Wednesday, March 25, 2009


No, this isn’t a chat about online dating websites! It’s about a method to quickly identify when, during a race, you “burnt a match”.

The concept of a “burning a match” isn’t a new one in cycling – basically it’s a metaphor for saying you did a hard effort, hard enough that it might impact on your ability to do other hard efforts later on in the race, since a match can only be used once. While not a perfect analogy, it’s not a bad one.

How many matches we can burn and how brightly and for how long those matches can shine for might be thought of as the size and quality of our personal matchbook. It’s one measure of our race specific fitness.

Performing well is as much about knowing when to "light one of your matches" as it is about doing the training to build up the size and quality of your matches and the number of matches in your matchbook.

Burning a match is also relative to the race in question. In a 3-week Grand Tour, a match might be akin to a long solo/small group break away or an attack on an alpine ascent, but in a 20 or 30-km points race on the track, it would be an attempt to gain a lap, or go for lots of sprints.

Typically, burned matches are attempts at race winning moves
or, for some, race survival moves.
In any case, often riders like to analyse their power meter files to see when, how many and how big were the matches they set off while racing. It's one way to help assess, post-hoc, the tactical decisions taken while racing (did you really burn an unnecessary match going on a fruitless escapade?) or simply as a way to assess how many and how often matches were lit before you cracked, or even an indication that your matchbook is getting bigger with training.

In the book, Training and Racing with a Power Meter, it shows one way to locate such efforts by using the fast find feature in the WKO+ software. I’m going to show another method, one, that with just a little bit of spreadsheet help, is pretty easy to do and which shows up matches quite clearly.

Again, as with much of what I write, it’s not an original thought. It is based on the Normalised Power concept developed by Andy Coggan and just such a chart can be seen on Slide 15 of his PowerPoint presentation hosted on Google docs:
“Making sense out of apparent chaos: analyzing data from on bike power meters”

OK, so let’s look at an example of what I’m talking about.

Let's take my points race at the recent State Masters track championships. Here’s what the power meter trace looks like for the race:

It shows my power output for the race, as well as horizontal lines showing my Functional Threshold Power (FTP) and my Maximal Aerobic Power (MAP). As is typical with these sorts of races, the power output is highly variable and while you can see some spikes, it is difficult to make all that much sense out of the information presented like this.

But with a little bit of maths applied to the power data, here is a plot of exactly the same race:

Now this really shows up where I burned my matches. In this instance it clearly shows the 6 sprints in which I either contested (the first and the last) or simply needed to put the power down to stay in the race (the other 4). It also shows the periods where I didn’t sprint at all (sprints are every 10 laps in these races about every 2.5 - 3 minutes depending on the race speed), which is when I was in a chasing group as laps were being taken/lost by various riders.

By showing the data in this way, it is really clear when matches were lit. My first one was a pretty big flare, as not only was there a strong sprint but it was clearly an extended effort, perhaps covering an attempted attack after the sprint. But it also shows that after a couple of sprints, I simply didn’t respond when the inevitable attack came. I needed another match but my book was getting a bit thin at that stage.

Here are a couple of other similar plots:

This one is from the State Masters Criterium Championships in 2006:

As is evident from this plot, the first 10-minutes were pretty brutal with some very hard efforts necessary. In this period the field was whittled down to a break of just 6 riders. Then the break settled somewhat, before some more attacks started in an effort to ‘break the break”. It is also clear that on this course, if you couldn’t repeatedly make such hard efforts, you would be toast, limping back to the hotel for an early shower. It was a "repeatedly go hard and recover" kind of course.

Here’s another example from a different type of crit race:

This time I made a solo break very early, then was joined by another rider after about 15-minutes or so and we stayed together up to the finish. You can see the large match early on, and then the smaller efforts while solo, which diminished somewhat when I was joined by my break companion and we established and consolidated our lead. This enabled me to save a big match for when it really counted - the final sprint.

So how are the above plots made?

Well it’s not hard and if you know about Normalised Power, then you’re well on your way.

1. Just take a normal power meter file and open it in Excel (or your preferred spreadsheet software).
2. Then calculate a rolling 30-second average of the second by second power data.
3. Then raise that rolling 30-second power value to the 4th power (watts^4).
4. Then chart that 30-second power raised to the 4th power by time.

That’s it.

I also added lines to show both FTP and MAP raised to the 4th power.

Hint: The chart is a x-y scatter plot, with horizontal (x) values being time and corresponding vertical axis (y) values the power^4 values.

Why 30-second averaging and why raise to the 4th power?

Well, in essence the 30-second averaging and the raising to the 4th power is because (and I quote from Andy's own item on Normalised Power):

  1. the physiological responses to rapid changes in exercise intensity are not instantaneous, but follow a predictable time course, and
  2. many critical physiological responses (e.g., glycogen utilization, lactate production, stress hormone levels) are curvilinearly, rather than linearly, related to exercise intensity

As for #1, as Andy has shown us in the Google docs presentation referenced above, the half lives of many physiological responses to the intensity we are riding at (i.e. our power output) are indeed not instantaneous. The time period for such responses to show up is typically around 30-seconds or so. Some, and I quote, include: PCr kinetics, heart rate/cardiac output and sweating all having half life response times of around 25 seconds. VO2 ~ 30-seconds, VCO2 ~ 45 seconds, ventilation ~ 50-seconds and core temperature changes ~ 70-seconds.

So from the point of view of assessing our body’s responses to intensity (power output), it makes sense to view power meter data as a rolling average power over a 30-second window*. It doesn’t actually have to be 30-seconds but changing the duration of rolling average (to say 25-seconds or to 40-seconds) doesn’t have a large impact on the outcome of the plots.

* except perhaps when assessing maximal neuromuscular sprint type efforts, since the energy systems in use fatigue over a handful of seconds (although our "recoverability" for sprints is still linked to our aerobic or "matchbook" fitness).

This is also why we sometimes refer to things like heart rate as being a "lag indicator" of effort and is one reason why HR is a poor guide to managing shorter harder efforts while training.

As for the #2, the research Andy shows suggests an exponential relationship exists between blood lactate concentration and power expressed as a ratio of 1-hour power (power:FTP, often referred to as the Intensity Factor). In the same presentation one can see (on slide 13) that the best fit for the data shows a relationship very close to the 4th power (3.91).

Again, the use of a nice even number of 4 rather than say 3.9 is simply more convenient and choosing numbers either side really doesn’t affect the nature of the plot all that much.

So if you cracked in a race, or couldn't go with the winning break, then what does your match analysis look like? Did you not have a match when it counted, or did you not use them wisely enough?

Just remember, playing with matches is dangerous, so take care out there!

Read More......

Friday, March 20, 2009

Championship Comeback

This past weekend, I participated in the NSW State Masters Track Cycling Championships. I say participate as I was not really likely to pose a threat to the podium in any of the three events I entered.

On Friday night was the Time Trial, which for my category is a 750 metre (3-lap) blast around the Dunc Gray Velodrome. Pretty simpe event really. Once the starting gate releases your bike, you go as hard as you can until you complete the 3 laps.

I was hoping to break 62-seconds.

Just before my final warm up I went to put on my leg but something was wrong, the cleat was twisted out of place.

Thinking it was just the lower clamp bolt that for some reason was loose, I got on the bike to get the cleat angle right and have a buddy tighten the clamp for me.

Except it didn't look too good to tighten.

That's when we discovered a stress crack/fracture in the lower clamp on my prosthetic bike leg attachment which meant the lower clamp-cleat section was completely loose (you could pull it off the end of the pylon). Wasn't there day before.... I know, I gave it a good look over and clean.

Managed to carefully tighten the bolt enough so the section wouldn't move with hand pressure at the same time as to not widen the crack anymore. But I gotta say, it was at risk of major fail in the TT. Bugger it I thought. You came here to ride. If you crash, so what? I've had worse.

But it didn't fail, although I was probably a little tentative out of the gates.

I ended up going better than a 62 with a 60.646 second ride. I figure if Chris Hoy has to do 4 laps to my 3, I'm in with a shot!!

The track time trial is not my natural event but the post-accident Alex is rediscovering all things bike, and so I'm going back to riding a range of events rather than necessarily specialise at this time.

Besides, this year is all about re-building a broken body.

Fortunately, I was able to find a spare clamp on an old "foot" I had in the cupboard, so that went on the leg and all is good again. Who'd have thought you'd need spares for your leg! LOL.

Next up was the 20km Points Race on Saturday afternoon. Now this is my natural event but nowdays I have an unnatural body. So what would happen? I stuck on a pretty tall gear (for me) - a 96.4" (50x14) as I was pretty sure I would need it.

Small field of 13. Still some class riders with world and national masters champs amongst them. Mission: well just get out there and race, and what will be will be.

So with that in mind we rolled out, and before long the first of 8 sprints (which are every 10-laps of the 80-lap race) was on in earnest. Hang on, what am I doing at the pointy end? Hmmm, OK well you're in the slot, so roll with it. 2 points - hey the race has only just started and you're on the board already!

Then it started to get interesting and most of it between laps 10 and 70 is a bit of a blur. But one by one the laps ticked off and I was still there. More than half didn't finish, being spat out the back. I did end up losing a lap (and 20 points with it) along with a small group when the bunch was split. In the end, only 6 riders finished and I was thereby placed 6th, equal on points with Tony from Canberra. Somehow I even managed to help out a couple of team mates along the way. That was fun. Hard but fun.

So after that lot, I had to back up for the 3km Individual Pursuit on Sunday morning. Of course I had crunched the numbers and set myself a, for the time being, challenge of going 4:06 for the 3,000 metre event.

I did a 4:10 but that's all I had in the tank. I would liked to have ridden faster but legs doth protesteth and I knew I had to just run what I brung.

But then again, it was a 4-min power PB so one can't be too harsh on oneself.

It was good to be back.

(and before all you power meter freaks out there are worried about me not posting any data - fret not - I'll have some more techno geekery analysis to share in the next installment).

Read More......

Sunday, March 08, 2009

Old Skool

Time trialling is a popular part of the sport of cycle racing. The "race of truth" is deceptively simple. Ride a set course by yourself and try to finish it in the shortest time possible. Fastest time wins. Easy.

As always with this sport, there are a number of elements that go into riding the fastest time you can. But in essence, they come down to the three P's:

1. Power to the pedals
2. Piercing the wind
3. Pacing the course

If you don't address all three "P's", you are riding sub-optimally from a "least time" perspective.

Power is all about the training you do to improve your fitness and your ability to ride harder. The power and training part of the equation is something I've written lots about on this blog.

Piercing the wind is all about the aerodynamics of your bike and especially your position on the bike. Given this is about the physical elements that resist our forward motion, I also include a few other items, like having the best tyres for the lowest rolling resistance and an efficient drivetrain so power is not "lost" between the cranks and the rear wheel.

Pacing is the skill part of the equation and is all about knowing how hard to push yourself at various times on the course. Pacing is something I've also written quite a bit about (but not here on my blog), and for which I have developed some mathematical modelling. I'll write a bit here about that one day.

Anyway, what's all this fancy pants maths got to do with "Old Skool"?

Well last weekend I rode a TT. It's our "local" TT run by the NSW branch of the Australian Time Trials Associations (ATTA) which is held on the first Sunday each month on the Calga course, about an hour's drive north of Sydney.

I decided to get back to doing some TTs for a few reasons. Firstly, since they are a race against the clock, once you have set a time then you have a benchmark for improvements. In essence you are racing only against yourself. They are an excellent fitness marker, especially when you are measuring your power output.

Secondly, I have also decided to race the paracycling national road championships (late April-early May), which includes a road TT (and a road race). So getting some similar length TTs under my belt seems like a good idea.

Finally, as I keep on with my return to the bike, I need to rediscover the type of competitive outlets that will better suit my new "body". So going back to riding many different types of events is one way to re-learn all about that.

Anyone who's ridden a TT, or even just watched one, will know that many riders use special bikes which are designed to improve the aerodynamics of the rider and bicycle. Much thought goes into this as it can result in big savings in time for the same physical effort.

However, some riders decide to ride TTs "Old Skool" style, meaning they use a standard road bike with standard wheels and helmet. No special effort is made to improve aerodynamics. That's what I did last week.

For some fun, below is the graphical summary of my ride and pacing analysis.

There's a bunch of stuff on there I won't go into today but in essence it shows the course elevation profile (it's an out and back course, with the first half mostly climbing) with elevation data obtained via two methods I won't describe here now, and two lines showing power - the yellow line is my actual power output (a 40-second rolling average) and the green line is a theoretical optimal power output for that segment of the course. By comparing the two lines you can see how well or otherwise I dosed out my effort along the way. So thats the Power and the Pacing elements.

There are also a range of other numbers shown, one of which is an estimate of my Coefficient of Aerodynamic Drag and Effective Frontal Area (CdA), which is a measure of how well (or not) I slice through the wind. Thats the "Piercing the wind" element.

OK, so how did I go?

Time: 42:55 @ 35km/h.
Average Power: 264 watts
Normalised Power: 268 watts

Not too bad given it was on standard road bike, helmet and wheels and the constantly up 'n' down Calga course. That placed me 25th out of 45 riders.

I was hoping for more like 270-280W (my first 20-min were at an average of 277W and holding myself in check) but I struggled with my prosthetic leg getting loose after about 15-20 minutes which saw 2nd half power fall somewhat and quite a bit of stump pain in the latter stages. My stump was very sore afterwards.

But that was exactly the point of the exercise - to find out the issues I face and need to fix when doing this sort of an effort. On a course like Calga, which is never flat and the power is somewhat more variable as a result, then the leg does tend to work loose a bit more quickly than say on a flatter road or on the ergobike where the effort is far steadier.

Main thing is to work on a solution for keeping my leg packed firmly in the socket during a longer hard effort. Otherwise it hurts and more importantly, you lose power. Maybe I needed to use more foam packing (which is what I'm doing at the moment). I probably didn't put enough in. Then there is the socket itself, which my rehab doc and prosthetics specialist says no longer fits properly (I have lost weight and my stump has changed its shape and volume again).

Of course I am also getting a special cycling leg made up, thanks to the generous donations and fund raising by the track cycling community here in Sydney. That is now being planned. I have the appointment to recast my stump on 2nd of April and from there we can begin the production of a new leg socket and a new general purpose leg. I can hardly wait!

So at the Calga TT next month I may be able to borrow a TT bike and see what difference it can make to my time. The analysis above shows a CdA of 0.33m^2 (square metres), which is very high for anyone doing a TT. Certainly riding on a TT bike I would expect that number to be much lower. Even at a modest 0.28m^2 (certainly readily attainable on a TT set up) I could expect to ride that same course in the same conditions (wind, temperature & air pressure) at the same power about 2-minutes faster.

We'll see.

Read More......