London Marathon 2020 Prediction
2 Men Under 2 Hours? A Scientific Analysis.
Not only will it be a historic day in London on Saturday with the 40th version of the marathon in the UK’s capital city, but we could also see something very special indeed in the men’s and women’s races.
With both the men’s and women’s world record holders toeing the start line on Saturday, it is feasible that both records could fall. In addition, there is a reasonable chance that one or even two men could go under the magical 2 hour mark as they chase the win.
On the men’s start line is the current world record holder and the first man to go under 2 hours, albeit in a race that did not count for the record books because of the way it was set up, but it did perhaps give an insight to what we might see this weekend.
The Kenyan Eluid Kipchoge’s run of 2:01:39 in the 2018 edition of the Berlin Marathon was truly impressive and is still the current world record. Although, Ethiopian Kenenisa Bekele’s run in the 2019 Berlin Marathon of 2:01:41 gave Eluid a run for his money and underlined his undoubted running pedigree. Saturday’s race brings not only two of the fastest ever, but they are tantalisingly equal in terms of ability.
In my nearly two decades of working in high performance sport, I have had the honour of working with both of these athletes in one form or another. Eluid in 2004–2006 when I spent multiple weeks at his training camp in Kaptagat in the west of Kenya. And Keninisa in 2007 as I helped him prepare for the upcoming IAAF World Championships in Osaka (where he went on to win the 10,000m). Both athletes, were absolutely supreme human specimens and I learnt a lot from being in their environments. Knowledge and experiences which in many ways helped me to support Sir Mo Farah to his 4 Olympic titles and 6 World titles between 2010–2017 when I worked at British Athletics.
Whilst it is pretty easy for anyone to simply state that both these men could go under 2 hours at the weekend, I want to make a case for it using some science. As a physiologist by trade, my mission has always been to try and understand performance from a first principles persepctive.
The physiology of running in many ways is quite simple. As you go faster, your heart rate and breathing rate increase which reflects your muscles requiring more energy to stoke the engine so to speak and at some point it cannot go any faster or keep going any longer which is your capacity. But in many ways it also complex. In fact the human body is an aboslute marvel. And when it exercises it becomes one of the best designed machines ever created. Professor’s spend their entire careers working out how the body works during exercise. One of those is Andrew Jones at Exeter University. Andy was one of the physiologists at British Athletics when I was there in that role 2007–2013 and he is a world expert on endurance physiology. He has written a lot (and I mean a lot) of papers on the subject. One of which is on the use of something called critical speed in running (you can see the detail here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5371646/).
The main points for the purposes of this article is that critical speed describes the speed at which a runner can run for an about an hour or more. It represents the shift from steady state to non-steady state exercise (this is when you go from being quite comfortable to starting to use up your finite energy stores…or to put another way, start breathing out your backside). Jones et al (2017) suggest an elite marathon runner can hold their critical speed for about 96% with less experienced runners probably nearer 90%.
Whilst it can be an academic exploration in terms of how you measure it (normaly requiring multiple visits to a laboratory), you can in the real world estimate critical speed by using an athletes personal best times over a range of distances (at least three) and plotting them on a graph. The slope of the line is the critical speed.
Eluid’s critical speed from his best times over 5000m (12:46.53), 10,000m (26:49.02) and half marathon (59:25) results in a critical speed of 5.74 m/s. Assuming Eluid can run at 96% this predicts a marathon time of 2:07:40. Similary, Kenenisa’S times over the same distances (12:37.35, 26:17.53, and 1:00:09) results in a critical speed of 5.61 m/s which converts to a marathon time of 2:10:35. Obviously both athletes have run quicker than their predicted times and therein lies one of the challenges of critical speed for prediction. It assumes that the times the athletes have run are their best efforts. So while Kenenisa’s 5000m and 10,000m times are unquestionable (as they were until recently world records), his half is probably not a good reflection of what he can do (I estimate he should be doing mid 58 minutes). Similarly, one suspects Eluid could run quicker in the 5,000m, 10,000m and half marathon than he has (something like low 12:40’s for 5,000m, low 26:20’s for 10,000m and something similar to Kenenisa in the half, mid 58 minutes). Now whilst that is a big factor, a further factor in the “error” are the distances chosen to calculate critical speed. If I were to chose shorter distances it will shift the slope of the curve and signifcanlty increase the critical speed. Indeed, using their shorter races seems to get them closer to the 96% marathon prediction. So why don’t I use them in my prediction? Well, when comparing to the normal runner, it is unsual that they will have true 800m and 1500m personal best times if they are a recreational marathon runner. So that’s why I went about looking for a new model.
I sampled 108,000 runners who had a 5000m, 10,000m, half marathon and a marathon time. I calculated critical speed from the slope of their 5000m, 10,000m and half marathon times and then looked at what percentage of critical speed each runner ran those respective races at. There were obvious differences between groups of runners (i.e. beginner, intermediate, sub-elite, and elite) and between sexes in what percentages races are run at but for elite men it looks like they tend to run marathons at closer to 97.3% of their critical speed.
As this model I have created covers a range of abilities, I can use the results of this data exploration to build training programs for recreational runners through to elites. So for example, if you are a beginner and are female, I know with reasonably good accuracy what percentage of critical speed you will run at and therefore I can also calculate training zones. For example, I can tell you if you want to run a 2 hour half marathon what each of your training runs should be paced at. Together with some Olympic level coaches, we have some programs that fit the model. If you are interested in getting your hands on this novel type of training program, we are launching an app in the coming months. You can register interest with the link at the bottom of this page.
So 97.3% is better but it is still way out for what these guys can actually do. Probably a result of lower than expected personal bests over the distances and possibly their ability to hold a higher percentage of their critical speed than the average elite runner. So I started to think about this another way. If we look at the new 5000m world record (just set by the Ugandan Joshua Cheptegei last month — 12:35.36) this may give a good indication of where the record may go. I say that as I believe that the Nike shoes that they are all wearing now undoubtedly have moved the sport on. A back of the envelope calculation suggest almost 3.5 minutes improvement for the marathon! That debate on whether this is helpful for the sport is not for here, but they are definitely having an effect. So taking Joshua’s new record and applying my new model this suggests that a 2:00:34 is possible on Saturday. For what it is worth, the model also suggest Joshua could break the 10,000m record when he runs it the week after next in Valencia with a 26:09 possible and the half marathon could go to 57:57 when it is next run in perfect conditions. But sticking to the marathon for now, 2:00:34 is tantalisingly close to the 2 hour mark and with some factors going their way it could just happen.
The course for any marathon is massively important, too many hills, twists and turns and it can take time away from even the best in the world. But Saturday’s race is on a looped course in St Jame’s Park and finishes on the iconic Mall which arguably is as good as it gets, especially in London. The weather is also a factor, too hot and windy and it too will take time away. However as it is October in England, it shouldn’t be too hot and hopefully with some good pacers the effect of any wind can be reduced. Sticking to pacing, it also has to be paced well and have a group of pacers that can go as far as possible. Word on the street is that the pacing should be spot on. And lastly, the athletes have to be in peak shape on the day. Given that both are very experienced marathon runners, assuming no injuries in the build up, will be ready to go on race day. There are of course lots of other factors such as nutrition and fluid but again these guys are pros and will know exactly what they are doing. So all going well there is a good chance we will see the best of these guys this weekend and if they both start racing towards the end of the course (and on pace) we really could see something very special as they are spurred on to break that 2 hour mark.
My prediction is that both men go under 2 hours this weekend. Boom there you go.
If you are intersted in learning more about our new training app that is being launched in the next couple of months you can click this link:
