Some time ago I came across a really fascinating article (https://medium.com/running-with-data/perfect-pacing-at-the-london-marathon-95f09db2ac5c) written by Barry Smyth, a data scientist from Ireland. At the time the article wasn’t stuck behind a paywall, so to summarise what Barry did as the introduction outlines was:
- An analysis of more than 215,000 London Marathon runners during the period 2011–2016;
- How do runners pace the London Marathon? Where are the fastest and slowest sections of the race?
- How does the pacing of elites compare to recreational runners and those running a PB?
- Putting what we learn into practice we create an optimal pacing chart for recreational runners that is tailored for London.
The article itself was a great read, Barry took all of the above data, plug it into a spreadsheet to model of how both elite and non elite’s paced a marathon by 5k splits. What he then produced was the % spread from beginning of the marathon to the end. For the non elites he found that people started 4% faster than their average pace and slowed over the course of the run to finish 4.3% slower (an 8.3% swing) – so a 3hour marathon runner would start around 4h 6m / km and finish at 4m 27s / km. He found that the elites were far better at staying close to goal average pace / deviating by no more than 2% faster at the start and 2% slower at the end (a 4% swing) which he was less surprised at, due to the nature of better training / internalised pacing / experience. As I read this, it stuck in my head and certainly rang true. Looking at the vast majority of runners marathon splits, post race, the vast majority go out faster than they finish, with a steady slow down – with what looks like a more considerable drop off over the last few KM.
With this in mind I wondered about reverse engineering his research – my thought process being that if this is how people are running marathons – and most likely against their initial goals at the beginning and succumbing to a gradual slowdown over the course of the marathon. It seemed like a good starting point to think about how to pace a marathon. With this in mind I took his numbers and flipped them, so used the slowest pace (4.3% slower than average) as a starting point for the first 5k and looked at how paces would progress over the course of the marathon.
The result was great, as it gave a far clearer guide to how a marathon could look like if paced more conservatively from the start. My initial thoughts were – it looks too slow at the start and too fast at the end, which I then observed was curious as this is exactly what is currently happening in reverse. It also led me to realise that training for the marathon had to have far more of this cutdown nature in longer marathon based sessions, in the 6-8 weeks leading up to a marathon. Finishing on sections at 4% faster than your goal pace key to ensuring you could achieve your goal pace.
I then put this properly into practice over my own last marathon block – and also for 2 of the guys I coach. Success for all 3 of us, far better pacing, negative splits for 2 of us and most importantly all 3 of us avoided the large drop off in pace that can happen at the end of a marathon. So the last 10k becomes one were you are are running past people, feeling confident you can hold your pace and the mental boost this gives is wonderful.
The calculator is here:
So to use above, you enter the goal time you want to finish and it spits out the 5k splits in time & average pace. I’ve added in 10k splits & half splits.
The way I use it is as a starting point, a guideline, not something to be stuck to religiously. It’s a good place to begin a conversation with a non elite athlete as to what they want to achieve and what they need to be able to run, based on a huge wealth of research of previous marathon runners.
