Big Data states the obvious for run-training insights. But…

Big Data gives marathoners new insights

A couple of days ago I looked at a training app from AI Endurance which closely looks at your individual performances and works out exactly what caused the good and bad adaptations in the past. The theory is that, in the future, you can improve your training by tailoring what you do more around workouts that have historically proved to be effective for you.

Another way of discovering what is effective is to look at much larger, population-wide data sets. The theory here is that you find a large group of similar, mid-pack runners (or whichever pack you are in) and then mimic aspects of their training or the aspects of training from similar athletes to you but who beat you 😉 . This is the BIG DATA approach.

Here’s a full article about this at which I found ‘obvious’ and ‘interesting’, both at the same time. Obvious that it tells us many things we already knew but interesting that it quantifies some of our beliefs with a dose of pesky science and offers up an odd mini-insight or two.

These were the key take-outs for MARATHON training which you could assume apply in principle to long-distance triathlons too:

Thorsten Emig looked at data from lots of athletes who have logged their data on Polar Flow. He found 4 things. I’ll mention these 3.

  1. Run More – and you can only run significantly more if you run slowly.
  2. Run faster for short distances and mitigate injury risks
  3. There is a point above which more training can be detrimental. This is 27,000 TRIMPs for the 6-months-to-date. (Google: TRIMP, even Strava’s relative effort uses it)

Then Barry Smyth (University of Dublin) looked at some big data from Strava. He found that a 4-week cycle seemed to be most effective. The themes of each week followed the Hard, Hard, Easy, Moderate pattern.

Easy means 2/3rd or even half the volume of other weeks and Hard means that 5k pace or Tempo pace work is included at least once. And remember this is for MARATHON training. So this broad approach seemed better than progressively and slavishly aiming to raise your weekly mileage and intensity.




Original Article:


Reader-Powered Content

This content is not sponsored. It’s mostly me behind the labour of love which is this site and I appreciate everyone who follows, subscribes or Buys Me A Coffee ❤️ Alternatively please buy the reviewed product from my partners. Thank you! FTC: Affiliate Disclosure: Links pay commission. As an Amazon Associate, I earn from qualifying purchases.

3 thoughts on “Big Data states the obvious for run-training insights. But…

  1. How on earth do you look at your cumulative TRIMP? Flow can display Cardio load build up for the past 6months but it only lists TRIMP on individual activities.

    1. In the training report bar graph you can see TRIMP by month in Polar Flow. I’ve already hit 28,000 for the year so I can now stop training.

Comments are closed.