Wednesday, 4 February 2015

Data, Data everywhere nor any time to Think

(with apologies to Tenysson)

Take a close look at many club cyclists and most triathletes these days and they and their bikes will be kitted out with an array of electronic devices that not long ago would have only been available in well equipped sports laboratories. Heart rate monitors, power meters and the like all communicating with a central unit that can display this information to the cyclist in real time as well as broadcasting it further afield. The data is also recorded so that it can be analysed post ride. Most of the head units are also GPS enabled so that the rider can plot their ride on mapping software.

But other than bragging rights (colloquially known as "willy waving") what's all this data used for? Well for many, not a lot. Online systems such as Strava will use the GPS data to map out your rides and show the times over whatever segments you have passed through potentially handing out "cups" or "King of the Mountains" should you be particularly quick but other than displaying the rest of the data not a lot is done with it. Training applications such as TrainingPeaks or Golden Cheetah will provide basic analysis of the personal data but ignore most of the GPS derived data. There is then the problem (and it really is a problem) of how to do the analysis of the data and how to draw the correct conclusions from it, i.e. convert data to information.

With the generally crap weather that the winter months have brought us recently I've been spending a lot of time on the turbo trainer which while hardly interesting is better than doing nothing or sliding around on icy roads. Simply getting on a turbo and spinning away is deadly boring, after twenty minutes you've had enough. Fortunately there are ways round this with varying degrees of cost depending on what kit you already have. Many GPS units let you set up workouts that let you vary the effort you are doing so you can set them up to make/encourage you to do two minutes' hard effort every ten minutes or whatever takes your fancy. It's a bit basic though. There are also video services such as Sufferfest - you set up a computer in front of the turbo and ride along to race footage along with exhortations to go for it.

I settled on Trainerroad (TR) which is sort of halfway between the two: it consists of predefined workouts but has on-screen instructions and tips to help you along. Another advantage was that it is power based rather than heart rate based so no dealing with the lag between changes in effort and your HR going up or down. Although I don't have a power meter TR has power profiles for lots of turbo trainers in their virtual power section - not 100% accurate but so long as the values are consistent then you can work against them and it's a bit cheaper than £600+ for a power meter! The TR training plans also include regular fitness tests as each workout is adjusted to your current fitness level - an estimate of how much power you can sustain for an hour known as Functional Threshold Power (FTP).

After a couple of weeks following one of the training plans based on my estimated FTP all was going well. The workouts were hard but then that's the point but I was completing them OK and either matching expected power or exceeding it. Some of the workouts in the plan are 90mins long - there's no way I could sit on a turbo for that length of time without some stimulus but with things to do and targets to meet . Then on Monday things went a bit wrong and I had to finish the workout without completing the last effort. Although slightly different in number and length of efforts, the workout was no harder than one the previous week but I felt as if I'd been trying to sprint through the efforts rather than a measured output. Time to look at the data.

A quick look at the data in Strava (it shows HR, power, cadence, etc but doesn't do any real analysis) showed my heart rate was constantly climbing from effort to effort. Cardiac drift isn't unusual in exercise but rather than a sawtooth profile where each effort is similar to the previous one, each effort carried on from where the previous one had left off. By the third effort my heart rate was at 95% of my theoretical maximum for nearly the entire ten minute effort! Although the rest intervals had my HR dropping back to reasonable levels it was obvious that I wasn't really recovering.

Moving over to Golden Cheetah and comparing the distribution of my HR with a previous workout showed the problem. Whereas the previous week I'd spent less than 30% of the workout in upper zone 3 or higher, this workout it was over 60% and that while producing less power. My power output was even more telling: at the end of the third effort my power dropped by over 30% with no corresponding drop in heart rate. Plotting heart rate vs power was even more telling. I was starting with a cold or some form of infection.

Rather a lot of time spent in Zone 4

This is from a decent session, virtually no time in zone 4.

Sure enough by Wednesday morning I'd got the tell-tale tickle at the back of my nose/throat. By Wednesday evening it hadn't really developed so I chanced doing the planned workout which I completed OK though my power tailed off for the last effort. By Friday it was back and even short efforts on a ride over to Wycoller on Saturday had my HR soaring so it might be best to lay off the training for a while.

Actually the efforts on Saturday showed that the limited turbo training I've done so far has already made improvements: there was very little muscular stress and the limiting factors were my heart and lungs going in to overdrive! This is somewhat unusual for me: my muscles are normally screaming while heart and lungs are ticking along.

Went out for a quick ride this evening it was a bit icy so times weren't brilliant, especially on the downhills where you had to be careful of suddenly coming across sheet ice. The climbs though felt good, not as fast as my fastest times but those were done in daylight and on dry ground rather than at night with ice and snow.

I wasn't expecting anything outstanding but was pleased with how things went especially with regards muscle efficiency and lack of tiredness in my legs. Will be nice to get out of the other side of this bug.


  1. Interesting blog as usual Bob, good use of data to make a prediction! IIRC the England Rugby Sevens team use GPS and heartrate monitoring to tell their players when they are coming down with a cold and tailor their workouts accordingly. Best wishes!

  2. Back in the day I used to measure my HR every morning. If it was 10bpm or more above normal then I knew I'd either been overtraining or was starting with some illness. I remember seeing the piece about the England Rugby team.