Two Wheel Blogs

So I've been watching this thread entitled Visualization of TSS Composition. I don't know if you can view that link without signing up for wattage, sorry if not. Then again, if you're reading this you're probably already signed up on wattage. Or you should be.

It must have been 3 years ago when I first thought of this very same idea. I think that coming up with the idea is part of the natural progression a cyclist goes through when he or she starts training with power. I think this is why it comes up about every year, just about the time a new year's crop of power users starts to get this training with power thing.

Now, let me be clear about this. All this discussion is a good thing. I hope that somewhere along the line it will lead some much more intelligent and creative mind than mine to devise a more useful tool with which to visualize TSS. I don't think it's impossible. I do, however, think most people don't immediately understand how difficult a problem this is. That is, if it's really a problem at all.

So, back to my so-called discovery some three years ago. I can't remember if I wrote to or spoke to Andy about it, but his reply was the same as it is today. What method can you use to identify when an interval begins and ends? For example does this particular 3½ minute effort stand on its own as a L6 interval, or is just part of some longer interval where you happened to be riding in the wind? How do you decide whether this 3½ minutes was something you intended to do at that point in time; or was just a consequence of the weather, or how well you slept last night?

Now, you probably easily define what some particular interval in training time really is for yourself under controlled circumstances. That's particularly easy to do on a trainer, for instance. But how do you write a macro in Excel or how do you write a program which can look at a ride's wattage data and decide what that workout is, or was?

To illustrate what I'm talking about, have a look at these two charts. They're both the same chart. They're just oriented at different angles and zoom factors to give you some idea how complex the problem of analyzing ride data can be. The charts both show one of my recent group rides. The X-axis is time in seconds. The Z-axis represents the percentage of Functional Threshold Power I was working at.

The Y-axis is the one that's a little harder to grasp. It's a succession of power output curves. Each successive curve is the same power data, but it is displayed with a longer smoothing factor over the one before. The most jagged, spiked end of the chart, where y=1, is unsmoothed %FTP. The most smoothed end uses a smoothing factor y=3600 seconds, an hour.

I've set the colors to transition roughly from deep blue out to deep red according to training levels we all know and love. Blue is roughly recovery and endurance. Red is L6 and beyond.

So, can you see the problem here? Where exactly do you start to cut this monster up to decide what kind of workout it was? If you smooth it out completely you end up with something that looks like a tempo ride. Cut it up into 30 second chunks and now it looks a bit more like a VO2max workout. It isn't really logical to account for some segments using a 30 second smoothing and others at 5 seconds, or 120 seconds, is it? Myself, I can't really say which is the best smoothing factor for this ride. Any recommendations?

And unless you can say what smoothing factor to use, how can you apportion TSS? And even if you did, you'd always end up either compiling TSS points separated in time into one (How can that be meaningful?), or you'll undercount TSS in the upper intensity levels. Under counting isn't a bad idea if you're doing long intervals; but if you're workout goal is L5 or L6 work, you really don't want to do that.

Here's an example of what I'm talking about. Recognize this one? It should be easier to interpret. It's a 2x20 interval workout I did earlier this year. You can see the 1200 second blocks right in the middle of the chart. I did these intervals outdoors. Like any outdoor workout, there is plenty of variability. But if you smooth this one out using a 20 minute smoothing constant, you see that my intervals were just about right in the L5 space - sort of an orange-red. Using a smaller smoothing constant turns this 2x20 minute workout into something that might look more L5-ish, or even L6-ish.

But it obviously wouldn't serve the purpose of my workout goal to smooth this by a shorter duration than 20 minutes. If I did though, I'd find that I'd earned a bunch of my TSS points in L5 and even L6 during my 2x20 intervals. So what was it, a Threshold workout or an aerobic power workout?

And that's my point in all this. Only by knowing what it was you wanted to achieve in a workout can you evaluate whether you achieved it. No software can uncover the mystery goal for every workout. Even if it could, would it do any good? After all, if you don't have a goal for your workouts, why are you calling them workouts? Put another way, why do you need software to tell you after the fact what your plan was for your workout?

The more I think about it the less I care about where my TSS comes from. I don't believe in wasting my time accumulating junk TSS by riding junk miles. I believe as long as my workouts follow my plan, every TSS point is meaningful, and it does me no further good to try to classify it. What matters to me is not so much what kind of TSS, but how much TSS.