Quote:
Originally Posted by
An old Guy 
I don't consider Internet bulletin boards to be science. I made no comments about the content of the paper. I limited my comments to the abstract. That seems to be proper.
(Actually, I said I accepted the figure you posted from the paper as proof of whatever.)
---
I did not know people had to consult you before implementing your algorithm. I implemented your model on a spread sheet. I must have it just as wrong as the implementers of PowerAgent.
I had values for 3 rides actual rides and also math models of the rides. The actual numbers agreed with the model:
1) 75% FPT for an hour yields TSS of 56 or so.
2) 100% FTP for 1 hour yields TSS of 100. (By definition I believe.)
3) Alternate 0% and 150% intervals for an hour yields a TSS of 120 or so. (The length of the intervals don't matter in the computation so 30 or 60 seconds (for each of the 0% and 150%) gives the same results.)
The results of those 3 rides should demonstrate that your TSS concept has errors.
I think you or Allen Hunter wrote a paper indicating that the first and third rides above were very similar in demands - both have average 75% FTP for the same period. Yet they have vastly different TSSs.
---
It is amazing that you have the "right" model. There are a large number of other models that make claims similar to your claims. I am sure that they are claimed to be correct, but in the end they have flaws just as serious are your model.
I am not the first person to point out that your model is wrong. Other people have given examples.
Critiquing (or citing) a paper without actually having read it isn't "proper", it is lazy.
...
The algorithm for calculating TSS is actually quite simple, so you'd think people would be able to get it right w/o any help from me. Nonetheless, Saris was unable to do so, at least at first - hence the reason I suggested you should be careful interpreting the data provided by PowerAgent. It seem unlikely you would make the same mistake as they did when you hand-coded things, though, so presumably the numbers are at least in the right ballpark...assuming, of course, that you actually did such rides, and aren't just making things up. (Even if you did, you clearly are making things up when you claim that alternating every minutes between 150% and 0% of FTP for 1 h is an easy workout.)
...
You are confused (again): different TSS in the same duration = different normalized powers = different demands. The fact that the average powers are comparable obviously means nothing (which is why normalized power was invented).
...
I don't claim to have the "best" model - just the best I've been able to come up with so far (well, and publically share), and obviously better than anything you (or any other critic of normalized power/TSS/the Performance Manager) has ever proposed (because, in fact, you - and they - have never proposed any alternatives).
...
Your stance reminds me of that of Kraig Willett, who liked to claim that normalized power and TSS were horribly flawed and that it was easy (for him) to generate an "NP buster", but who never offered any alternative and who, when pressed, refused to provide evidence of the latter...
