Wednesday, April 25, 2012

Empirical Fitness Potential Equation

One of my recent triathlon data analysis projects has been trying to determine, analytically, how far away I am from my fitness potential.  In other words how much faster could I have raced if I had been truly "race fit".  This is off-course, especially from a theoretical point of view, a fairly subjective concept.  However, I thought by examing my actual historical race data, I could get a real world, empirically derived, standard of how well I do when I do my best.  so, I came up with the idea of an "Empirical Fitness Potential Equation".  Here is how it works:

I measure my %-tile finish against the Overall pool of racers in each triathlon.  I do this for the race as a whole and for each of the SBR legs.  I then take the top 5 (roughly my top third each year) %-tiles in each category and average them and set that average as a measure of my best potential performance for the year.  Once I establish this empirically derived %-tile potential, I can compare any race to it and determine how far away I was from racing to my potential.

To illustrate, let me show you the calculation for my 2011 swim legs.  My top 5 %-tile (Overall not Age Group) finishes in 2011 were:

1. Walker Minnesota Sprint: 98.6 %-tile
2. Tuckahoe Sprint: 96.4
3. Bassman Sprint: 95.3
4. Philly Sprint: 93.6
5. Lake Lenape Sprint: 92.8

The average for these five races is 95.3 %-tile.  Here is the same calculation for the swim in my ten full year seasons in triathlon:

2002: 65.4
2003: 63.2
2004: 80.9
2005: 81.2
2006: 84.3
2007: 80.5
2008: 83.3
2009: 85.4
2010: 96.4
2011: 95.3

There three distinct periods, as far as my swim potential goes.  2002/2003 when I was just learning the ropes.  2004-2009 where I became I pretty solid swimmer.  In then, after my off-season commitment to fundamentally overhaul my stroke my developing the potential to be a front of the fact swimmer in 2010 and 2011.

As I think about the 2012 season, it's reasonable for me to assume that I'm racing close to my full fitness potential when I finish at the 95 %-tile in the swim leg.


A similar analysis for the bike leg yields:

2002: 94.8
2003: 98.0
2004: 99.4
2005: 98.9
2006: 98.3
2007: 98.3
2008: 97.0
2009: 97.4
2010: 98.1
2011: 98.0

The bike data shows how naturally the bike comes to me.  I started at a high level and very quickly reached the point where I could deliver the fastest bike leg OA.  By this data, I peaked in 2004 and have been on a very shallow decline since--which reflects how it feels subjectiveluy to me.  I'm still relatively very good at the bike, but I'm not as good as I was 7-8 years ago.

I'm assigning full fitness potential value for my bike leg in 2012 of 98 %-tile.


Next, the run analysis:

2002: 79.1
2003: 72.4
2004: 82.4
2005: 88.4
2006: 80.6
2007: 86.9
2008: 77.6
2009: 78.7
2010: 77.4
2011: 80.7

At first blush, the run data is harder to decipher.  However, my history of knee (and related problems) drives these numbers.  In 2005 and 2007 I was healthy for the whole year and I have the highest %-tiles.  My 2006 number suffers from my shoulder and knee surgeries that year.  The last four years I've been plagued by an assortment of problems and have had significant training interuptions in March-May each year.  That's why I'm so focused on not getting injured this year and am more slowly developing my run.

I'm going to assign a number of 80 %-tile for 2012.  I still hope if I can stay healthy I might be able to do better than that.  On the other hand, my run has been the most negatively impacted by my age so maybe 80 will turn out to be optimistic.  We'll see.


Finally, my OA finish %-tile:

2002: 87.1
2003: 89.2
2004: 93.3
2005: 95.6
2006: 93.8
2007: 94.9
2008: 93.2
2009: 93.0
2010: 95.4
2011: 95.0

You can see how my bike leg has always carried me.  Even in the beginning when my rubn and swim were poor, my bike saved me.  As I brought my swim up in 2004 on I was able to move up into the 90+ zone and in 2005/2007, when my run was at it's best, I scored relatively higher.  Lastly note, the very beneficial impact that the improvements in my swim have had on my overall competitiveness in 2010 and 2011--you can teach an old dog new tricks sometimes!

I'll assign 95 %-tile as my overall race full potential number.


So my Empirical Fitness Potential Equation (EFPE) for 2012 is:

S +B+R=OA

(95%-tile)+(98%-tile)+(80%-tile)= 95 %-tile

So what?  Well, given this standard I can compare my 2012 races and determine how far away I am from this full fitness potential!

My first race this year was Smithfield:  Here are my %-tiles from that race:

swim: 75.2
bike:  94.9
run: 71.8
overall: 85.9

So obviously, I was not yet at my full potential fitness for this first race--no surprise, that is my training strategy!  So the interesting question that can now be answered, is how much faster would I have been had I been at my full potential?

If my swim had been at 95%-tile vs. 75.2 %-tile, my time official time would have been reduced from 5:18 to 4:18, a savings of 1:00.  This is not surprising, given it was a non-wetsuit swim.

Here are my total improvements as calculated by the EFPE:

Swim: 5:18-4:18=1:00
Bike: 26:37-25:23=1:14
Run: 22:35-20:49= 1:54
Total: 1:00+1:14+1:54=4:08

If I had been 4:08 faster, I would have finished 33rd OA (93 %-tile) vs. the 65th that I actually finished.

Now 4:08 is 7.2% of my total race time of 57:16.  In other words, I'm 7.2% away from my full potential fitness!  Also, note that almost half of my improvement potential is in the run.  All of this jives very well with my subjective assessment of my fitness level.

So, I expected to see a modest improvement come my second race at Bumpass.  How did I do?  My %-tiles were:

swim: 84.9
bike: 95.9
run: 69.2
overall: 88.6

The EFPE yields the following improvements:

Swim: 14:01-12:36= 1:25
Bike: 31:10-30:37= 0:33
Run: 25:27-23:21=2:06
Total: 1:25+0:33+2:06=4:04

If I had been 4:04 faster I would have finished 23rd OA (96 %-tile) vs. the 64th I did.

4:04 is 5.5% of my total race time of 73:53.

So on the surface, I've gone from being 7.2% away from being truly race fit to now 5.5%.  I seemed to make my biggest improvements in the swim and bike and now I'm going to have to start making some progress on my run if I want to reach my potential fitness.

So there it is.  A lot of questions about this analysis arise of-course.  For example, why not just look at at my FTP in my training tests to gauge my fitness on the bike.  Or my 100 yd repeat times on the swim.  Well, of course I do, but these measure my relative physical fitness in each of the three disciplines.  What I'm trying to measure is my RACE FITNESS--which includes my physical fitness plus things like tactics, pacing, etc.

I'm sure you have questions as well but as I've thought through all of the places that error could be interjected in this aanalysis, I've conmcluded that it's actually pretty valid.

We'll see how it holds up in 2012.  I'll sure be looking to get inside of 5% at Devilman.  I'm also going to back test the equation at some point so that will be interesting as well.

That's it for now!

No comments: