The Lifting Apologists....

[RapDaddyo]

I think Rap's app mod idea would be useful for pulling data points from files if you only included the testing files. I'm not sure why you'd put a duration constraint on it if this was what you were doing though.

Do you want to include every observation > x watts? For example, if the lower boundary is 1000W, you want to include every obseration >1000W? I was thinking about 2 criteria (user definable), time (e.g., 5sec) and power (e.g., 1000W). As to the trailing edge cutoff, I would just snip it when watts falls below the lower boundary (e.g., 1000W). Why put a (maximum) time limit on it? If you look at the watts time line of typical sprints, they look like Mt. Everest on both sides (steep slope up, steep slope down), so they don't tend to sort of "wave hop" just above the lower boundary.

I think this is a good idea, although, as mentioned above, I think you would only want to apply it to actual testing or racing files (eg from flying 2's and kilo's).

That would really be up to the user, to select the ride files to be included in generating the master file of individual efforts.

Also, I'd like to know why you think it's necessary to put a minimum time on the efforts. I think it would be important to 'snip' the data at 6s to stop inclusion of data where you have 'fallen off the curve'.

The minimum duration criterion would be up to the user. If you don't include a minimum duration, you could end up with efforts consisting of a single observation (e.g., 1.26s for PT).

So here is what I think the batch app might look like:

Enter minimum time P recorded for (optional)
Enter maximum time before snip (optional)
Enter minimum AP

Rest like normal...

But, this only speaks to which efforts to be included. My question is what data to extract for the included efforts?

 

[Lucy_Aspenwind]

But, this only speaks to which efforts to be included. My question is what data to extract for the included efforts?

Rap - at a bare minimum, we would need enough to do the analysis Andy pointed out - so that would be at least cadence, torque, and power. I'm sure there are other data points that might be helpful.

I might be interested in knowing how many times I exceed a certain power/duration criteria.

This seems quite relevant to the discussion and would appear to be rather valuable information.

Hey Lucy - don't want to make you jealous but I have at least 6 tracks within 2 hours drive (maybe more DGV, Goulburn, Adcock Park, Lidcombe, Hurstville, Tempe, Unanderra...) with one 10 minutes away (Tempe outdoor 333, just refurbished) and DGV 30 minutes away (OK 50-60 min at peak hour). Add another hour and there would be another half dozen in the regional towns. Most small towns here have a track of sorts, ususally goes round a sports oval and not steep banked like a velodrome but they're everywhere. We have indoor wooden board velodromes in Sydney, Melbourne (x2), Perth, Adelaide & Launceston. Other major centres would have an outdoor velodrome.

Right. Not wanting to make me jealous mate? Go figure, I live in a country with a population almost 15 times larger, but with barely half as many velodromes.

No, not even a wee bit of envy here, even if I do look unusually green in this post!

 

[RapDaddyo]

Rap - at a bare minimum, we would need enough to do the analysis Andy pointed out - so that would be at least cadence, torque, and power. I'm sure there are other data points that might be helpful.

Well, I need to be a lot more specific than that. One really has to work backwards from the data analysis that one wants to do. Think of it this way. Go to your ride files that have max effort sprints. Each effort has from 4-16 observations (at the PT observation frequency of one observation per 1.26s). Each observation has several data elements: watts, torque, speed, cadence, HR. Several statistics can be computed for each variable (e.g., min, max, avg, rate of change from prior observation, etc.). Now think about collecting a large number of such efforts and their associated statistics for further statistical analysis. What exactly do you want to do and therefore what exact data do you need? What charts do you want to produce? Just saying, "cadence, torque and power" is not sufficiently precise and specific. Think deeper.

 

[Lucy_Aspenwind]

Well, I need to be a lot more specific than that. One really has to work backwards from the data analysis that one wants to do. Think of it this way. Go to your ride files that have max effort sprints. Each effort has from 4-16 observations (at the PT observation frequency of one observation per 1.26s). Each observation has several data elements: watts, torque, speed, cadence, HR. Several statistics can be computed for each variable (e.g., min, max, avg, rate of change from prior observation, etc.). Now think about collecting a large number of such efforts and their associated statistics for further statistical analysis. What exactly do you want to do and therefore what exact data do you need? What charts do you want to produce? Just saying, "cadence, torque and power" is not sufficiently precise and specific. Think deeper.

Sorry mate, but I'm not a statistician, my background is in liberal arts.

I'd say cadence, torque, and power are basic, but then so would the data Andy suggested earlier......Avg, Min, Max, theoretical max....Pedal Force, Pedal Velocity. Given that those two would help quanitfy what is the limiting factor in the sprint....speed or force.....compared to the data for the AIS sprinters.

What else?

Much as I would like to, I don't know enough about the subject now to be able to formulate the basis for a thorough analysis.

 

[RapDaddyo]

Sorry mate, but I'm not a statistician, my background is in liberal arts.

I guessed.

Much as I would like to, I don't know enough about the subject now to be able to formulate the basis for a thorough analysis.

If I thought you did, I would have been much tougher above. Don't worry, we've got enough data maestros on the forum. We just have to catch their attention. The summary statistics for each variable (max, min, avg) over the entire duration of each effort that meets the criteria (whatever they are) is an easy one -- just do them all. It's the statistics that attempt to capture peak efforts that get more interesting and complex. For example, let's say you want to plot the average of the highest 2 consecutive watts observations against the corresponding averages for torque and cadence. You can't do this from the typical average statistics. You have to test all pair of observations, find the largest, then average them and the corresponding variables. You really have to reverse engineer the master data file by assuming you have unlimited computational capability and asking yourself, "Okay, what do I want to know and why do I want to know it?"

 

[Roadie_scum]

Do you want to include every observation > x watts? For example, if the lower boundary is 1000W, you want to include every obseration >1000W? I was thinking about 2 criteria (user definable), time (e.g., 5sec) and power (e.g., 1000W). As to the trailing edge cutoff, I would just snip it when watts falls below the lower boundary (e.g., 1000W). Why put a (maximum) time limit on it? If you look at the watts time line of typical sprints, they look like Mt. Everest on both sides (steep slope up, steep slope down), so they don't tend to sort of "wave hop" just above the lower boundary.

I was saying time limit because, in my understanding, this exercise requires only maximal data be used, and only the start of the effort will be truly maximal. It's not like a wingate where you are also interested in the trail-off, although that might also be interesting in a different context.

That would really be up to the user, to select the ride files to be included in generating the master file of individual efforts.

The minimum duration criterion would be up to the user. If you don't include a minimum duration, you could end up with efforts consisting of a single observation (e.g., 1.26s for PT).

If the single observation is maximal and occurs whilst fresh I think this would be a valid data point for the exercise Andy is discussing (correct me if I'm wrong please someone).

But, this only speaks to which efforts to be included. My question is what data to extract for the included efforts?

Cadence, Torque, Power, Speed - For every datapoint

Avg Speed, Avg Power, HR (jk) - For each individual effort

 

[mikesbytes]

nope on both counts! i live in uk and ride it most days on my winter bike why do you want one so badly

I have the rear wheel and would like to score a front one. Usually riders wipe out the rear wheel and then end up buying a set.

 

[RapDaddyo]

I was saying time limit because, in my understanding, this exercise requires only maximal data be used, and only the start of the effort will be truly maximal. It's not like a wingate where you are also interested in the trail-off, although that might also be interesting in a different context.

But, by definition these efforts are very short so we're not talking about many observations even if you capture the entire effort > x watts (assuming you set the lower boundary high enough).

If the single observation is maximal and occurs whilst fresh I think this would be a valid data point for the exercise Andy is discussing (correct me if I'm wrong please someone).

I was thinking that excluding single observations would prevent the occasional erroneous data observation from the PT hub. I've never been able to achieve a near-max power observation with a single burst. The pattern of my sprints is typically 3-4 observations to reach peak power. But, it doesn't really matter because what I have in mind leaves the minimum duration up to the user (and zero minimum duration is an option).

Cadence, Torque, Power, Speed - For every datapoint

Avg Speed, Avg Power, HR (jk) - For each individual effort

So, you're really talking about two output files. One file is one row per observation, effectively a subset (albeit a very small subset) of the original ride file, with each row of each qualifying segment. The second file is one row per segment, with summary data about the segment.

Let's cut to the chase. What are you going to do with the data? Are you thinking about a regression analysis? Which variables? What charts do you have in mind? The subsequent analysis and charting requirements really define the data requirements in the most pragmatic way.

 

[RapDaddyo]

I was saying time limit because, in my understanding, this exercise requires only maximal data be used, and only the start of the effort will be truly maximal. It's not like a wingate where you are also interested in the trail-off, although that might also be interesting in a different context.

If the single observation is maximal and occurs whilst fresh I think this would be a valid data point for the exercise Andy is discussing (correct me if I'm wrong please someone).

Cadence, Torque, Power, Speed - For every datapoint

Avg Speed, Avg Power, HR (jk) - For each individual effort

Here are my first thoughts about a research app for sprint efforts using ride files from standard power meters (as opposed to special lab equipment such as that discussed in the pubmed document that Andy referenced). The ride file segments to be extracted for further analysis would be those that met a user-definable combination of duration (in seconds) and power (in watts). But, since the duration criterion can be zero, the criteria can be reduced to simply power. The segments would be truncated when power drops below the power criterion (with no time limit). So, the criteria can be set as loose or as tight as you want. It's equivalent to slicing the tops off of a range of mountains. If you set a high boundary, you're not going to get many mountains and you're going to get less of each mountain.

The app would generate two files for the efforts (segments) that meet the criteria. One file would be a detail file, with one row per observation. So, if the app extracts 100 segments with an average of 5 observations per segment (e.g., 6.3 secs with a PT), the file would have 500 rows. The 2nd file would be a summary file, with one row per segment. In the above example, the 2nd file would have 100 rows.

My thoughts on the variables in the 1st file are:
file name
date
time - this is basically the unique index within the file
each of the variables from the power meter (watts, cadence, torque, etc.)
tss - basically, the cumulative TSS of the ride at that point (proxy for cumulative fatigue?)
prior 5mins tss - basically, the TSS of the 5mins preceding the effort (another proxy for fatigue?)

My thouhts on the variables in the 2nd file are:
file name
date
time of the first observation (to facilitate locating the effort in the ride file and to enable linking with the detail file)
duration
avg, max & min of each of the power meter variables (watts, cadence, torque, etc.)
tss - basically, the cumulative TSS of the ride at the start of the segment (proxy for cumulative fatigue?)
prior 5mins tss - as above

After thinking about the ride files and efforts to be included, my thought is to let the data speak for themselves. Do you want to limit the analysis to max efforts (based on RPE) or max performance (based on observed performance)?

Thoughts?

 

[Alex Simmons]

After thinking about the ride files and efforts to be included, my thought is to let the data speak for themselves. Do you want to limit the analysis to max efforts (based on RPE) or max performance (based on observed performance)?

Thoughts?

I'm trying not to think too much today.....

 

[RapDaddyo]

I'm trying not to think too much today.....

LOL. Okay, when your brain starts working again, let me know your thoughts. But, don't wait too long. First good golf tournament that attracts my attention and it'll be done. There's a tournament coming up with Tiger Woods and a few others. It will be done by the end of that tournament.

P.S., the tournament starts tonight!

 

[Lucy_Aspenwind]

LOL. Okay, when your brain starts working again, let me know your thoughts. But, don't wait too long. First good golf tournament that attracts my attention and it'll be done. There's a tournament coming up with Tiger Woods and a few others. It will be done by the end of that tournament.

P.S., the tournament starts tonight!

Hey, so I had a go at what Andy presented earlier in this thread. Basically the whole Maximum theoretical CPVmax & AEPFmax discussion that was mentioned previously.

I used the best 3 seconds of data from 5 standing starts and one lead-out sprint. Also, I used the intercept function for the x & y data range to obtain the max values for the pedal velocity & force.

Basically I came up with something on the order of AEPFmax = 1218 N & CPVmax = 2.561. Based on this, given it may be speculative, leg speed seems to be much more of a limiting factor.

I may well have made a mess of things and don't know that these numbers are correct - it is all getting a bit sophisticated for my statistical abilities.

 

[mikesbytes]

Hey Lucy - don't want to make you jealous but I have at least 6 tracks within 2 hours drive (maybe more DGV, Goulburn, Adcock Park, Lidcombe, Hurstville, Tempe, Unanderra...) with one 10 minutes away (Tempe outdoor 333, just refurbished) and DGV 30 minutes away (OK 50-60 min at peak hour). Add another hour and there would be another half dozen in the regional towns. Most small towns here have a track of sorts, ususally goes round a sports oval and not steep banked like a velodrome but they're everywhere. We have indoor wooden board velodromes in Sydney, Melbourne (x2), Perth, Adelaide & Launceston. Other major centres would have an outdoor velodrome.

Sydney use to have more velodromes in the past. We may of lost a few but we also gained 2: Tempe in the early 80's? to replace camperdown? and Dunk Gray Velodrome (DGV) built for the 2000 Olympics.

Alex, I assume with have included Maitland under the extra hours drive. Its susposed to be the tightest outdoor velodrome in Australia?

Where the hell is Adock Park, Hurstville and Unandarra velodromes?

Right. Not wanting to make me jealous mate? Go figure, I live in a country with a population almost 15 times larger, but with barely half as many velodromes.

No, not even a wee bit of envy here, even if I do look unusually green in this post!

Come on Lucy admit it...

 

[Alex Simmons]

Sydney use to have more velodromes in the past.
Where the hell is Adock Park, Hurstville and Unandarra velodromes?

I said tracks, not velodromes:
Adock Park - West Gosford - racing there this Saturday night - Gosford track open
Hurstville - er, Hurstville
Unandarra - Wollongong

We have lost over time Camperdown and Wiley Park, no doubt others. Merrylands as well I think.

Yeah - Maitland and Canberra's Narrabunda velodromes are an extra hour, Newcastle as well. So many tracks, so little time....

 

[mikesbytes]

I said tracks, not velodromes:
Adock Park - West Gosford - racing there this Saturday night - Gosford track open
Hurstville - er, Hurstville
Unandarra - Wollongong

We have lost over time Camperdown and Wiley Park, no doubt others. Merrylands as well I think.

Yeah - Maitland and Canberra's Narrabunda velodromes are an extra hour, Newcastle as well. So many tracks, so little time....

Marrickville as well, you can still see it from the air as you land at the airport coming in over the city for to the north south airport, they mounted lights in the middle of it and then covered it with seats.

Yeh I figured that Hurstsville was in Hurstville, but I thought the only thing there was one of those dredded Westfield shopping centres.

 

[acoggan]

Hey, so I had a go at what Andy presented earlier in this thread. Basically the whole Maximum theoretical CPVmax & AEPFmax discussion that was mentioned previously.

I used the best 3 seconds of data from 5 standing starts and one lead-out sprint. Also, I used the intercept function for the x & y data range to obtain the max values for the pedal velocity & force.

Basically I came up with something on the order of AEPFmax = 1218 N & CPVmax = 2.561. Based on this, given it may be speculative, leg speed seems to be much more of a limiting factor.

I may well have made a mess of things and don't know that these numbers are correct - it is all getting a bit sophisticated for my statistical abilities.

Hmm. Those results don't really fit with your vertical jump being 60 cm. Can you post a scatter plot showing the raw data points?

Another thought: since the PowerTap records data at a relatively low frequency (i.e., once every 1.26 s), your cadence (pedal velocity) may be significantly underestimated in this situation since it represents the average over that time interval. I've never tried such a test using a PowerTap, though, so I'm only speculating here.

 

[Lucy_Aspenwind]

Hmm. Those results don't really fit with your vertical jump being 60 cm. Can you post a scatter plot showing the raw data points?

Another thought: since the PowerTap records data at a relatively low frequency (i.e., once every 1.26 s), your cadence (pedal velocity) may be significantly underestimated in this situation since it represents the average over that time interval. I've never tried such a test using a PowerTap, though, so I'm only speculating here.

Andy, I don't have the scatter plot here in my office. Anyway, I excluded a few points because the numbers seemed too high, if I included those, I recall the CPVmax ~ 177 rpm's. I am uncertain if that fits in better with the vertical jump.

Back to the original question, this analysis was aimed at determining what was more of a limiting factor....pedal force or pedal velocity. Can I say one way or another at this point, which is more of a constraint for me?

You know the 1.26s recording interval seems more than suitable for conventional road riding, but as we have had this conversation, I see it has some limitations for track racing/riding.

 

[acoggan]

I excluded a few points because the numbers seemed too high, if I included those, I recall the CPVmax ~ 177 rpm's. I am uncertain if that fits in better with the vertical jump.

Not really, as neither the rider whose data are shown here:

http://home.earthlink.net/~acoggan/misc/id6.html

or yours truly:

http://home.earthlink.net/~acoggan/setraining/id3.html

have much in the way of neuromuscular power. Yet, the slope of your AEPF-CPV relationship is actually greater (more negative) than ours, when you'd expect exactly the opposite.

Back to the original question, this analysis was aimed at determining what was more of a limiting factor....pedal force or pedal velocity. Can I say one way or another at this point, which is more of a constraint for me?

Pedal velocity, clearly, but I don't believe the numbers are correct. Indeed, if I use the slope and intercept you indicate I calculate a maximal power of only 780 W, or less than half of what you can generate.

 

[Lucy_Aspenwind]

Pedal velocity, clearly, but I don't believe the numbers are correct. Indeed, if I use the slope and intercept you indicate I calculate a maximal power of only 780 W, or less than half of what you can generate.

Ughhh, I'm having rather chilling flashbacks to my graduate school statistics course. Even as I write this, I shudder....

Seeing as my statistical abilities are no better than they were before. Moreover, last time I checked my MMP chart, I could do 780w for almost 30 seconds.

I think I'll work on leg speed since even my haphazard analysis suggests that, not to mention I noticed it today on several sprints where my legs felt sluggish. Hence, new thread on improvements for leg speed

Thanks for being a good sport and trying to lead the mathematically challenged

 

[n crowley]

Quote:
Originally Posted by n crowley
What is the most important limiting factor to a fully trained TT or pursuit rider's pedal power application ?


Don't know if I understand the question? Road TT or a track TT (completely different ride demands). For a pursuit it's a matter of exceptional power at VO2 Max and/or a high AWC. If you're lucky you have both. I'm not aware that either of these can be influenced by weight training. They certainly can be improved by riding your bike. Easiest method would be to pick your parents to start with....

I am referring to only the muscles used in generating and applying the pedal power, those that could be used in resistance or weight training. The answer is the absence of a direct and firmly fixed resistance base which could instantly counteract and reinforce all increased force used in applying crank torque when using the higher gears especially in that initial thrust of power at the start of the powerstroke while using the perfected unweighting technique.