Training Stress Score

Today we're taking a break from our regularly scheduled blogging (Romeo and Juliet Pt 2 will have to wait) to discuss an article from VeloNews.com by Dirk Friel. His credentials far exceed mine, but that doesn't mean I can't snipe at him from a quasi-anonymous blog.

So he wrote a pretty interesting article entitled Tracking Power at the Tour of California. The long and short of it were that ProTour riders put out a lot of power. Duh. Still, it was cool to read just how powerful these guys are, and it was a well-written article.

I'm not pissed that the article was, at times, a thinly-veiled advertisement with the pretense of exposition. I get it, that's business. What gets me is the Training Stress Score introduced by Friel.

So Power = Force x Velocity... all this really means is that one can push a high gear at a low cadence or a low gear at a high cadence and be producing the same amount of power. Boring. You can use power feedback in real-time, glancing at your cute little monitor to gauge your exertion, or you can look at it post-hoc on your computer (I think only Handloff and Hoffman know the gesture that accompanies that word).

Well, if you're looking at things post-hoc, you might as well quantify the workout further. Work = Power x time, which seems simple enough, and it is. Tada, you have a measure of how much work you did, measured in kJ (kilojoules, which mustn't be confused with JigaWatts). This isn't rocket science; it's been around for quite some time.

Now here comes Friel with another metric of your workout, Training Stress Score (TSS). What's the difference between TSS and Work?

"The benefit of tracking TSS points over kilojoules is that TSS points are relative to one's threshold power. A kilojoule on the other hand is the energy it takes to produce watts over time, but does not take into account individual efficiencies, or lack thereof," says Friel.

Whoa whoa whoa, Dirky-poo. You've introduced a metric that does nothing more than normalize an already existing metric. You've scaled Work. You've divided.

Why does this piss me off so much? Hell, I sorta make a living concocting metrics (I called the last one G. It's the inverse of a previously existing metric). I guess the difference is that TSS takes information out of the users' hands. If you look at kJ and you know your Functional Threshold (FT), you know your TSS... but you also know your kJ and your FT. You can compare today's kJ with that of last year, and then you can compare kJ/FT. This will give you two pieces of information. TSS is only one. I know that not everybody has the training or resources to write the 3 lines of MATLAB code (or 3 hours of evil evil Labview) that would completely automate this process... but who would want to sacrifice the information afforded by already-existing metrics?

What it boils down to, for me, is this:
Knowledge is power (no pun intended), and by reducing the amount of information we analyze, we're reducing our capacity for self-knowledge. It is endemic of a fast-food society, where everything must be presented in neat, processed packages for our consumption. God forbid we get our hands a little dirty, playing with simple, meaningful numbers.

(I really hope Friel doesn't Google himself like Litespeed does.)