RChung

I do.

PepeM

Interesting. Still, I guess you take into account the way you get there (long slow ride vs some sprint intervals for example) rather than just do whatever gets you to a specific number.

RChung

Yup. TSS is a summary, sort of a short-cut to understanding the quantity-quality trade-off. Over time I've figured out what kind of TSS volume I can sustain, and the boundaries on quality (IF) and quantity (hours) that I can trade-off to get to my target. When time is short I can substitute a bit more intensity (within reason) and when I've got more time I can substitute more volume (within reason).

Seattle Forrest

I do this, too. Based on the PMC, where I am now, and where I want to be for an upcoming event.

kkapdolee

Thanks. I think I understand now.

PepeM

That is sort of what I was trying to get at. TSS tells you something about your session, but setting it as your goal and then doing whatever it takes to achieve it wouldn't be beneficial. Using it to know how much load you can take is fine though and actually I believe is the point of the metric in the first place. But what matters is what you do within the TSS you're allowing yourself for the session. Or at least that is how I understand it, I am far from an advanced user.

Seattle Forrest

Yeah it's definitely not "whatever it takes to get to 125 TSS" or whatever the number is. Generally I think if I don't need much TSS, I'll probably do hill repeats. Mostly though I'm after bigger numbers, which mostly come from Z2 rides. After several days of long, steady rides, my TSB is trending up and I can afford a day of less TSS, which means shorter and more intense efforts.

When I describe it like that it doesn't sound like chasing TSS at all. Maybe it's more fair to just describe it as polarized training. I guess the difference is that since I got a PM I've been thinking less in terms of miles, minutes, or heart rate, and more in terms of power. So it isn't "ride for 90 minutes" it's "ride for 100 TSS."

RChung

Coggan says he first started with the idea of the PMC, which was a way to think about the balance between "fitness" and "freshness." So he needed a way to quantify training load or training stress, and for this he needed a way to get equivalent load from different kinds of rides. So in a big picture macro view kind of way, he started off with PMC, and then long-term training stress or load (which he calls CTL) to model fitness, and short-term training load or stress (which he calls ATL) to model freshness or fatigue. So he needed TSS to build up his CTL and ATL. And he needed NP so he could compare steady and highly variable rides in order to get his TSS. Each step is a simplified model, and they're chained together to get to the PMC. If you understand the models and the chain, you can figure out the critical parts.

DaveLeeNC

While Autopause will certainly affect your #'s, I'm not sure that turning it off is the way to get the best data. But I guess it depends on what you are trying to learn.

Take two cases

1) A full out one hour session at a constant power output = ftp (IF = 1.0)

2) The session above divided into two sessions of 30 minutes with a period of time between 30 minute sessions that will be autopaused out (if autopause is enabled).

The IF for #1 is (by definition) 1.0. And the IF for #2 is some amount less than 1.0 (depending on how much dead time you put in between the sessions). The TSS score for #1 will be 100. And if autopause is on then the TSS score for #2 will also be 100, and I think that most folks would agree that these are not the same training stress.

HOWEVER, if you leave autopause OFF, while the IF will go down the TSS score will go UP in all cases. This happens because the amount of time in the total exercise increases more than the resulting IF (squared) is reduced. Just do the math - it isn't horribly complicated.

So in my mind leaving autopause off yields TSS data that is 'more wrong' than when you leave it on. BTW, I don't think too many people would leave off power data on a long descent (where you basically are not pedaling) from their data. And I'm not sure why a long stoplight enforced pause is any different.

dave

ps. The Coggan/Allen Training and Racing with Power Book states a somewhat overly complicated equation for calculating TSS once you know NP and duration. It is

(S x NP x IF)/(FTP x 3600) x 100 (where S is duration of exercise in seconds)

A simpler form would be 100 * IF**2 x D where D is the duration of the exercise in hours (as a decimal #).

Bah Humbug

I've never found myself able to recover at the rate predicted - a 2x20 at FTP is something I can only do a couple of times a week, but it's well under 100TSS. It's certainly not something I can do every day. Achieving the same TSS with a longer ride certainly allows recover much more quickly, but that means I recover from workouts of the same TSS differently based on how the TSS is achieved, which... goes rather against the whole point.

I've never had anyone be able to explain that to me, other than telling me the formulas are perfect and clearly I'm wrong about how trashed I am and I'm just a wuss for not being able to repeat the workouts.

asgelle

How are you determining FTP?

RChung

Right, we pointed this out back in, oh, 2003. A less convoluted example is when you do two workouts a day, such as when you commute. This observation was the genesis of a long argument of when to split a ride into two, and what the "consequence" of small differences in TSS were (that is, at the time we were trying to figure out how sensitive training was to small differences in TSS). Note that this is a problem related to the underlying issue of how to cumulate TSS over a period of several days. For odd historical reasons, we referred to those arguments collectively as "the beer and burrito issue."

Yeah, Coggan's original presentation used the former formulation to explain the underlying logic. When he explains using the computationally simpler version, people question the IF^2 term. When he explains using the longer heuristic, people point out the IF^2 form.

Bah Humbug

The 20-minute protocol. It predicts my triathlon pacing fine, and the ability to run after them. I can also climb and chase others of known FTP as I'd expect. A 2x20@100% though - that kicks me square in the quads, and trying to do it again the next day leaves my quads on fire before I'm halfway through the first and then they fail a few minutes later.

DaveLeeNC

Well, that is interesting. When I first encountered these equations it occurred to me "what is this - it is like someone is trying to hide what this really is". Guess I was (kind of) right.

Fascinating and it reminds of something that I did back in my engineering days. I was dealing with electronic circuit delays and we had good data on nominal performance, but the worst case performance was nebulous (and my problem). So for each circuit configuration I had to come up with a 'worst case factor' which you multiplied against the nominal performance to get worst case performance. Do to various correlation things, one factor came out to pretty much EXACTLY 1.0. If I published (to the engineers in our company) a factor of 1.0, I would have spent the rest of my life explaining to every damn one of them why worst case was not worse than nominal. So I just fudged it and published 1.05 and never got a 'call back'.

Thanks for the info.

dave

RChung

You doing the testing and the 2x20's on the road, or on a trainer, or a mix?

Bah Humbug

Trainer.

It's all rather a moot point now, honestly (though I remain curious), since I haven't put wheel on road in eight months or so, and might not again.

RChung

I suspected that when you mentioned your quads as the problem. 2x20's are really designed to develop your aerobic system, and the TSS you get accrue from that shouldn't be so great that you'd be limited to only doing that twice a week (I hate 2x20's so the mental load is huge but in terms of my aerobic system it's not really a problem) . On the other hand, the crank inertial load on a trainer is quite different than what most riders experience on the road and some people are much more sensitive to crank inertial load than others. So when you mentioned muscular stress rather than aerobic stress, that was a clue.

I suspect that if you were to do 2x20's on the road you wouldn't have difficulty recovering.

Bah Humbug

Oh, interesting, thanks for the explanation! That bugged me for years.

And so by "inertial load" you mean the nearly-200lbs of me + bike moving down the road makes the pedaling easier?

RChung

Almost. The inertia is what keeps you rolling down the road (or resists your change in speed when you accelerate). The crank inertial load is the feedback you get through the cranks and pedals, and that's a function both of your speed and your gearing (if you were going down the road at 20mph while pedaling a 50x13 your inertia will the same as if you were going down the road at 20 mph while pedaling a 34x28, but your crank inertial load would be very different).

kkapdolee

Hmm... I just did the math as I was skeptical and I think you're right. TSS goes UP when there is a break in between sessions with autopause OFF. That is odd and doesn't seem to make sense in physiological stress point.

I need to keep autopause on then to be 'less wrong'.

DaveLeeNC

Just a general comment here. I was not intending my observation regarding 'breaks' and TSS to be criticism. As RChung said "don't let good and useful get sacrificed in the search for perfect" (kind of what he said - too lazy to go look it up ).

dave

PepeM

https://www.facebook.com/TrainingAnd...49275731761396

DaveLeeNC

I would just say that if you are going to focus on the TSS #'s that are typically calculated by various software packages then NEVER/NEVER/NEVER turn off autopause.

If you want to know your true average speed then always turn it off. If you want to know how fast you ride when you aren't waiting for stoplights, then turn it on.

If you want to know your true NP over the entire workout, then turn AP off. If you want to know your relative effort level when you are riding then turn it on.

It depends on what you want to know. I will often add some errands onto the end of a 'real ride' (with no pauses) and will end the 'real ride' at the grocery store or bank ... Then the data from the real ride has meaning and I, quite frankly, have no idea how to set anything to yield useful (to me) training information about those last 4 miles, wandering past stoplights and in/out of traffic other than it is some # of additional training minutes.

You just need to know what is being calculated and how.

dave

Seattle Forrest

You could break those last 4 miles into a 2nd ride to collect information on them while separating them from your "real" riding. If you don't use laps, you could create a lap 4 miles from the end and then get your NP and TSS for each section from the splits.

DaveLeeNC

Sorry for the poorly worded post, as that is exactly what I do (separating the last 4 miles into a separate ride) even though I don't find that data very revealing.

dave