Bike Forums

Bike Forums (http://www.bikeforums.net/forum.php)
-   Track Cycling: Velodrome Racing and Training Area (http://www.bikeforums.net/track-cycling-velodrome-racing-training-area/)
-   -   Power File Analysis for Sprinters?? (http://www.bikeforums.net/track-cycling-velodrome-racing-training-area/923278-power-file-analysis-sprinters.html)

Quinn8it 11-25-13 06:27 PM

ok. not totally sure i got the answer i was looking for... but thanks.

i think that this quote does a nice job of illustrating what I'm getting at:
Quote:

Originally Posted by carleton (Post 16277874)
Imagine that you are trying to analyze your athlete's Standing Start. You have an old digital camera that can only take continuous photo every 1 second. Now imagine getting one that took frames every 0.5 seconds. THEN you got one that took 30 frames per second (video quality). Which one would you like to use?

Then what if you had a camera that could do this...filming arguably the best Man 1 in the game:

I understand that faster sampling is state of the art, and i understand what is basically the overwhelming argument in this thread for faster sample rate: "More=More and More=Better"
I am also well aware that the best cycling development teams use the state of the art technology to analyze the fastest starter that ever lived-

my issue is with the seemingly universal consensus amongst the knowledgeable members here that anything short of the absolute best is a totally not valid training tool. I am sorry to say that regardless of whether you are sampling at 1x per second or 60x per second there is a real possibility that the resultant data is identical.
remember that i am using both SRM with .5" sampling and a stages/Garmin with 1" sampling- i am not seeing any glaring difference in power data.

anyone who did not already have an opinion on this subject would likely read through this thread and come the the conclusion that they needed a $3500 power meter and a $1000 head unit

I think that this is an unfortunate situation.

my personal methods of collecting power data and analyzing power data has led to results that i am very happy with- I can pedal a bike pretty damn fast! sometimes i wonder if people lose sight of that goal, and focus more on trying to appear to know the most about training methodology?

carleton 11-25-13 06:39 PM

Quote:

Originally Posted by Quinn8it (Post 16277960)
my personal methods of collecting power data and analyzing power data has led to results that i am very happy with- I can pedal a bike pretty damn fast! sometimes i wonder if people lose sight of that goal, and focus more on trying to appear to know the most about training methodology?

Man, I think you are combining how you feel about data collection and data collection. I don't think anyone is passing judgment on you and your choices or saying that you should buy something.

SRM has been the gold standard for years to which all other power meters are compared. They have a well-deserved good reputation. You can get SRMs used at a reasonable price AND SRM will service models that are over a decade old. I've been to their US HQ and during a tour it was explained to me that they still take old units and repair and/or upgrade them.

What is good enough for you and makes you happy is one thing. What is better in terms of data collection is another. Ask any scientist or data analyst if he would like X or 10X data, he would take 10X data in a heartbeat.

You posted that the Stages system can record 64/second. That blows SRM out of the water. Last I checked, the very expensive Scientific unit could only record 10/second. Stages has downsides, though.

I'm sorry that you aren't convinced. I'm not shooing you off, but maybe you should post in the famous Wattage forum which has Andy Coggan and similar scientists as regulars and get the answers you need: https://groups.google.com/forum/#!aboutgroup/wattage

Racer Ex 12-07-13 01:48 AM

Quote:

Originally Posted by Quinn8it (Post 16277960)
anyone who did not already have an opinion on this subject would likely read through this thread and come the the conclusion that they needed a $3500 power meter and a $1000 head unit

While I'm over here I'll comment on this.

What you need is a reliable, consistent unit (relative to prior readings) with a high enough sampling rate to allow you to see differentials.

Example: If you are doing standing starts at the pointy end of the stick where 1/10th of a second here or there might be the winning differential, then having a high sampling rate might let you choose between the pedal at 9:45 vs. 10:15 for the maximum peak output and launch.

For the short track stuff like a 200m or 500/1000m this matters a lot more than if you're doing 4k's.

The reality is most people aren't at the level of consistency where they can benefit from fine analysis like this; even guys who have age group records are all over the map when it comes to the shorter distance stuff.

bitingduck 12-09-13 05:20 AM

I'll add my $0.02 to the sampling rate as well:

Find the duration of "events" (in this case power variations, not the race duration) that are significant to the outcome of the race, and then pick a sampling rate that puts at least 2 data points, and preferably more like 5-10, inside your shortest event duration.

Easier said than done. Suppose the kilo becomes a huge international moneymaking event and every kid on the street wants to be the next Theo Bos and you want to sell them all power meters. You can afford significant development money in that case, so what you'd do is get a fast data acquisition (say 1 ksample/sec, which really isn't that fast) system and Theo, put him on a bike and have him ride a bunch of kilos. Then you'd select a sampling rate such that if you sampled any slower you'd start clipping the peak of the standing start (and thus be under measuring the total energy spent), and if you're nice you'd sample a little faster than that so people don't complain about your product.

A different way of saying this is that for a particular race you're measuring the total energy spent (and when) and how much error can you tolerate in that measurement? In a flying 200 m, you're only racing 10-12 s, so at 2 samples/sec you're only getting 20-24 data points during the timed part, and if you look all the way back to the final acceleration it's maybe 15-18 s, and 30-36 data points. But your power might be varying a lot over the course of 2-3 seconds before you hit the first tape, and you want really good data there. But you only have 4-6 points to fit, so there might be a peak that gets clipped and you could have the energy spent in those 2-3 seconds off by 10 or 15% if the samples are at exactly the worst times. So you need a faster sampling rate. In physical terms, if you're pedaling at 120 rpm and sampling at 2 samples/sec, you're getting one power sample per revolution, which probably isn't nearly enough-- given that you're doing two "power strokes" with your legs during that one revolution, you need at least twice the sampling rate, preferably 4-5 times. The power meter could still get the total power right by using an analog integrator for each point, but if you're a sprinter you probably want the details of the peaks.

Typically, the longer the event, the longer a sampling rate you can tolerate. Even though mass start races are won and lost on the accelerations, the peaks aren't as high or as sharp in a standing start and clipping one isn't going to bother a mass start racer because position in the pack matters more than a joule here or there. And a pursuit should be pretty steady power after the start, so you might want a higher rate at the start and then a lower rate for the rest. But it's probably cheaper to make a unit with a faster ADC and more memory (and changeable rate so you can sell it to randonneurs, too) than it is to try to make it sample faster at some particular point (e.g. the start).

edit: and FWIW, I've never even used a computer on the track, let alone a power meter. I had a power meter on the road briefly (I got an SRM cheap used) and it was mostly just entertaining. You can train very effectively with a cheap clock, and then evaluate the cost/benefit of anything beyond that.

bitingduck 12-09-13 05:35 AM

Quote:

Originally Posted by Racer Ex (Post 16308846)
Example: If you are doing standing starts at the pointy end of the stick where 1/10th of a second here or there might be the winning differential, then having a high sampling rate might let you choose between the pedal at 9:45 vs. 10:15 for the maximum peak output and launch.

You also have to be careful of bias that a not-quite-high-enough sampling rate can introduce-- how many samples are you getting per pedal stroke at the start? If it's only 3-4 it's not much better than 1 and might deceive you into thinking that you're producing more power at different times than you really are, depending on when the samples are taken in the stroke (which meters don't record).

slindell 12-09-13 02:41 PM

For high sampling rate you can look at something designed for the highest power short races that riders actually make a living at - http://g-cog.com/G-Cog.htm - BMX poser meter in the cog that will sample 250/sec and looks like it records at that rate. http://g-cog.com/Versus.htm - in the graph the start is showing the peaks for each pedal stroke.
A study says the numbers do not line up with Powertap or SRM which might show the differences in sampling/recording in sprints or some issue with the G-Cog. http://www.ncbi.nlm.nih.gov/pubmed/23254482

carleton 12-09-13 03:22 PM

Quote:

Originally Posted by slindell (Post 16315307)
For high sampling rate you can look at something designed for the highest power short races that riders actually make a living at - http://g-cog.com/G-Cog.htm - BMX poser meter in the cog that will sample 250/sec and looks like it records at that rate. http://g-cog.com/Versus.htm - in the graph the start is showing the peaks for each pedal stroke.
A study says the numbers do not line up with Powertap or SRM which might show the differences in sampling/recording in sprints or some issue with the G-Cog. http://www.ncbi.nlm.nih.gov/pubmed/23254482

I remember G Cog from a couple of years ago. I messaged the owner asking if he'd make a unit for track and he said that they weren't going in that direction. To be honest, there is more money in BMX than track cycling.

Anyway, 250 samples per second is awesome. This GCog chart showing the variances that occur in a 1 second are amazing. The red dots show only what a SRM records every 0.5s.

http://i.imgur.com/pqDQEzd.png

The GCog team rightfully calls out the guys who published that scientific report. I would have, too. Also, PowerTap is not considered a "gold standard" recording device ESPECIALLY for recording efforts like the 6-8s efforts in the study.

http://g-cog.com/research.htm

Racer Ex 12-09-13 04:02 PM

Quote:

Originally Posted by bitingduck (Post 16313777)
edit: and FWIW, I've never even used a computer on the track, let alone a power meter. I had a power meter on the road briefly (I got an SRM cheap used) and it was mostly just entertaining. You can train very effectively with a cheap clock, and then evaluate the cost/benefit of anything beyond that.

I've found it pretty helpful for pursuiting (2k and 4k) and the associated training. According to USA Cycling I'm too old to ride a kilo and I suck at the 500m 80% of the time I ride that event so I couldn't comment on that stuff. My 200 would probably work with a sun dial.

bitingduck 12-09-13 04:25 PM

Quote:

Originally Posted by Racer Ex (Post 16315557)
I've found it pretty helpful for pursuiting (2k and 4k) and the associated training. According to USA Cycling I'm too old to ride a kilo and I suck at the 500m 80% of the time I ride that event so I couldn't comment on that stuff. My 200 would probably work with a sun dial.

I'm morally opposed to anything that doesn't allow drafting. I've been known to occasionally compromise my morals by doing a pursuit at states, but only under duress. Hence no real need for a computer or meter. All that really matters is where I am with respect to everyone else every 250 after they ring the bell.

Racer Ex 12-09-13 04:29 PM

Quote:

Originally Posted by bitingduck (Post 16315631)
I'm morally opposed to anything that doesn't allow drafting. I've been known to occasionally compromise my morals by doing a pursuit at states, but only under duress. Hence no real need for a computer or meter. All that really matters is where I am with respect to everyone else every 250 after they ring the bell.

Team pursuit. Your new bestie.

bitingduck 12-09-13 04:39 PM

Quote:

Originally Posted by Racer Ex (Post 16315647)
Team pursuit. Your new bestie.

Yes, I like those. Much drafting, but I have to remember not to up the pace a teeny bit everytime I hit the front. Annoys the teammates.

Racer Ex 12-09-13 05:16 PM

The BMX powermeter is intriguing BTW. Didn't know that was out there.

carleton 12-09-13 06:35 PM

Quote:

Originally Posted by Racer Ex (Post 16315800)
The BMX powermeter is intriguing BTW. Didn't know that was out there.

It's good stuff. Very well thought out.

[video=youtube;RM73rjmuCVw]http://www.youtube.com/watch?v=RM73rjmuCVw[/video]

More info here: http://www.youtube.com/user/GCogPowerMeter/videos

carleton 12-09-13 06:40 PM

I think that sprinters can learn a lot from BMXers. There are A LOT of BMXers that rose to the top ranks of track cycling: Victoria Pendleton, Chris Hoy, Craig McLean, Jamie Staff, Shanaze Reade etc...

Quinn8it 12-09-13 09:12 PM

Quote:

Originally Posted by bitingduck (Post 16313777)
A different way of saying this is that for a particular race you're measuring the total energy spent (and when) and how much error can you tolerate in that measurement? In a flying 200 m, you're only racing 10-12 s, so at 2 samples/sec you're only getting 20-24 data points during the timed part, and if you look all the way back to the final acceleration it's maybe 15-18 s, and 30-36 data points. But your power might be varying a lot over the course of 2-3 seconds before you hit the first tape, and you want really good data there. But you only have 4-6 points to fit, so there might be a peak that gets clipped and you could have the energy spent in those 2-3 seconds off by 10 or 15% if the samples are at exactly the worst times. So you need a faster sampling rate. In physical terms, if you're pedaling at 120 rpm and sampling at 2 samples/sec, you're getting one power sample per revolution, which probably isn't nearly enough-- given that you're doing two "power strokes" with your legs during that one revolution, you need at least twice the sampling rate, preferably 4-5 times. The power meter could still get the total power right by using an analog integrator for each point, but if you're a sprinter you probably want the details of the peaks.

good stuff!
thanks for that Bitingduck! very well explained..


Quote:

Originally Posted by bitingduck (Post 16313777)
but if you're a sprinter you probably want the details of the peaks.

I think i must be the only sprinter out there that is not that concerned with the short duration stuff... in my eyes the 200m (our shortest event) is a 25+second effort.. i get faster when 25" power goes up.. the peaks don't really seam to matter so much..

Racer Ex 12-11-13 12:18 AM

Quote:

Originally Posted by carleton (Post 16316039)
I think that sprinters can learn a lot from BMXers. There are A LOT of BMXers that rose to the top ranks of track cycling: Victoria Pendleton, Chris Hoy, Craig McLean, Jamie Staff, Shanaze Reade etc...

Cut to 2:33. This was pretty much exactly me in '71 :)

[video=youtube;YHIHj8VwMV8]http://www.youtube.com/watch?feature=player_detailpage&v=YHIHj8VwMV8[/video]

carleton 12-11-13 01:07 AM

Quote:

Originally Posted by Racer Ex (Post 16320060)
Cut to 2:33. This was pretty much exactly me in '71 :)

[video=youtube;YHIHj8VwMV8]http://www.youtube.com/watch?feature=player_detailpage&v=YHIHj8VwMV8[/video]

Hahaha! Nice!

jmikami 03-25-14 03:38 PM

Quote:

Originally Posted by Quinn8it (Post 16316408)
I think i must be the only sprinter out there that is not that concerned with the short duration stuff... in my eyes the 200m (our shortest event) is a 25+second effort.. i get faster when 25" power goes up.. the peaks don't really seam to matter so much..

I hate to review old threads, but I am going to start using a stages power meter for my track training ... as soon as the sun comes out for an extended visit and was curious if you have made more progress with your ability to analyze your rides. I currently use a powertap on the road and love being able to toss my intervals into graphs and look at how my power improves or not with training and create new targets. I am looking to do the same on the track this year and would like to avoid re-inventing the wheel. However my first go around with testing the stages power meter has shown a few gaps.

I am currently sending my files to golden cheetah and then back to my powertap software to do the analysis or to excel and creating my own files, but that is a pain. I am finding sending my files directly to my powertap software creates issues and while I only have a couple of workouts to review, I will soon be training on the track and hoping to avoid wasting time.

As for what coaches say ... blah, they mostly trained without power and sprint coaches seem to be the last people on the planet to switch over to the new world. The new world is all about training with power and the sprint coaches will eventually care what your 5sec, 12sec, 20sec and 30sec power is. They likely avoided the issue because of the heart rate training which was totally useless for sprinting, and are just technology adverse. I doubt they will care much about your max power because that is only good for drinking at the bar after the event, even a standing start is more about what you do after the first 5 pedal strokes. I personally think that adjusting my training to max out my 20 and 30 second power in particular will be very helpful in my sprints. Although I plan to focus on races one kilo and longer, I enjoy any day that I can ruin a true sprinter.

So, while I don't have the data files or experience to help on this yet, I hope to contribute to the stages power meter/track analysis soon.

Quinn8it 03-25-14 04:34 PM

i will start by saying that i am very happy with my Stages.
not only did i put one on my Track bike, but i added one on my Road Bike and also put a Compatible crank on my Single Speed that i do Crits on, so i can move my Rival Arm back and forth.. i know there are people out there with issues with them- most have never even seen one, much less tested them.. ill just say that i am totally happy with mine..


as for track training with power:
since most training efforts for most sprinters are maximal- i think all the data from a power meter can be a little useless in terms of dictating training- they are more for recording progress. like- "wow- my power/speed/RPM's fade in the second half of a flying kilo"… well Duh! you didn't need a $1000crank to tell you that.. and now that you know can you actually change it?

where the Power Meter has been invaluable for me is in calibrating track efforts with ergo efforts, so that i can make the best use of my time and train at home..
i can do a series of test efforts at the track- and replicate those efforts on my ergo with the same RPM and Power range for the time duration and get in a relevant workout..
Ive got a power file from a PR 200m last fall, basically a 25" effort. so i can watch my best efforts and get an idea of what time ill be shooting for once i get back on the track- this can also help me with gear choice.

I also like comparing blocks. So up until i got sick this week i was having a good little run numbers wise. I was able to pull up my current 6-week power curve and a power curve from a 6-week block that ended just before my final taper last season, and determine that i am actually ahead in terms of output.. despite Zero track time

i dont understand your situation with analyzation.. why not do everything in GC?
i like to use GC for pulling the files off my device and storing the files, then i use the upload feature to send the files to Strava, which has a longer history of power files and allows me to see data from any computer...

jmikami 03-25-14 05:20 PM

Quote:

Originally Posted by Quinn8it (Post 16610907)
i dont understand your situation with analyzation.. why not do everything in GC?
i like to use GC for pulling the files off my device and storing the files, then i use the upload feature to send the files to Strava, which has a longer history of power files and allows me to see data from any computer...

I use a power meter to figure out if it is a good day for efforts, if I can't get it up, I just take a day off and swap easy and hard days around until I can get good power output. I figure there is no reason to load up on a 80% wattage day, when tomorrow or the next day I will be able to load them up at 100% for that distance. I am all about what can I get in during my 10 to 12 hours on the bike this week and how many intervals or efforts can I handle right now. Wattage is my indicator if I am rested enough to handle another hard day, usually I combine that with HR for the longer efforts to ensure I am on track for a good interval day. For that I am currently using the iphone wahoo app, and it works ok, but I might need to customize the readouts some more. I focus on 30 second, 1 minute, 2 minute and 3 minute numbers as I am more of a pursuit rider, but still work on sprint and kilo type efforts as well.

After that I use it like you said to review the effort and to see if all my efforts were the same or if I was better at the start, middle or end, etc. Then I concentrate the next sessions on working both areas where I am weak, and continue with my strengths as well so that I can continue to max out my power and load up my legs. Where I am having the most issues is my post analysis with sending the files out via email from wahoo (where the issue likely lies) then loading them up on software on my pc. I might be doing something wrong as I have only done this twice so far, but PowerAgent from cycleops will not take the file that wahoo sends. I first have to load it to golden cheetah and then to poweragent. Also I get weird zero data that I have been deleting to not mess up my files further as the zero data has anything but zeros at times. Overall just odd, but as I said I am 100% a newbie with the stages meter and with golden cheetah not to mention the wahoo phone app. I have loved my powertap and I am happy with poweragent, so that is what I have been using.

Maybe I need to get use to golden cheetah and start using that more. I sent one set of data to strava, but I don't use that as much either, and I figured all of the agents should work with the same data so why not use what I know ... and that is poweragent. Overall I am excited with what the stages will bring to my training this year and hope that the additional time on the track leads me to better times at nats this year. Maybe I need to focus on the head unit and figure out something better than the wahoo app on my iphone.

Quinn8it 03-25-14 05:42 PM

Just as a heads up-

i recorded a single road ride off a stages meter simultaneously on a Garmin 500 (Ant+)and via the Strava app on my iPhone (Bluetooth)
the file recorded by the phone had crazy transmission gaps- basically useless.

also-
Play around with the PMC chart on the Home tab of Golden Cheetah. Interesting fitness/fatigue tracking. It will take a while uploading files before the readings are complete- but you may be able to isolate when you will have a good day before it happens..
Me I just train hard no matter how I feel ;)

carleton 03-25-14 05:57 PM

Quote:

Originally Posted by Quinn8it (Post 16610907)
as for track training with power:
since most training efforts for most sprinters are maximal- i think all the data from a power meter can be a little useless in terms of dictating training- they are more for recording progress. like- "wow- my power/speed/RPM's fade in the second half of a flying kilo"

+1

Sprinters use power data in a different way than roadies or enduros. As a sprinter, I used the power data to see what or how something happened, but as Quinn rightfully pointed out, every effort is maximal. So, power data just shows what my max was (or wasn't) that particular day on that particular gear for that particular effort.

Roadies/enduros often use it to make real-time decisions. Basically, "PUSH HARDER. You have room." or "BACK OFF before you blow."

A secondary benefit (that Quinn also mentions) is tracking fitness and freshness. If you have lots of data, you can see how your numbers stack up against previous efforts (either previous days or previous efforts in the same day). You can also track training volume.

To a sprinter, your speed and cadence data is more valuable than power...in my humble opinion (I'm a guy with power files going back to 2009).

Quinn8it 03-25-14 06:59 PM

Wait-Wait!
Did Carleton and i just agree on a bunch of stuff?

or did i take too much cold medicine?

jmikami 03-25-14 07:40 PM

Quote:

Originally Posted by carleton (Post 16611133)
To a sprinter, your speed and cadence data is more valuable than power...in my humble opinion (I'm a guy with power files going back to 2009).

If you are doing flying 200s and your power is off 10 to 20% would that not be an indicator that you are blown and to stop. I agree you could do this with cadence and speed, but power would be just as good. Once you train with power, I think it beats both speed and cadence as track and wind can affect speed/cadence, but not power. Power can also transition to different surfaces like training on the road for track, and using slight inclines or declines while training to purposefully change things up. I sprint uphill, downhill and on the flats, as well as low and high RPM. Power is the only constant with all of these, speed and cadence are worthless ... same with motor pacing for those who do it. I think that once sprinters really get the idea of power and how it can work they will never go back. But it seems to be a slow uptake.

Also another issue I see with using power on the track or with shorter efforts is that all the TSS/Fatigue/fitness numbers are endurance based. I am a numbers geek and am trying to come up with a way to show that a madison is by far the hardest workout anyone will ever do, IMO, and so the TSS for that, even at 45 minutes will blow doors on almost any 3 hour road ride, again IMO. Same with sprinters doing 100 and 200 meter efforts and then laying down and doing it again and again. The models I have seen are just too enduro focused to truly measure the impact of sprint/anaerobic efforts. I know that after the six day I am way more blown than any stage race I use to do, and those were often 5 and 6 hours in the saddle. Different tired for sure, but there is something about a good track work that just doesn't show up on my TSS right now.

I look forward to my year with my stages power meter and hope to see more people use them so that we can get better modeling data.


All times are GMT -6. The time now is 01:43 AM.