Average Weekly TSS?

germanboxers said:

Frenchyge, I think we are on the same page, probably even the same sentence of the same page. I was just responding to whoawhoa.

Here's a great example of an apples-to-oranges TSS comparison: rode for 39 miles with a reasonably fit friend in Jefferson County, MO (hilly ride) and averaged around 17mph for the 2h15m ride. We hammered up some of the hills. (TSS=161, IF=.822). The week before I rode 50 miles in NE Indiana (hilly, but not MO-like hills) with two slow rec riders averaging a little over 13mph for the ride. I did hammer ahead at times for bathroom and directional reasons and sprinted up a few hills, but in general my heartrate was literally as low as walking heartrate for most of the ride. (TSS=221, IF = .778). The day after this ride I felt guilty that I hadn't "trained". The day after the MO ride, I felt somewhat fatigued. Obviously, the fact that I hammered ahead and sprinted up some hills on the slow ride increased my IF significantly above what it would have been if I rode the same speed as my riding partners that day, but from a training load standpoint, I don't think it's even close to the MO ride, yet the MO ride had a significantly lower TSS. I think this illustrates (to me anyway) that the value of TSS lies in the normal range of training, not at the extremes (excessively long easy rides).

g/b,
What was your average power (AP) on both rides or what was your ratio of NP to AP? How about average HR on each?

On the ole Wattage list we call NP/AP VI (for variability index). This has been explored over there ... perhaps w/o concensus but IMHO there are a few loose ends with TSS ... epecially for slow/sloppy rides that could stand to be cleaned up.

rmur

rmur17 said:

g/b,
What was your average power (AP) on both rides or what was your ratio of NP to AP? How about average HR on each?

On the ole Wattage list we call NP/AP VI (for variability index). This has been explored over there ... perhaps w/o concensus but IMHO there are a few loose ends with TSS ... epecially for slow/sloppy rides that could stand to be cleaned up.

rmur

I think that's a good point, rmur. The NP/AP ratio on the very slow ride was much more than the faster ride. It's interesting also in that the faster ride in MO was pretty much up and down the whole ride. I would think this would lead to a higher NP/AP ratio because of the high variability in power. The slower ride consisted of plenty of smaller hills, but basically was steady SLOW SLOW, with a couple of sprints sprinkled (complete boredom), a couple of hill sprints, and a couple of reasonably high 2-3 minute intervals.

I may track this VI index...in conjunction with other info, sounds like another nice index to interpret the ride. Thanks.

BTW, anyone know when the next version of CyclingPeaks is coming out? I use and SRM Pro and unfortunately, with CP vers 1, I still have to open up the SRM software to delete the datablocks. Hunter mentioned this summer that ver2 would be able to do this directly from the CP.

germanboxers said:

it seems to me that the real value in TSS is comparing successive blocks of training, say daily, weekly, monthly for an individual so that within a particular period (base, build, taper, etc) an athlete can progressively increase training stimulus and load without overdoing it. I would bet that comparing TSS scores from base2 (more L2/L3 rides) with a build2 (L4, L5 intervals) or taper phase would not be as meaningful as comparing the TSS between a base2 and base3 or a week in base2 to a successive week in base2.

I won't claim that TSS is perfect, but I think it may be a bit better than you imply above. Here's a couple of examples of why I say that:

1) as you may know, uber-MTBer Dave Harris is a long-time TSS user, and hence has accumulated considerable experience using it to monitor his training in scenarios like you describe above. That experience leads him to schedule only a 10-15% reduction in TSS when moving from a "base"-type focus (e.g., long level 2 rides, tempo workouts, etc.) to VO2max focus (e.g., level 5 intervals done in block format). At the same time, his total training volume (hours) goes down by ~50%. So, while the current formula for calculating TSS may overweight duration and underweight intensity slightly, I don't think it does so by much.

2) my wife came close to "cracking" twice during the 2002 season while training for the pursuit at track nationals. On both occasions, her average weekly TSS had ramped up at a rate of ~5%/wk for 6-8 wks, maxing out at ~1000/wk. The first time, though, her training 20-25 h/wk of level 2/3/4 sessions, whereas the second time she was doing "only" 10-15 h/wk of level 5/6/7 work. Thus, much like Dave's experience above, comparable TSS levels appear to reflect fairly comparable training loads, even when the training focus changes considerably.

Having said the above, I nonetheless agree that you should never lose sight of the "composition" of the training that results in a particular TSS value, if only from the specificity perspective...

germanboxers said:

BTW, anyone know when the next version of CyclingPeaks is coming out?

All I can tell you (which is all I know) is that the programmer is hard at work, but no news on a release date yet.

germanboxers said:

I use and SRM Pro and unfortunately, with CP vers 1, I still have to open up the SRM software to delete the datablocks. Hunter mentioned this summer that ver2 would be able to do this directly from the CP.

While it is more work, there are a couple of advantages of downloading via the SRM software, then importing into CyclingPeaks (which is how I do it): 1) you automatically create a back-up copy of each file that way, and 2) you retain the ability to adjust the slope and/or offset used to calculate power should you determine in the future that this needs to be done (e.g., you've recalibrated your SRM, and discovered that the slope was off).

acoggan said:

... but I think it may be a bit better than you imply above.

Dr. Coggan, thanks for replying. I didn't mean to imply TSS wasn't useful for a wide variety of training, just that longer L1 rides (like riding with cyclists way below your fitness level) results in much higher TSS scores than I would have guessed based on feel. In looking closer at my files, an IF of .778 does suggest I rode "harder" than the 13.5mph average we maintained.

I certainly will defer to your judgments regarding comparisons between L2/L3 periods and L5/L6 periods. I haven't really structured my training this year in blocks so I only assumed those comparisons may not be as meaningful as would week-to-week periods within the same phase.

What is the highest IF you have heard of for a training ride of ~1h, assuming FT is correct? I would have guessed that it would be ~1.0 by definition, but seeing how significant hard efforts are to NP, and not having done a short VO2max interval ride this year (with the powermeter), I think it may be slightly higher than 1?

BTW, do you get out to Jefferson County, MO (south of St. Louis) for many rides? My in-laws live down there and I discovered some great roads down there.

Thanks,

Jordan

IMO, IF and TSS all derive from NP. The heart of all three computations is the power weighting embedded in NP. After studying NP and studying many of my rides of various durations, I have a high degree of confidence in NP. So much so that all of the considerable work I am doing on variable power pacing (that Andy thinks good cyclists do instinctively) is built on top of NP and the power/duration curve, plus of course a small matter of a little physics.

germanboxers said:

To put it more mathematically, although TSS is proportional to IF^2, the domain (likely range of numbers for IF or variation in IF) is significantly less than the domain for duration.

Sure. I fundamentally agree with you and frenchyge, but I guess in my experience, finding training rides outdoors such that you can finish a multi-hour workout with IF greater than 0.7 is difficult. You really need to plan such a workout carefully to involve sufficient hill work, etc., like rapdaddyo does (I believe), not have too many traffic lights, do it almost like an endurance race (pace-wise).

On the other hand, it is very "easy" to have a trainer workout of around one hour with an IF of 1 to 1.1. If you look at the squared ratio of 1.1 to 0.7, you get a result close to 2.5. That, in my mind, is large enough that saying "you'll get a lot more TSS through long rides" may be accurate theoretically, but maybe not so for the average committed enthusiast's training schedule constraints.

Berend

P.S.: I should have done this before I picked a number out of my, well, you know where, but after a quick perusal of my own training logs, I think 0.8 is a more representative IF for the "average committed cycling enthusiast's" multi-hour outdoor rides. That brings down the comparative squared ratio in my discussion above to 1.9.

squidwranglr said:

Sure. I fundamentally agree with you and frenchyge, but I guess in my experience, finding training rides outdoors such that you can finish a multi-hour workout with IF greater than 0.7 is difficult. You really need to plan such a workout carefully to involve sufficient hill work, etc., like rapdaddyo does (I believe), not have too many traffic lights, do it almost like an endurance race (pace-wise).

I agree with this observation. After studying a number of my training rides (as opposed to recovery rides), I concluded that I was over-recovering from my pushes (intervals). I'm not sure but this may be a common tendency for road rides. It may be that trainer rides are more tightly structured, possibly because it is so boring to ride a trainer that we want to minimize the session while realizing the training goals. The other point about intensity vs. duration is that beyond an hour FT is very misleading. I know that my 2-hr MP is significantly lower than FT and my 3-hr MP is even less than that. So, a 3-hr ride at IF>0.75 is a really hard ride. I know because after I re-designed all of my training rides I regularly have IF>0.75. So, I don't think TSS overstates the stress of duration vs. intensity. I find it quite helpful in evaluating my re-designed road rides.

squidwranglr said:

Sure. I fundamentally agree with you and frenchyge, but I guess in my experience, finding training rides outdoors such that you can finish a multi-hour workout with IF greater than 0.7 is difficult. You really need to plan such a workout carefully to involve sufficient hill work, etc., like rapdaddyo does (I believe), not have too many traffic lights, do it almost like an endurance race (pace-wise).

On the other hand, it is very "easy" to have a trainer workout of around one hour with an IF of 1 to 1.1. If you look at the squared ratio of 1.1 to 0.7, you get a result close to 2.5. That, in my mind, is large enough that saying "you'll get a lot more TSS through long rides" may be accurate theoretically, but maybe not so for the average committed enthusiast's training schedule constraints.

Berend

P.S.: I should have done this before I picked a number out of my, well, you know where, but after a quick perusal of my own training logs, I think 0.8 is a more representative IF for the "average committed cycling enthusiast's" multi-hour outdoor rides. That brings down the comparative squared ratio in my discussion above to 1.9.

It's basically impossible to reach IF=1.1 for a solid hour. If you're doing so, you've underestimated your FT. 1.05 is the max and I've hardly ever exceeded 1.03. In fact, when I exceed 1.02 IF for any hour in a hard training ride/race, I re-evaluate my FT and invariably it has increased.

But .. in general I agree it's possible to rack up more weekly TSS points in L2 than at L5/6 ... not by an awful lot ... perhaps 10-15% IME. But who has the time to ride that much or the opportunity this time of year?

'Sloppy' L1/2 rides are for recovery or riding with the family - not really training in my book.

I try and stick to the 0.85-0.95 ride IF (upper L3-lower L4) range to best balance the TSS load and the key training stimulii which lead to improvement in FT. After three full years training with power, I still believe FT is the single most important metric. Raise it and you lift the entire Power-Duration curve (in the aerobic region at least).

Hey RD, if you can sustain that 1500 TSS/week without rest weeks - let us know how you're doing it 😉 That average daily TSS load of 214 points is way higher than anyone else I've heard of.

rmur

rmur17 said:

I try and stick to the 0.85-0.95 ride IF (upper L3-lower L4) range to best balance the TSS load and the key training stimulii which lead to improvement in FT. After three full years training with power, I still believe FT is the single most important metric. Raise it and you lift the entire Power-Duration curve (in the aerobic region at least).

I agree with the comment about FT as the single most important fitness metric. But, I think IF needs to be viewed in the context of the ride duration. I learned the difference between my MP at 1 hr vs. 2 hrs in the Mt. Charleston Hill Climb. I think I believed previously that my MP would remain relatively flat >1 hr. Well, I was wrong. At least in my case, my power/duration curve declines steadily out to at least 4 hrs. An IF near 1.0 assumes a ride duration of ~1 hr.

RapDaddyo said:

I agree with this observation. After studying a number of my training rides (as opposed to recovery rides), I concluded that I was over-recovering from my pushes (intervals). I'm not sure but this may be a common tendency for road rides.

I'd started upshifting and pedalling down hills on my endurance rides because I got tired of seeing wimpy IF's at the end of all my outdoor rides. It requires a good bit of concentration to keep AP in L3 in the rolling terrain around here.

RapDaddyo said:

The other point about intensity vs. duration is that beyond an hour FT is very misleading. I know that my 2-hr MP is significantly lower than FT and my 3-hr MP is even less than that. So, a 3-hr ride at IF>0.75 is a really hard ride.

I wish I was at home so I could pull some actual numbers (or that Cycling Peaks could be used on more than one computer so I could see my files at work 😠 ) but those numbers seem pretty low to me. Last Saturday I rode solo with the intent of doing some outdoor tempo, and maintained AP = .89FT for 2 hrs (I believe NP was ~.93FT due to the rolling terrain). It was a tough ride, so maybe we're just talking about a matter of degrees here (although I was able to do a set of jumps at the end and rode about 30min home after that). I'm pretty sure I could ride for 3hrs+ at IF>.85, and 6hrs+ at IF>.75.

frenchyge said:

I wish I was at home so I could pull some actual numbers (or that Cycling Peaks could be used on more than one computer so I could see my files at work 😠 )

You can if you're willing to pay a monthly fee. gotomypc.com offers that service. I know because I use it. I can access all the files (and programs) on my PC from anywhere in the world, even from public computers.

frenchyge said:

those numbers seem pretty low to me. Last Saturday I rode solo with the intent of doing some outdoor tempo, and maintained AP = .89FT for 2 hrs (I believe NP was ~.93FT due to the rolling terrain). It was a tough ride, so maybe we're just talking about a matter of degrees here (although I was able to do a set of jumps at the end and rode about 30min home after that). I'm pretty sure I could ride for 3hrs+ at IF>.85, and 6hrs+ at IF>.75.

I think my MP declines ~5%/hr after 1 hr. For me, a 2 hr NP of .93FT would be ~98% of my 2hr MP. My main point is that FT is simply one point (~1 hr) on the power/duration curve and longer (or shorter) rides shouldn't really be compared to FT but rather to the power/duration curve at that duration.

RapDaddyo said:

You can if you're willing to pay a monthly fee. gotomypc.com offers that service. I know because I use it. I can access all the files (and programs) on my PC from anywhere in the world, even from public computers.

That'd be a good work-around for what I consider to be an overly restrictive security policy on a program I've purchased. Another option would be to purchase another copy of Cycling Peaks, but I'll step away from that soapbox for now. 😉

RapDaddyo said:

I think my MP declines ~5%/hr after 1 hr. For me, a 2 hr NP of .93FT would be ~98% of my 2hr MP. My main point is that FT is simply one point (~1 hr) on the power/duration curve and longer (or shorter) rides shouldn't really be compared to FT but rather to the power/duration curve at that duration.

Agree. I've always considered "Functional Threshold" to be a good term because to me it implies flexibility to the context in which it's being used (ie, the power threshold for the particular function in question: 10k TT, 50mi Road Race, 24hr endurance race, etc.), but it sounds like Andy intended it to mean ~1hr TT power and nothing more.

frenchyge said:

That'd be a good work-around for what I consider to be an overly restrictive security policy on a program I've purchased. Another option would be to purchase another copy of Cycling Peaks

And remember to copy your WKO files to both computers. Too much trouble.

frenchyge said:

Agree. I've always considered "Functional Threshold" to be a good term because to me it implies flexibility to the context in which it's being used (ie, the power threshold for the particular function in question: 10k TT, 50mi Road Race, 24hr endurance race, etc.), but it sounds like Andy intended it to mean ~1hr TT power and nothing more.

Yes, it is nothing more than that. I mean, 1 hr is a good point on the power/duration curve to represent fitness, not too long and not too short. And, it intersects with the 40K TT (at least for a relatively flat course with moderate winds), another common benchmark for fitness. But, in looking at individual rides of durations greater than or less than 1 hr, I ignore FT and look at the ride relative to my MP at that duration. That's why I do my intervals at 90%MP for that duration rather than as a percentage of FT. It makes a difference for L4 and L5 intervals in particular, where there is a pretty large duration range.

rmur17 said:

It's basically impossible to reach IF=1.1 for a solid hour. If you're doing so, you've underestimated your FT. 1.05 is the max and I've hardly ever exceeded 1.03. In fact, when I exceed 1.02 IF for any hour in a hard training ride/race, I re-evaluate my FT and invariably it has increased.

I think you're right rmur. Looking through the logs, the highest IF I see for a one hour interval session on the trainer is 1.03.

Berend

germanboxers said:

Dr. Coggan, thanks for replying. I didn't mean to imply TSS wasn't useful for a wide variety of training, just that longer L1 rides (like riding with cyclists way below your fitness level) results in much higher TSS scores than I would have guessed based on feel. In looking closer at my files, an IF of .778 does suggest I rode "harder" than the 13.5mph average we maintained.

I certainly will defer to your judgments regarding comparisons between L2/L3 periods and L5/L6 periods. I haven't really structured my training this year in blocks so I only assumed those comparisons may not be as meaningful as would week-to-week periods within the same phase.

What is the highest IF you have heard of for a training ride of ~1h, assuming FT is correct? I would have guessed that it would be ~1.0 by definition, but seeing how significant hard efforts are to NP, and not having done a short VO2max interval ride this year (with the powermeter), I think it may be slightly higher than 1?

BTW, do you get out to Jefferson County, MO (south of St. Louis) for many rides? My in-laws live down there and I discovered some great roads down there.

Thanks,

Jordan

Hi Jordan,

First, no need for the "Dr." - I'm just another bike rider like you. 🙂

As Rick Murphy pointed out, an IF of >1.05 for ~1 h is unusual, and suggests that either 1) functional threshold power has been underestimated, and/or 2) the algorithm has failed to do what it is supposed to do. By periodically polling the wattage list, I've managed to collect about two-dozen such rides - but that's from hundreds of workouts from 1000+ subscribers. So, while this can hardly be considered a random sample, I would argue (in fact, have argued) that their relative rarity is evidence that the algorithm is quite robust (which doesn't mean that it can't still be improved, however).

I actually live w/in a mile of Jefferson County, but to get there I'd have to ride (or drive, and I hate driving to rides) across the Meramac River on a six lane divided highway with no shoulder and tall concrete barriers on each side. Consequently, I do 99% of my riding in west St. Louis County, in what is now Wildwood, MO. Next time you're in town, though, drop me an email at acoggan at earthlink dot net and maybe we can get together for a JeffCo ride...

thurs 10/27 through thurs 11/3 - TSS was 542.

friday 11/4 through thurs 11/10 - TSS was 708.8

wow, that's kind of a big jump. that last week included a few pretty long rides on the weekend... of course i followed that 2 week block up with 3 totally off of the bike rest days this past weekend. so far this week (yesterday, today) i'm at 191.6 but i've been doing group rides, going way harder than i planned.

joule said:

Messed that poll up! Missed 600-800! Sorry.

Can you edit the poll to include that range??

rmur

acoggan said:

Hi Jordan,

I actually live w/in a mile of Jefferson County, but to get there I'd have to ride (or drive, and I hate driving to rides) across the Meramac River on a six lane divided highway with no shoulder and tall concrete barriers on each side. Consequently, I do 99% of my riding in west St. Louis County, in what is now Wildwood, MO. Next time you're in town, though, drop me an email at acoggan at earthlink dot net and maybe we can get together for a JeffCo ride...

Andy, six lanes, no shoulder, concrete barriers?? I can see why you head west! Around 15 years ago, I used to ride out in the Wildwood area, well reasonably close I think? We would take off from the Smokehouse (??) outside of Chesterfield and ride out and around St. Albans. There were some real nice climbs that usually broke my knees when I tried to ride them with an 12x18 7spd cluster (foolish 23 year old used to riding less challenging hills). I think one of the roads was "Wildwood"? I also remember climbing a hill we called "The Klondike", but I'm not sure if that was the same ride as the St. Albans ride?

I'll e-mail you next time I'm in town and hopefully get out for a ride. Probably won't be until next spring though.

rmur17 said:

Can you edit the poll to include that range??

rmur

No. Once a poll is created it cannot be edited. But it served a better purpose, to inspire a discussion on the merits of TSS and IF.