 April 201011.813.814.415.6
2 April 19958.39.61214.1
9 May 198210.511.113.414
23 April 201711.712.613.613.9
18 April 199311.812.213.113.8
12 April 199211.312.113.213.8
13 April 201410.511.312.713.7
22 April 20128.810.412.513.4
13 April 19979.911.21213.3
13 May 19848.610.412.213
21 April 201378.910.412.8
23 April 200611.111.312.112.7
26 April 199811.111.21312.7
10 May 19871010.61212.4
17 April 20057.79.410.912.3
16 April 20006.68.79.911.3
18 April 20049.610.210.311.2
14 April 20027.69.210.611.2
29 March 198110.110.210.210.9
22 April 19909.41010.210.9
22 April 20017.68.48.610.4
23 April 19898.48.58.710.1
17 April 198310.11010.410
13 April 200810.412.18.99.5
21 April 19915.87.19.49.1
26 April 20158889
24 April 20166.87.489
21 April 198566.988.6
18 April 19995.96.87.68.4
17 April 19945.96.67.57.6

As Hugh Brasher, Event Director, stated in the press release of 23rd April not only did the event see its highest number of finishers but the number of DNFs was only 169 higher than in 2017.

The 2018 race percentage finisher rate was slightly down on previous years but again, if you look at that as an overall percentage this doesn’t to me seem like a massive difference.

However take a look at the finishing times. Now I’m not going to tell you that this is high level statistics and there may well be people out there much better qualified than I to analyse the results but just as a simple approach I looked at ten years of results (all that are still available) and took the 10th, 25th, 50th, 75th and 90th percentile finisher times. So for example with 40,273 I looked at the 4,027th finisher to give the 10th percentile place and so on. I was hoping to see whether there is any sort of link between the temperatures for the race and the performances of these percentile points or not. I was also interested to see whether there was any link between the general trend for one percentile point and others in the same year.

London Marathon Finishers Percentile 2009 to 2018

 10th percentile25th percentile50th percentile75th percentile90th percentile
26 April 200903:26:0903:54:2004:28:2805:04:5305:43:45
25 April 201003:25:4303:53:5404:27:5405:05:4805:46:21
17 April 201103:26:4203:57:2404:34:0205:14:0005:54:31
22 April 201203:24:2003:53:3904:28:5305:09:0005:49:56
21 April 201303:20:5303:50:3804:27:2905:07:2105:49:54
13 April 201403:20:3403:49:5004:25:4405:06:5805:48:42
26 April 201503:15:3003:44:4804:19:0704:57:3305:39:31
24 April 201603:14:4203:45:0704:20:5505:01:4105:46:06
23 April 201703:15:4703:47:2104:26:5705:09:2505:54:16
22 April 201803:28:0504:02:0604:47:5605:34:2006:17:03

And my broad analysis is…….. Yes the hot race in 2018 showed much slower times for all data points than previous years. The 10th and 25th percentile mark had been on an improving trend until this year; whereas the 50th, 75th and 90th percentile points seems to have varied quite a bit with no significant data trend.

One thing we do know is that the number of ‘fast finishers’ ie subs 2.20 / sub 2.30 was much higher at the race in the 1980s than now. An Athletics Weekly article from 2016 demonstrates a number of stats on these benchmarks for British runners but looking at the times for the 10th and 25th percentile finisher times I wonder where general trend is for improvement is coming from?

My only suggestions are two fold. The first is that we seem to be having another running boom over the past few years and that is bringing more people into the sport. The second could be the improvement in the standards of other big city marathons in the UK. I have been a runner since the early 1980s and there were no shortage of big city marathons at that time but when you look at the professionalism of the big races now being staged (Brighton, Manchester, Bournemouth etc) it is allowing more runners to race to their potential and thus create a wider base of faster running talent in the UK.

VM London Marathon 2018

When I look back over what I have written above I’m not sure that I’ve come up with any definitive answers nor solved the ‘big’ questions about the state of distance running. I do however hope that like many runners, a bunch of stats may be of interest to you! One thing is for sure – the heat of 2018 has not deterred a record 414,168 people from applying for the ballot for 2019. Maybe just getting into the race is the hardest part……

Enjoy your running – whatever the pace

Mike

London Marathon Applicants, Successful applicants, Starters and Finishers 1981 to 2018

YearTotal applicantsSuccessful applicantsStartersFinishersWheelchair finishersIPC World Cup finishersTotal finishers incl WC & IPC% finishers
29 March 198120,0007,7477,0556,255006,25588.66%
9 May 198290,00018,05916,35015,1160015,11692.45%
17 April 198360,00019,73516,50015,77617015,79395.72%
13 May 198470,00021,14216,99215,64926015,67592.25%
21 April 198583,00022,27417,50015,84132015,87390.70%
20 April 198680,00025,56619,26118,03037018,06793.80%
10 May 198780,00028,36421,48519,54541019,58691.16%
17 May 198873,00029,97922,46920,88943020,93293.16%
23 April 198972,00031,77224,45222,65249022,70192.84%
22 April 199073,00034,88226,50024,95360025,01394.39%
21 April 199179,00033,48524,50023,39342023,43595.65%
12 April 199283,00034,25024,50023,78350023,83397.28%
18 April 199368,00035,82025,00024,44847024,49597.98%
17 April 199472,00037,37926,00025,19448025,24297.08%
2 April 199579,00039,09727,00025,32651025,37793.99%
21 April 199668,00039,17327,13426,76145026,80698.79%
13 April 199778,00039,81329,50029,13554029,18998.95%
26 April 199896,00042,22830,66329,92448029,97297.75%
18 April 199987,00043,77431,58230,80940030,84997.68%
16 April 200093,00042,59632,62031,65840031,69897.17%
22 April 200192,00043,51731,15630,28632030,31897.31%
14 April 200299,00046,08333,29732,90644032,95098.96%
13 April 2003111,00045,62932,74632,28143032,32498.71%
18 April 2004108,00045,21932,74631,98329032,01297.76%
17 April 2005132,00047,96935,55735,26040035,30099.28%
23 April 2006119,00047,02033,57833,22426033,25099.02%
22 April 2007128,00050,03936,39635,69831035,72998.17%
13 April 2008120,00048,63035,03734,60235034,63798.86%
26 April 2009155,00049,99535,88435,36638035,40498.66%
25 April 2010163,00051,37836,95636,63234036,66699.22%
17 April 2011163,92650,53235,30334,83636034,87298.78%
22 April 2012170,15050,20037,22736,77438036,81298.89%
21 April 2013167,44948,32334,63134,311462434,38199.28%
13 April 2014169,68249,87236,33735,911412535,97799.01%
26 April 2015172,88851,69638,02037,641539937,79399.40%
24 April 2016247,06953,15239,52339,091613239,14099.03%
23 April 2017253,93053,22940,04839,487633239,48798.60%
22 April 2018386,05054,68541,00340,273TbcTbc40,27398.22%