Sunday, December 09, 2012

An interlude - indoor sessions for Team Time Trial training

Despite my intention to finish the incomplete previous post this evening, a day spent largely discussing the specifics of training for a team time trial (TTT) led to me having a moment of clarity during my solitary walk home through the parks regarding training.

Whilst the methods for preparing for a TTT as a unit as simple in principle, and might be covered at a later date, I was initially stumped when I considered how to prepare for the demands of the race. However, eventually I realised that you can reduce the problem down to series of assumptions (that may be race-specific). Furthermore, these assumptions can be turned into parameters with a little more careful thought.

 As such, let's get going. We are going to use the real life example of the BUCS TTT race that is a target of mine for April of next year, and make the following statements:

- The race is 50km long.
- There are three men per team, and all three must finish.
- The difference between being Man 1 compared to Man 2 or 3 is 30% (this assumes that we can ride 6-12 inches from the guy in front, but that just requires practise)

These two statements can be turned into parameters, as will be demonstrated later. To complete the picture, we make three more assumptions:

- The course is flat.
- There is no wind.
- There is little difference between sitting in second or third man in the pace line.

From these six pieces of knowledge, we can construct a representation of the specific demands of the event:

- Each rider takes turns. Using that the race is around 70 minutes long, we suggest that each rotation of riders should be somewhere between 3 and 6 minutes. Any shorter and the energy wasted switching riders will be costly, much more and people won't be able to ride hard enough during their turn on the front.

- In the 'default' case that each rider is of identical strength, turns of one minute each are about right. If fully fit, one should be able to go slightly faster than their 10 mile race pace when on the front (i.e. around 115% FTP) and correspondingly softer when in the pace line (around 80% FTP)

- If one rider is stronger or weaker than the others, it is better to weight the TIME spent on the front rather than the INTENSITY. That is, whatever the relative strengths of your three athletes, your speed should remain constant throughout (given the flat road and no wind). This is the same principle that is applied to an individual time trial, and is not revolutionary.

From this, we can construct session that replicate the event specifically that can be done at home (on the turbo). Due to the nature of them, they are best done with power, as using HR or RPE will be very unreliable when going over threshold intensity. We now use the fact that 'default' sessions to prepare for a race of roughly an hour are 2x20' or 3x15' at race pace, or alternatively 3x10' or 6x6' at slightly above race pace.

The relative strengths of the rider in question compared to his two teammates is a continuum, and we can use 'feel' to map this onto the scale of percentage time spent at the front of the group - the extremes being that the rider is too weak to take any turn at all, or so strong that he must stay on the front for the entire race. Clearly in these situations, there is no need to undertake such training.

In addition, it is a good idea to put upper and lower bounds on the amount of time spent at the front. For example, if 'by numbers' the rider should only spend 5% of the time on the front, that corresponds to only 18 seconds even using the upper bound of 6' rotations. In this case the energy wasted by switching people is not counteracted by the recovery that the stronger two men would get from our 5% chap being on the front.

Somewhat arbitrarily, I will therefore place these upper and lower bounds at 20% and 80% respectively. It therefore remains to work out how much intensity the rider will need in each situation. If on the front for 33% of the time, in the case of equal riders, we have our first session:

- 1' @ 115%, 2' @ 80% 

Repeating this 5 times gives a 15' effort, for instance, but it could easily be done as 2x20' or whatever one feels is appropriate. We now change this session based on the amount of time spent on the front as follows:

- As you spend more time 'on the front', your % FTP will fall (for both 'on' and 'off' sections), so that once you reach 100% of time on the front, you are doing 2x20' (or 3x15') @ 100% FTP (sound familiar?) as and 'on' section.

- Conversely, as you spend less time on the front, your % FTP will fall, again so that once you reach 2x20' @ 100% FTP (but this time it will be 'off')

We now have these three data points, but how do we move between them - is it simply a linear relationship between time on the front and % FTP for 'on' and 'off' sections? If so, we would have the following:

- ON: 33% on -> 115%, 100% on -> 100%. Thus, 'on' FTP = 122.5 - 0.225*time on (%)
- OFF: 33% on -> 80%, 0% on -> 100%. Thus, 'off' FTP = 100 - 0.6*time on (%)

This may sound like a nice, mathematical way to approach the problem, but we are neglecting one fact: we know the difference in power between the 'on' and 'off' sections MUST be 30%, i.e. the savings due to drafting. How do we use this?

Simple: one more assumption! We assume that, whatever percentage of time we spend on the front, the AVERAGE power for the interval is to equal our FTP (i.e. 100%). This is perhaps not as good as assumption as working with Normalised Power (NP), which requires us to: 

1) starting at the 30s mark, calculate a rolling 30 s average.

2) raise all the values obtained in step #1 to the 4th power.

3) take the average of all of the values obtained in step #2.

4) take the 4th root of the value obtained in step #3. 

However, this does not allow for an easily calculation. Given an hour and some MATLAB time I might look into the NP method, but for simplification I will run with average power. Our problem is now reduced:

- We spent a proportion x of total time working 'on' at a power P
- We thus spend a time 1-x as 'off'
- 'off' is done at 70% of the power than 'on' is
- The average power is ones FTP

The formula is thus:

Average power = x * P + (1-x) * P * 0.7 = FTP

Rearranging, we have 

P = FTP/(0.7 + 0.3*x)

or in terms of % FTP and % time spent on the front (X): 

P = 100/(0.7 + 0.003*X)

We now have the ability to check the figures - how does this stack up with what we know? Clearly, if x=0 or x=1 we have the 'correct' answer. If x = 0.33 (as in our initial example), we then calculate than

P(33.3) = 125

This is significantly higher than our initial estimate done by 'feel'. In fact, doing a minute at 125% FTP followed by 2 minutes at 87.5% FTP is probably too tough to be done in training (untapered) for 2x20'. Plugging in more random numbers (20% corresponding to a weak rider and 60% corresponding to a strong one), we have:

P(20) = 131.5 and P(60) = 113.6

Looking at these, I know which I would rather do (let's say 1' on @ 131%, 4' off @ 92% vs 3' on @ 113.6%, 2' off @ 79.5%). In fact, as the intensity of the 'on' section is increasing, so is the punishing nature of the session. This brings us back to the concept of Normalized Power, which takes into account the variability of the effort and 'rewards' a more disparate series of efforts with a higher number.

Clearly, we now have two choices:

- Use NP to calculate our intensities. This will require a small amount of coding, but should be very doable.
- Change our assumption that the whole session must equal a certain % FTP to some other criterion.

In the interests of science, I am going to try out option (1) first, but haven't ruled out that giving spurious results either.

One more thing (before I sign off) that strikes me is that our formula for the session power can be altered to reflect different coefficients of drafting savings. For example, if you have a more upright stance than your teammates, the draft factor may be higher than 0.7, and vice versa. It is also possible to see how the effect of hills might be modelled - the drafting effect will be vastly reduced going uphill, and the overall intensity will be increased and correspondingly decreased for downhill sections.

I fear that this post may also need a 'part two' - I will endeavour to write it while the idea is fresh in my mind!

No comments:

Post a Comment