Challenge Roth to Ironman UK – Recovery, Tapering and Over-Analysis

Less than two weeks to Ironman UK and the limited amount of training I’ve done so far has clearly established one thing. There’s not going to be much quality training the next week or two! I had some illusions that I might throw in some decent sessions, but so far no luck. My legs are coming round, but as my short, easy run last night showed they have a way to go. If I feel the same next Friday I’ll start worrying.

All this extra time allows me to catch up on reading, patch the damage to my wet-suit and spend hours analysing data and speculating for Ironman UK. My speculation can be divided into two categories, firstly equipment choice and secondly potential performance. What’s about to follow is a series of charts, graphs and lightweight analysis from Challenge Roth. From this data I’ve then foolishly attempted to extrapolate something about the Ironman UK course.

Racing by Heart Rate and Power

Averaging a heart rate of only 132 BPM on the bike course at Challenge Roth had surprised me. Previous Ironman races had seen averages in the region of 142 BPM. Ironman Lanzarote was the only instance with such a low heart rate and I was injured there. My bike time was perfectly acceptable given the set-up (heavy box rim training wheels) so I didn’t feel that I needed to be too concerned about fatigue as a factor.

Challenge Roth - Power and Heart Rate

I decided to compare the relationship of power to heart rate for the race. A bit of analysis in excel produced the graph above. What it shows is a consistent relationship between the two. As heart rate increases so does power. Take a line around 132 BPM and sure enough it produces roughly my average power of 210 watts.

When planning a race I usually aim to hold my heart rate in the 140-145 BPM region. Let’s be a little more cautious and expand that range to 135-145 BPM. Allowing a bit of recovery too! From my Challenge Roth graph I could extrapolate a range of power from 220 to 260 watts. Playing on the safe side again it would suggest racing above 220 watt and below 250 watt for the most part. The data suggests I could maintain a comfortably aerobic heart rate in this power range.

Training - Power and Heart Rate Relationship

Wanting to back this theory up a little I decided to apply the analysis to some of my training data. Taking a range of data from long rides over the past few months and again plotting the average power for heart rate. The result was the graph above, based off around 50,000 data points from training. I’ve put a quick linear trend line in place to give a rough idea to values. The graph seems to back up the data from Roth, again suggesting for a heart rate of 135 to 145 BPM a power range of 220 to 250 watts would be expected.

A simple conclusion from this is that when racing Ironman it wouldn’t be unreasonable for me to expect to average 240 watts. That’s a good 30 watts higher than I managed in Roth. However my normalised power was a much closer 230 watts. This suggests that there was enough work above the average to produce a higher normalised figure. Likely caused by the climbing and wind. I have data for the alititude, but nothing for the winds, so the next step was to look at the relationship between heart rate, power and gradient.

Challenge Roth – The Bike Course Profile

Challenge Roth - Course Profile, Heart Rate and Power

I’ve updated my graph of Challenge Roth power and heart rate to include altitude as well. It’s here more for general interest as over the time-scale of five hours it’s hard to see much more than a very general pattern.

Total altitude change for the race is an ascent of 1210m with a descent of 1240m (T2 is separate from T1) according to my Garmin (my GPX file available on With data including elevation recorded every second by the Garmin I had a detailed profile of the course to analyse. Taking the change in altitude over time I quickly established some general figures. Identifying that 44% of the course is descending, a mere 1% flat and 55% is ascending! I was a little surprised by this as Challenge Roth is known for its fast course. As mentioned I can’t factor in wind speed or direction to this analysis of course.

Challenge Roth - Course gradient distribution

Whilst overall the course is predominantly climbing plotting the distribution in change of altitude between data points shows that the majority of that climbing is very subtle. To the point that with bigger buckets you would consider most of it flat. As the graph shows whilst there is a wide range of negative gradients the inclines are mostly very small. Factor in tailwinds and the like and it’s easy to see why the course is fast. A section of flat or tiny incline following a descent will be very fast, especially with the wind behind you.

My conclusions from this? Well primarily it starts to back up the view that Roth is fast due to rapid descents and a lot of easy terrain. Throw in useful tailwinds on those smooth roads and it’s not too hard to see where the speed comes from. Knowing I prefer a relatively low cadence to most ad that on the day my 11 cog was skipping forcing me to avoid it I suspect that this would result in lower than expected heart rate and power averages. I simply couldn’t turn-over the cranks fast enough to generate the heart rate or power my training might lead me to expect.

Relating gradient to Heart Rate and Power

Challenge Roth - Power and Heart Rate by Gradient

The final step of analysis I’ve performed is to look at the relationship between the gradient and my heart rate and power at the time. Essentially taking the range of gradients encountered on the course and then averaging my heart rate and my power for each of them. This should then let me see how the terrain was affecting my ability to raise my heart rate. From this I get the graph above.

It’s not surprising that both power and heart rate rise as the road goes up. Whilst gearing choice will affect how much I think most of us work much harder against an incline than on a descent. I’m a great believer in trying to maintain consistent effort throughout a race, but I’m far from a perfect practitioner!

The trend isn’t surprising then, but what I was more interested in was how this related to the 132 BPM figure and the 210 watt averages. Looking at the lines it does seem that around the zero mark is where 132 BPM and 210 watts occur. Once the gradient is below zero and I’m descending I get the drop off in heart rate and power. Above zero and I’m much closer to that 220 – 250 watt region and the 135 – 145 heart rate.

I’ll openly admit there’s a lot of limitations to this analysis. I don’t think gradient alone explains everything. To my mind there’s a relationship at least at Roth. Perhaps I’ll try and analyse some more training data in the future to see how it holds up. The downside with analysing training data is there’s a greater percentage of freewheeling on descents so this may skew the results. Not being able to factor in wind or fatigue in any clear fashion into this analysis clouds the results further.

Implications for Ironman UK

The main message I’m taking from the analysis is that if I freshen up for the big day I should expect to hold a power of 220 – 250 watts without problem. I won’t be racing with my power meter, opting for faster race wheels this time so I will use heart rate instead. Training and racing data suggest I should aim a a heart rate of 135 – 145 BPM to achieve that power range.

Ironman UK - Bike Course Profile

The other message is that the nature of the course may affect my ability to maintain these values. Taking the profile of a single lap of the UK course above it seems downhill might be quite common. To check this out I used a GPS file (again from to apply similar analysis to the course. It appears that total climbing for the course with be around 1600m and with T1 and T2 being the same descent will match. 31% of the course is descending, 40% flat and 29% climbing. That’s not the impression given n the profile above, I suspect the relatively low number of data points in the GPS file used contributes here. It’s certainly no where near as accurate as my Roth course.

Based purely on the profile my impression is of a relatively fast course. A couple of steeper sections of climbing, but followed by long periods of descending over rollers. I can’t comment on how technical the course is or the influence of that on speed. Wind and road surface will obviously be major contributing factors too. A quick graphs looking at the distribution of gradients doesn’t clarify things too much.

Ironman UK - Bike Course distribution of gradients

A lot in the flat region with a wide distribution of gradients from that. The smaller spread of data points and subsequent greater change in altitude makes it harder to compare directly to the Roth graph. My impression is of a much more rolling terrain with some notable climbs and descents mixed in.

What’s clear is until I see the course on the Thursday before the race there’s only so much I can draw from this data. I suspect it will suit me a little more than Roth with less opportunity to run out of gears.

Conclusions and Limitations

You can tell I’m in recovery/taper at the moment. You have to have a lot of time on your hands to do something like this! How much it really tells me for my future races I’m not sure. It was interesting to at least attempt to backup some of my suspicions. Ironman UK is almost a test bed for what I’ve seen in this data!

Primarily what I’m taking away from this is that my 135 – 145 BPM target for heart rate is pretty reasonable. The power range I get for this is roughly 220 to 260 watts and that seems to hold true for training too. That’s a fairly pleasing 3 – 3.5 watts/kg for an entire Ironman bike. I think the top end of that power range may be pushing it at the moment, but it’s hard to be sure.

I’m more confident that terrain and conditions that aid cycling. So descents and tailwinds to an extent work against my ability to hold my target heart rate. Certainly I need to ensure my highest gear is working perfectly on race day in the future! My natural cadence works against spinning out on descents, I tend to hit my limit quite quickly. Perhaps a bigger chain-ring in the future!?

For Ironman UK I will continue with the heart rate strategy I’ve used in previous races. Hopefully the nature of the terrain is such that this will be easier than in Roth. My view is that whilst having to work harder and maintain a higher average wattage likely means a slower bike time it should mean a higher bike position. Not being able to reach your target zones will certainly have reduced my relative position in the field.

Onto the limitations. I think they’ve mostly been mentioned already, but are worth emphasising again. This isn’t the most scientific of analysis, but rather something I’ve done to at least explore how I’ve raced on the bike.

Wind is a significant factor in bike riding. Anyone who’s trained in Lanzarote can confirm that one! At no point is it factored into these calculations as I just don’t have the data. Similarly I can’t factor in variables like road surface, rolling resistance etc. There are plenty of other factors that might affect the ability to work at a given heart rate or power on race day not least injury or fatigue.

Aside from the missing external factors this analysis in itself is very simplistic and almost certainly flawed. Whilst I studied statistic as part of my Biology degree it’s fair to say that was a long time ago! At best here I am very generally looking at data and drawing conclusions.

I should also mention that whilst the 135-145 BPM figures chosen seem sound it’s relatively arbitrary to choose them. They are based off aerobic values for my training, but at the same time there are other factors to consider in choosing the range. You could look at the 132 BPM in Roth and argue I had a much better run off this time compared to the 142 BPM in Ironman Australia or Western Australia. Also it’s worth considering if my ability to absorb my nutrition was significantly improved by the lower average intensity.

I don’t think I’m alone in spending time attempting to analyse past racing and training data in the run up to a big race. You hope you’ll find something that will confirm you’ll have a fantastic day and a new PB. The reality is you never really do and sometimes come away with more questions than you started. At least in this case I think I’ve essentially ended up with a plan to continue as normal.

Ironman Training Library

From nutrition to pacing - a collection of CoachCox blog posts focused specifically on Ironman training and racing.