Power on the downhills (Page 2)
-
No new posts
Moderators: k9car363, alicefoeller | Reply |
|
![]() ![]() |
Champion ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() | ![]() Interesting discussion....some great insightful information. It took me a summer of TTs after having a power meter couple years to help me learn to pace evenly. I must be odd as I find it easier to keep power on downhills than most other terrain assuming the grade isn't so steep I run out gears and tuck in and coast. I find often down hills that aren't super steep is when I have the highest power output in a race or training ride and it feels the easiest. I can easily go 10-20 watts higher. I work hard and keeping my vi low in riding as I do long course. I had the experience in my first IM of riding to much of the ride above zone 3 due to not having proper gearing. I'll need to look at what Dave and Robert posted to see if that helps me understand more. |
|
![]() ![]() |
Master ![]() ![]() ![]() ![]() ![]() ![]() | ![]() The Simmons article that Michael posted doesn't seem to say how his Pacing Optimisation Index (POI) is calculated, and as far as I can tell it focuses more on easily quantifiable variables like gradient, wind, CdA, etc. It may be the case that these are indeed the most important considerations for the short TTs that the roadies are mostly concerned with. But that seems quite different from some of the main factors that folks here are raising, which may be more relevant to triathletes: -- effect of riding sections well above LT on subsequent ability to run well; that part is surely affected by varying individual abilities to clear waste products from their system during the remainder of the ride -- effect of even sub-threshold power fluctuations on whatever-the-hell-it-is that gets depleted during 5-6 or more hours of lower intensity IM riding; again presumably subject to individual variation and training I'm guessing that these are roughly the physiologic limiters that Dave L. is talking about experienced riders getting a feel for. But I wonder how any of that gets incorporated into the fancy models. |
![]() ![]() |
![]() ![]() ![]() ![]() ![]() | ![]() colinphillips - 2011-05-24 7:25 PM The Simmons article that Michael posted doesn't seem to say how his Pacing Optimisation Index (POI) is calculated, and as far as I can tell it focuses more on easily quantifiable variables like gradient, wind, CdA, etc. It may be the case that these are indeed the most important considerations for the short TTs that the roadies are mostly concerned with. But that seems quite different from some of the main factors that folks here are raising, which may be more relevant to triathletes: -- effect of riding sections well above LT on subsequent ability to run well; that part is surely affected by varying individual abilities to clear waste products from their system during the remainder of the ride -- effect of even sub-threshold power fluctuations on whatever-the-hell-it-is that gets depleted during 5-6 or more hours of lower intensity IM riding; again presumably subject to individual variation and training I'm guessing that these are roughly the physiologic limiters that Dave L. is talking about experienced riders getting a feel for. But I wonder how any of that gets incorporated into the fancy models. Every optimization scheme has to have some way to constrain the power you can generate. Alex's POI uses NP as its global constraint, and the pacing model "allocates" NP subject to course conditions (mostly gradient based on VE) in a way to minimize overall time. NP is nonlinear and it penalizes excursions above FTP far more than an equally long amount of time below threshold (i.e., the cost function is asymmetric). So the part you were missing is the NP constraint. But the NP constraint is just a convenience -- almost any nonlinear (asymmetric) cost function works as well. The key is as I said above: the best TT'ers turn out to have very very good pacing in a way that closely mimics what the models predict. IOW, the models appear to be validated in the real world by real riders in real races. |
![]() ![]() |
Master ![]() ![]() ![]() ![]() ![]() ![]() | ![]() RChung - 2011-05-24 11:08 PM But the NP constraint is just a convenience -- almost any nonlinear (asymmetric) cost function works as well. The key is as I said above: the best TT'ers turn out to have very very good pacing in a way that closely mimics what the models predict. IOW, the models appear to be validated in the real world by real riders in real races. Interesting. Thanks for contributing to this discussion! I'm curious: do the model-rider comparisons show the same effects for short TTs vs. longer events, e.g. IM? A couple of people have commented in this thread that even sub-threshold excursions become more damaging as the events get longer. In effect, the suggestion is that the asymmetric cost function changes with event duration. |
![]() ![]() |
Champion ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() | ![]() colinphillips - 2011-05-24 11:58 PM RChung - 2011-05-24 11:08 PM But the NP constraint is just a convenience -- almost any nonlinear (asymmetric) cost function works as well. The key is as I said above: the best TT'ers turn out to have very very good pacing in a way that closely mimics what the models predict. IOW, the models appear to be validated in the real world by real riders in real races. Interesting. Thanks for contributing to this discussion! I'm curious: do the model-rider comparisons show the same effects for short TTs vs. longer events, e.g. IM? A couple of people have commented in this thread that even sub-threshold excursions become more damaging as the events get longer. In effect, the suggestion is that the asymmetric cost function changes with event duration. I think that the main issue abot the relevance of these models has more to do with the penalty incurred on running. As I understand it, the models in question are focused on pure tt efforts. The run in a tri brings in another variable that is not necessarily covered in these models. That doesn't mean that they are irrelevant -- far from it -- but it does mean that we need to pay attention to another facet of this complex situation. That's just my inexpert reaction -- perhaps Robert or other experts will chime in on this point. IMO, the models are useful because they isolate at least some of the things we should be thinking about, and some of the things to which we should be attentive, when developing a pacing strategy. |
![]() ![]() |
Not a Coach ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() | ![]() The experts may provide better explanations, but the models are generally indifferent to whether you are doing just a bike TT or doing a long run after finishing your 112mi TT. All that changes are your constraints. Now figuring out what your constraint should be so that you can run well may not be an easy task, but it's a crucial input. There are some people who have attempted to stratify where those constraints appear to lie for many athletes (they've used TSS limits to help figure where NP should be targeted) and that's probably a reasonable place to begin for most--especially if you aren't in the tails of the finish time bell curve (in which case their models, unlike the one discussed here, DO appear to break down). |
|
![]() ![]() |
Master ![]() ![]() ![]() ![]() ![]() ![]() ![]() | ![]() RChung - 2011-05-24 10:49 AM It's because of the crank inertial load. When you're climbing you're typically using a gear ratio that produces low crank inertial load, while when you're descending you're typically using a gear ratio that produces high crank inertial load. Many (most?) riders have difficulty modulating their power when the CIL is high so they have to concentrate more to keep power steady (and high). Conversely, when CIL is low many (most?) riders have immediate feedback that helps them to modulate their power and keep it steady (and high). [Edit:] CIL varies with the fourth power of gear ratio. So, if you're climbing a hill in a 39/26 = 1.5 gear ratio that's low CIL while if you're descending in a 52/13 = 4.0 that's high CIL. Almost every time you post I learn something new. Thanks for the insights. Edited by sand101 2011-05-25 9:05 AM |
![]() ![]() |
![]() ![]() ![]() ![]() ![]() | ![]() JohnnyKay - 2011-05-25 6:36 AM The experts may provide better explanations, but the models are generally indifferent to whether you are doing just a bike TT or doing a long run after finishing your 112mi TT. All that changes are your constraints. Now figuring out what your constraint should be so that you can run well may not be an easy task, but it's a crucial input. Exactly. In a short TT you expect that your average power can be high so your constraint is high; in a 180km TT your average power is lower so your constraint is lower. If you need to run after that your constraint will be lower still. How you optimize given your constraint appears to be captured by the models pretty well, and it appears to mimic what real athletes, whether short course TT'ers or long course IM'ers, do in real races on real roads. The issue of your power-duration curve that sets the constraint is a separate issue than how one paces given a particular power-duration. There are some people who have attempted to stratify where those constraints appear to lie for many athletes (they've used TSS limits to help figure where NP should be targeted) and that's probably a reasonable place to begin for most--especially if you aren't in the tails of the finish time bell curve (in which case their models, unlike the one discussed here, DO appear to break down). Right. The "TSS budget" tends to break down if you're on the course for a long time. |
|