Power test time

Marty Moose
Posts: 1421
Joined: Sun Jan 17, 2010 7:00 pm
Location: W.A

Re: Power test time

Postby Marty Moose » Wed Nov 03, 2010 10:57 pm

I love these power threads I know nothing about all the acronyms :) :) Better read up I'm getting one real soon hopefully weeks away. So anyone will to post up all the tss type stuff thrown around ?? Or book sites to look at !!

MM

User avatar
toolonglegs
Posts: 15463
Joined: Fri Jun 29, 2007 7:49 pm
Location: Somewhere with padded walls and really big hills!

Re: Power test time

Postby toolonglegs » Wed Nov 03, 2010 11:22 pm

Marty Moose wrote:I love these power threads I know nothing about all the acronyms :) :) Better read up I'm getting one real soon hopefully weeks away. So anyone will to post up all the tss type stuff thrown around ?? Or book sites to look at !!

MM
Grab this book...very good read even if you don't train with a meter.
http://www.amazon.com/gp/product/193403 ... QE01S6S8RJ

User avatar
twizzle
Posts: 6402
Joined: Wed Aug 13, 2008 11:45 am
Location: Highlands of Wales.

Re: Power test time

Postby twizzle » Thu Nov 04, 2010 10:38 am

Out of interest TLL, what are your ATL/CTL constants set to, and did anyone ever tell you that they need to be tuned to each athlete?
I ride, therefore I am. But don't ride into harm's way.
...real cyclists don't have squeaky chains...

User avatar
DanielS
Posts: 512
Joined: Sat Apr 19, 2008 8:00 pm
Location: Adelaide
Contact:

Re: Power test time

Postby DanielS » Thu Nov 04, 2010 11:16 am

I never knew there was a training metric named after me! :lol:

User avatar
Alex Simmons/RST
Expert
Posts: 4997
Joined: Fri May 02, 2008 3:51 pm
Contact:

Re: Power test time

Postby Alex Simmons/RST » Thu Nov 04, 2010 12:55 pm

twizzle wrote:Sure, tell you what, I'll switch my training to riding Corin every day. That gives me 45 minutes of climbing, and then I get free TSS points for the ~10 minute roll to the bottom of the hill. How useless will the 'forest' be after 42 days? As for the 'forest', should we start discussing the need to tune the ATL/CTL periods to suit each athlete? Or why the default of 42 days suspiciously happens to be the same as the six week block recommended by Jack Daniels? Like any model : garbage in, garbage out.
Go for it, but you are missing the point(s). Indeed if that's what training you do do, then you would find the impulse-response model of your choice would behave exactly the same, just with slightly different absolute values.

So on what basis is one garbage and the other not?

You're the one that said there was a rationale for why Daniels Points are better, so all I wanted to know what that was (other than the subjective sense of the programmer in question that they felt the balance between intensity and duration wasn't right for them). No need to get all defensive. If you don't actually have a sound reason as to why it is better, that's OK. I just thought you might be able to help enlighten me because I haven't found one myself.

In fact it's kind of cool it's that way, because the modeling is not overly sensitive to these things, such that you can make up your own system, and provided it's not hideously off the mark in assigning some relationship with intensity and duration, then you can still make sense of it. Still, nothing wrong with looking for a better mousetrap. Dr Coggan has one up his sleeve in fact.

As for tuning ATL/CTL, well that's a bit of a red herring since training stress scores (DP, TSS, Bikescore, TRIMP etc) come before those models are applied. Besides, anyone who looks at the modeling for that process should know that the amount of formal testing required to make such a model truly predictive of TCs with error ranges that make sense is well beyond anything practical to implement (unless you plan to formally test yourself with a maximal effort every one, or maybe every two, days of your riding life).

Raceday does this analysis but what it doesn't do is show you the error range (which is often quite large, such that for the most part, one can with reasonable experience make just as good, if not a better estimate simply by knowing their athlete). Self coached riders simply need to inspect various ATL TCs to see under what circumstances they best reflect actual performance.

As a coach, I'd love a better way to do it, because it is time consuming to do such analysis but I wonder how many actually train with such consistency that these things matter? If you are an athlete who's CTL is south of 100 and has regular falls in CTL due to inconsistent training, well this really is the leaves on the trees of the forest.

The 42 days for CTL TC was chosen as it correlates with the typical time frame for relevant physiological adaptations and as nominal based on the scientific literature. It was also tested for sensitivity, such that changing the duration by a week or so either side really doesn't make a lot of difference to the outcome of the models. The ATL TC however is another matter.

If Daniel's came to a 6 week TC through observational experience/empirical means, it's probably no surprise that they are similar since anything wildly different most likely would not have worked.

As for threads, this is a debate that has gone on for many years, many times at wattage forum (google groups and in the pre google topica days) as well as other places.

lukas
Posts: 865
Joined: Tue Jul 24, 2007 4:32 pm
Location: Hurstville, Sydney

Re: Power test time

Postby lukas » Thu Nov 04, 2010 1:18 pm

Hey guys..

I just recently picked up a wired powertap. I’ve not done a formal FTP test, but have been doing 1 or 2 threshold interval sessions per week (2x20min). In the last 8 weeks I’ve gone from doing these at approx 215w to approx 245-250w. FYI I’m currently 72kg (aiming to get back down to 68-69kg which was pre european holiday, where I ate and drank lager a like a king)

I’m waiting on my copy of ‘training and racing..’ to arrive and at this stage, I’m a bit of a noob to it all still. Should I bother doing a FTP test or just use my 2 x20 min power as my FTP? I’m doing these intervals pretty hard.. generally pretty cooked by the end of the 2nd, but not totally obliterated.

The idea of an hour TT FTP test does not appeal to me at all. The 20min TT x 0.95 test sounds fine, but I’ve read it’s not always accurate. So is there any point, or should I just keep plugging away slowly trying to improve my 2x20mins, safe in the knowledge that my FTP is climbing with them?

FYI , I’m doing all the 2x20’s in the same condition, on my trainer in my garage. It’s tough finding suitable roads to do them on. Too hilly or too many traffic lights.
Cervelo R3
Canyon Ultimate AL
Image

User avatar
twizzle
Posts: 6402
Joined: Wed Aug 13, 2008 11:45 am
Location: Highlands of Wales.

Re: Power test time

Postby twizzle » Thu Nov 04, 2010 4:17 pm

I agree with most of what you have said, just a couple of points :
Go for it, but you are missing the point(s). Indeed if that's what training you do do, then you would find the impulse-response model of your choice would behave exactly the same, just with slightly different absolute values.
But for managing ramp rate etc., accuracy starts to matter - assuming the recommendations about TSB ranges for avoiding injury are correct. Although, once again, I suspect that's an 'average person'.
So on what basis is one garbage and the other not?
I was referring to the Performance Manager model, not the scoring model. And what I mean was that if the input is 'Garbage' due to incorrect inputs (both ride data and ATL/CTL constants), then the information coming out the other side is going to be incorrect.
You're the one that said there was a rationale for why Daniels Points are better, so all I wanted to know what that was (other than the subjective sense of the programmer in question that they felt the balance between intensity and duration wasn't right for them). No need to get all defensive. If you don't actually have a sound reason as to why it is better, that's OK. I just thought you might be able to help enlighten me because I haven't found one myself.
No, I said it makes more sense - to me, anyway. Obviously, that's subjective.
I ride, therefore I am. But don't ride into harm's way.
...real cyclists don't have squeaky chains...

User avatar
toolonglegs
Posts: 15463
Joined: Fri Jun 29, 2007 7:49 pm
Location: Somewhere with padded walls and really big hills!

Re: Power test time

Postby toolonglegs » Thu Nov 04, 2010 4:55 pm

twizzle wrote:Out of interest TLL, what are your ATL/CTL constants set to, and did anyone ever tell you that they need to be tuned to each athlete?
Just the standards....my 1st month on power was a bit pointless,but I was pretty fit so I started everything roughly where I thought I was...and I was still under where I should of been.After that it didn't matter that much...if I use my power meter again I will be starting with an FTP so low that it will hardly matter.But I was thinking of selling my power meter,but as it stands at the moment in can stay in the garage till I get back from NZ in a few months...if I am not riding by then I probably never will be.

User avatar
twizzle
Posts: 6402
Joined: Wed Aug 13, 2008 11:45 am
Location: Highlands of Wales.

Re: Power test time

Postby twizzle » Thu Nov 04, 2010 5:13 pm

toolonglegs wrote:
twizzle wrote:Out of interest TLL, what are your ATL/CTL constants set to, and did anyone ever tell you that they need to be tuned to each athlete?
Just the standards....my 1st month on power was a bit pointless,but I was pretty fit so I started everything roughly where I thought I was...and I was still under where I should of been.After that it didn't matter that much...if I use my power meter again I will be starting with an FTP so low that it will hardly matter.But I was thinking of selling my power meter,but as it stands at the moment in can stay in the garage till I get back from NZ in a few months...if I am not riding by then I probably never will be.
:( - on all counts.

Hunter Allen mentioned 'tuning' the constants recently on the wattage forum, I haven't digested the whole concept yet. Just another variable to try and understand... :?
I ride, therefore I am. But don't ride into harm's way.
...real cyclists don't have squeaky chains...

User avatar
Alex Simmons/RST
Expert
Posts: 4997
Joined: Fri May 02, 2008 3:51 pm
Contact:

Re: Power test time

Postby Alex Simmons/RST » Thu Nov 04, 2010 7:12 pm

twizzle wrote:I agree with most of what you have said, just a couple of points :
Go for it, but you are missing the point(s). Indeed if that's what training you do do, then you would find the impulse-response model of your choice would behave exactly the same, just with slightly different absolute values.
But for managing ramp rate etc., accuracy starts to matter - assuming the recommendations about TSB ranges for avoiding injury are correct. Although, once again, I suspect that's an 'average person'.
Well in that case, the ramp rate (or TSB) would simply be lower or higher if you use a different metric. All that's really changing is the overall relative daily scores.

These things are ranges, not absolutes. For CTL growth we suggest 2 to 8 points per week over the medium term. That's a heck of a wide band. Some I can't push more then 3, others I can push at 6+. But that comes from knowing your athlete, not by looking at some model.

If you think your riding typically results in "inflated" TSS (or at least in your subjective sense, but regularly), then the PMC and "ideal" CTL rates and any assessment of tweaking time constants simply adjusts to that input.

You might be able to sustain a ramp of 6 under one score method and 5 under another.

If your riding involves the occasional effort with "inflated" TSS, then it's just noise in a relatively insensitive model, and I would challenge one to know the difference between a weekly CTL ramp of 4.0 and 4.1.

I suppose I've been through all the playing with TSS variants some time ago now, and have learned to focus on what's important, and to filter out what's not.

User avatar
Alex Simmons/RST
Expert
Posts: 4997
Joined: Fri May 02, 2008 3:51 pm
Contact:

Re: Power test time

Postby Alex Simmons/RST » Thu Nov 04, 2010 7:12 pm

twizzle wrote:Get the development build of GC, the one dated August 15'th.
By the way, version 2.0 came out the other day.

User avatar
Alex Simmons/RST
Expert
Posts: 4997
Joined: Fri May 02, 2008 3:51 pm
Contact:

Re: Power test time

Postby Alex Simmons/RST » Thu Nov 04, 2010 7:20 pm

twizzle wrote:I was referring to the Performance Manager model, not the scoring model. And what I mean was that if the input is 'Garbage' due to incorrect inputs (both ride data and ATL/CTL constants), then the information coming out the other side is going to be incorrect.
Sure, GIGO applies to everything.

Still, the PMC and TSS do a remarkably good job of predicting stuff that it says it can. Even with default TC values and what some subjectively think are "inflated" TSS.

User avatar
twizzle
Posts: 6402
Joined: Wed Aug 13, 2008 11:45 am
Location: Highlands of Wales.

Re: Power test time

Postby twizzle » Thu Nov 04, 2010 9:40 pm

Alex Simmons/RST wrote:These things are ranges, not absolutes. For CTL growth we suggest 2 to 8 points per week over the medium term. That's a heck of a wide band. Some I can't push more then 3, others I can push at 6+. But that comes from knowing your athlete, not by looking at some model.
I've been doing three week blocks at 5 - 7 CTL ramp a week, but due to my crappy life schedule this means pushing TSB down between -25 and -30.
If you think your riding typically results in "inflated" TSS (or at least in your subjective sense, but regularly), then the PMC and "ideal" CTL rates and any assessment of tweaking time constants simply adjusts to that input.

You might be able to sustain a ramp of 6 under one score method and 5 under another.

If your riding involves the occasional effort with "inflated" TSS, then it's just noise in a relatively insensitive model, and I would challenge one to know the difference between a weekly CTL ramp of 4.0 and 4.1.

I suppose I've been through all the playing with TSS variants some time ago now, and have learned to focus on what's important, and to filter out what's not.
My worry is that I might be missing out on important training, going on long rides is O.K., but playing in the traffic is where I get the hard efforts and really push myself - but then I have blocks of time stuck at lights or caught behind slow riders. Or if I do hills, the big ones around here (local) are ~ 14 minutes for the climb and 7 minutes for the descent/turn, and I still haven't got a handle on whether 100TSS on the hills is anywhere close to 100TSS of track intervals.

I must be doing something right, I'm five seconds a lap quicker at Stromlo than I was six months ago.
I ride, therefore I am. But don't ride into harm's way.
...real cyclists don't have squeaky chains...

User avatar
Alex Simmons/RST
Expert
Posts: 4997
Joined: Fri May 02, 2008 3:51 pm
Contact:

Re: Power test time

Postby Alex Simmons/RST » Fri Nov 05, 2010 7:29 pm

twizzle wrote:I must be doing something right, I'm five seconds a lap quicker at Stromlo than I was six months ago.
Well there you go!

Who is online

Users browsing this forum: No registered users