Quinn8it

I am sure Carleton will have a lot of insight into my question- hopefully some others with power data experience will join in!

recently there has been some discussion here about power data for sprinters (lets say from a flying 200m- just to keep things simple) and talk of the need for hi rates of data sampling to analyze peaks in power, speed and cadence, no matter how brief they were.
I have been operating under the thought that the short duration max-wattage numbers were good for making your roadie buddies jaws drop at coffee after a road ride with a bunch sprint- but not really the end-all in sprint analysis.

last Sunday i rode a PR flying 200m, it was my first 200 with a power meter. I was pretty excited to check the file. when i looked at it my first instinct was to look at peak power.. it was good for me, but not the highest i have seen this year, or even in the last 6 weeks. next i isolated the 12" that ends at the end of the effort- so roughly the timed section of the effort... once again- nowhere near my best 12" effort- not even the best 12" wattage of the day (i match sprinted later)
in some ways this left me thinking- wow i could of gone faster! cadence was about where i like it, so maybe a bigger gear would be faster?

then i started thinking about the effort and went back and isolated the entire effort. A 200m on ADT is a little different from some tracks since you have to maintain a higher speed early so you don't slide off, but there was a clear point on the data where things got serious. isolating from there to the end of the effort was approximately 25" effort. At ADT that is likely from mid turn 1/2 after you see "2 to go".. about the last 11/4 laps.

all wattage numbers from 15" to 25" were PR numbers for the year...

my take on this effort is: my limiting factor was obviously my 25" power, since i did not hit maximum wattage in 1"/5"/10" metrics...
this leaves me to believe that improving 30" power will improve my 200m

obviously bringing up the shorter duration numbers will help as well...

anyone have a take on this logic?

can anyone educate me as to why the shorter duration numbers and the higher rate of sample to analyze these numbers is so popular among sprinters?

carleton

I'm going to ramble a lot...bear with me.

1) Congrats on the PR so early in the training year!

2) Peak wattage comes at the right intersection of Torque (how hard you push) and Cadence (how fast you are pushing hard). They are inversly related, meaning as one goes up, the other goes down, and vice-versa. Torque=high and Cadence=low at the beginning of let's say a Standing Lap effort. Torque=Low and Cadence=High at the end of the Standing Lap. The Perfect mix of torque and cadence comes about 8" in as you exit turn 2 at around 120RPM. That's the sweet spot. You probably feel a boost of speed at that point. This is often where you see Man1 gap off his teammates in a team sprint. Power.

3) You can hit that sweet spot easily using lower that race gears. If you want straight-up bragging numbers, do a standing 1/2 lap on like a 90 or 92" gear. But, if you tried that gear in comp, you'd spin out too quickly to be of any use. So, using race gears, your max power will actually not be that high. That's OK.

4) Some software will show you Torque. This is NICE. This shows you how hard you are pressing on the pedals independently of cadence (Remember, Power needs Torque and Cadence for its calculation). More on this later.

5) You won't see Max Power numbers in a Flying 200M effort ...but you WILL see max speed (the reason we are here, right?). The reason you won't see max power numbers during a F200M effort is because when you JUMP as you approach the start line, your legs are somewhat faded from the windup. Plus, you won't be able to apply as much torque as you might when you are 8" into a Standing Start effort. Remember, the Flying 200M is not a 10-12" event. It is a 1 minute event! It's almost like a Kilo chopped up and the segments are rearranged

....

carleton

Analyzing files...

In an effort to compare apples to apples, always select a 12" slice from your files. This is so that you can compare the averages of multiple efforts from multiple days or same day using different gears.

How do you know what 12" to grab? Well, assuming that you stop pushing on the pedals as you cross the finish line, count 12" backwards from where the power line in the graph takes a nose dive.

Here is a good example:



You can see is averages in the upper left of each segment (the multi color numbers). Also, don't get thrown off. SRM software does this thing where it will truncate the numbers on the left of the graph. So, "000" means 1,000W. "500" means 1,500W.

After you mark off the 12" of the timed part, mark off 30" before that, so you can see how much energy you spent getting up to speed.

Now start comparing the averages. In the file above, you can see that he averaged 138RPM for the effort. This is good. That's sort of right in the middle of two schools of thought on the subject.

Basically, the million dollar question for Sprinters is: What gear should I use for my flying 200M?

My answer has always been: The biggest one that you can spin an average 140RPM.

Old school guys (from back when 94" was a sprinters race gear) say that 150 should be the target to average. New school guys (modern big gear advocates) say 135-ish should be the target. Fact: Many of the qualifiers of the London Olympics clocked MAX RPMs between 130-140. That means that their averages were lower. But, they are not mortals. They rode gears of 104" or bigger.

carleton

Regarding Match Sprinting numbers vs F200M numbers:

Match Sprints are more like Standing Starts than they are like Flying 200s. Typically, you don't spend as much energy before the jump of a match sprint as you do before the jump of a F200M.

Adrenalin also plays a factor. In 2012, I clocked my season best max wattage at Masters Nationals in the first round of sprints during the jump. It was a number that wasn't matched in my F200M or Team Sprint during the same peak week.

Many coaches yell, "UP UP UP!!!!!" for a reason...it works I once had a teammate do that to me when I was just practicing rolling starts...boom, extra 100W out of nowhere. The body and mind are weird, man. Weird.

Impreza_aL

i know my 200 has improved from just riding more. i started riding a bike around this time last year and did a 12.6(94gi). I then hit a 12.1 (101gi) during our velodrome challenge. I then participated in the districts sprints and hit a 11.9 (96gi). No lifting, no specific sprint work just riding (looking at my data i did the volume of a track sprinter lol) and racing.

this year i plan to do specific workouts. i've already put in two huge blocks for base. probably more riding than i have done all year. i know my 200 is probably capped maybe i can knock out a few more tenths. i would probably have to do a lot more sprint specific work outs (lift weights ewww) if i ever want to see the 10s.

as for your question, it would of probably helped if you had a pm at the start to just look at the numbers. but i feel as you gain aerobic power it'll increase your 200. it isn't really 200m anyway since you have to ramp it up. there are also other factors to the 200. your line, wind, humidity, blah blah blah...

carleton

Key metrics to note from your 12" slice:

- Average speed: this is your flying 200 time: time = 200M/(avg velocity)
- Average cadence (key indicator. You want to define a sweet spot and run the biggest gear that keeps you in that sweet spot)
- Average power: this will go up as you get stronger and use bigger gears
- Max speed: Directly related to your time, and a GREAT indicator of what it will be. Use the max speed from 100M test jumps to learn what is working without killing yourself doing a full on 200M. On a test day, you can do 3 full-gas flying 200s before getting weak...or you can do 6 full-gas flying 100s and note the top speeds. More opportunity to fool around with windups, jump spots, and gears. This is no less precise than coaches timing F100M splits, doubling the time and adding 0.1" which is a tried and true way of evaluating how fast a guy is on the day.
- Max cadence: Too low and you are missing your "wheelhouse" of power. Too high and you are spinning out and not able to give lots of torque.
- Max power: Not really a key indicator here. This can vary due to gearing, jump location, and/or windup style (slow or hot)

queerpunk

People are susceptible to suggestion when fatigued. I tell people I'm racing against to "go!" when a good-looking move goes off, and fairly often they do. they think i'm saying it for the benefit of all of us but really i'm just doing it to get a free ride up to the front.

carleton

You are right...but for different reasons than you might think

First, load up the stopwatch app on your phone and start it. See how fast 1, 2, 5 tenths of a second can zip by. Fast.

The Flying 200M is a very nuanced event. I've seen guys shave off more than 0.8" between efforts on the same gear just by changing where on the track they rise out of the saddle to stand on the pedals.

And "The Line". Wow. That is so track specific.

Here's a story for you. Mike Baranoski brought Matt down to DLV for one of our big events and they were at the track the day before working out. Part of the workout was Matt doing Flying 100s and 200s. I was like, "This guy is riding the wrong line. Everybody at DLV knows that the right line is different than that..." So, I do my Flying 100 or 200 and come off the track and, being the super generous guy he is, Mike asked me why I rode the line that I did and I told him that that's how we were all advised to do it. He suggested that I try the line that Matt was using and explained why. Being open-minded, I did. BOOM -0.3" off of the previous effort. I repeated the same the next day in the event using the new line.

Impreza_aL

sprinter vs me. please don't laugh.

carleton

Haha, don't be discouraged! This guy used to be the fastest sprinter in the world:


Theo Bos

Also, I see that he (I think that is Farioletti) is checking his blood glucose. Either he's diabetic or simply monitoring his blood sugar to keep it from dropping due to the crazy exertion that sprinters do. This is why we can eat chocolate and sweets...to keep glucose levels high enough for gluconeogenesis (that's what I tell myself anyway, hahaha)

https://en.wikipedia.org/wiki/Gluconeogenesis

Impreza_aL

yes it's hard to compare all the 3 because i'm pretty sure i approached (where i kicked, rolling in speed, line...) them differently. i just don't practice it enough. i love doing these sprint events especially the keirin and i feel i'm better suited to doing them but mass start racing is what's offered more at my velodrome. so i'm trying to better myself at that. hopefully get my 2 on the track.

Impreza_aL

oh man theo bos! well he's 6'3 i'm 5'8 lol. i don't keep up with professional cycling but has he won anything on the road? i wonder how his transition from track to road has worked for him. he left because there wasn't much money in track? if he didn't leave would he still have been dominate?

yes that's farioletti. i didn't notice he was checking his glucose level. i thought he was playing with his phone charger. i'm always eating chocolate during race nights =)



here's my new hero in the track cycling world. brown pride!


sorry for taking this thread off topic.

carleton

Practice is the key to doing better flying 200s. Like, having an entire session dedicated to Flying 100s and Flying 200s every so often.

My buddy (ATRA CAT A; all-around type of racer) called me once after the monthly sprint night:

Buddy: "Man, I can't get my flying 200M under 13 seconds."
Me: "Have you practiced it since last month?"
Buddy: "No."
Me: "There you go."

Fast forward 1 month and I get a text:

Buddy: "Still over 13 seconds :("
Me: "Have you practiced it since last month?"
Buddy: "No."
Me: "There you go."

A month later...

Buddy: "12.7!"
Me: "Have you practiced it since last month?"
Buddy: "Yes"
Me: "There you go."

carleton

Hahaha, to keep it on track, Njsane's top cadence during his Olympic Flying 200M was 140RPM according to a photo of his SRM head from Turn 1 during the event. (max speed/cadence happens in turn 1 on a 250M track).

Quinn8it

Lots of responses... But
I'm not sure you guys understood the question.

I'm not looking for advice on how to run a 200 (I'm actually pretty good at them)
And I feel that I know what I'm looking for in a power file- so I'm not really looking for power file 101

My take on power file data for a 200 is that max watts and max speed are important metrics- but the real determining factor for a 200 is something closer to 30" power.. I don't really see how super high sample rates of power data change the reading of a 30" chunk of data...

I get that it's more data- but i suspect it's just more of the same data

So my original question again:

carleton

I've explained a lot. Now it's time for you to ask SRM and Stages why they offer it

Quinn8it

Carleton,
i appreciate your lengthy responses, but i don't think you touched on data sample rate in any of them. I understand that faster rate of sample is State Of The Art and that manufactures will always be looking to advance their products. And i don't doubt that faster rate of sample is better.

im asking you, or anyone else on this forum, how the faster rate of sample changes your data and how these changes benefit your analysis and training?
the price difference in the Garmin 500 with 1sec sample and the SRM with (i forget how much higher) much higher rate of sampling is about 400%-
im just wondering what a 40-something masters hack like myself will see for my money... this is an honest quest for education.

Velocirapture

The faster rate of sampling simply gives you more data, and more data allows more precise analysis of the data, and a better understanding. if you are not interested in more precision analysis, and averages/ max number etc are good enough for you, then faster sampling wont benefit you. The converse also applies

slindell

The biggest difference for a track bike is that the Garmin does not allow you to turn off auto calibration. Every now and then you notice that watts are a lot harder to get than they were a minute ago and realize that the head unit auto calibrated when things were not really at 0. On the road it works ok with powertap and quarq, but SRM seems to send calibration when still under light load - not sure if it is a low cadence thing.
As far as sampling rates with ANT+ my understanding is that the power meter is sending its latest values at some interval. The head unit is looking for the last packet at some interval. If the power meter is sending more often than the head unit some packets will be ignored. If the head unit is sampling faster than the power meter some will be duplicated. So it is somewhat of a lossy protocol with the expectation that the values are not that different and it all works out in the averages. As the intervals get smaller the samples in the average go down so the reliability of the numbers goes down too - kind of like looking at GPS speed data - get too short and the values jump all around but go over a longer time and it all works out.
The wired units can sync the send/receive to run a no-loss protocol if they wanted to easier than acking wireless signals which would require the sender to also listen.
If you really wanted fine grained data you would need multiple wheel magnets and cadence sensors to get a more frequent signal on the sensors. This is where the Stages and Garmin units might have an advantage with the motion sensors being able to determine intermediate crank locations. I think the track and high end SRMs used to have (still have?) multiple cadence switches to get half rotation values.
With a base SRM with once per crank rotation signal you are not gaining much power data if your cadence is low. From 60-120rpm you are dropping some rotations, over 120 you are missing half with 1/sec recording. So you might miss the first pedal stroke or 2 of a jump.
Powertap does a time based accumulation and sends once per second (wireless - 1.25 sec wired) so in theory you don't miss things with a 1/sec receive but clocks drift and signals are missed so a 2/sec receive would usually show 2 duplicate values but can avoid missing the stray value.
This is why if Stages is updating 64 times a second it is important to know what interval they are averaging over in the packets.

Quinn8it

Slindell- thank you for that reply! there is some really interesting stuff in there.

this is basically what i was thinking:
any idea where the tipping point is? at what interval size is 1" sampling sufficient?
what would be the symptom of missed data in a short interval that was collected at 1"? im assuming that for most applications we are talking about showing slightly lower peak power data?

thanks again for your response.

slindell

If you miss a reading you might miss a peak pedal stroke or two. but odds are you catch in the next time you jump and most of the time you will have a few revolutions at close to peak power. It is getting to the fine grain 1-3 second measurements when the 10-20 second average is what matters. For fun going the other way SRMs have a min cadence of 30 RPM so there is 2 seconds before finding out you stopped pedaling. Try going hard then coast suddenly and there will be 3 seconds of the same reading.
Calibrate the meter and use the same type and it all works out close enough until you are at the point where others are doing all the analysis for you.

carleton

Thanks Velocirapture and Slindell for expressing what I had a hard time expressing.

Quinn, I think you are asking, "Where is the point of diminishing returns in the increased frequency of the samples?"

I think that the value of the data is "good enough" at 2/second (every .5 seconds) and great at anything faster than that. In my experience, every one second is not good enough for accurately measuring sprint efforts (Standing 1/4, 1/2, full laps, Flying 200M, etc...). Also, when comparing 2 files of the same effort (let's say a Kilo), with higher resolution, you can more easily line up the files for comparison.

Back when I ordered my current Track SRM (early 2011), Wired Track SRMs came with 2 Reed switches. Wireless Track SRMs only come with 1 Reed switch. You can have two Reed switches custom installed into just about any SRM crank.

The Reed switch is the switch inside the crank that responds to a magnet on the bicycle frame. Everytime the Reed switch passes the magnet, the switch closes and data is transmitted.

https://en.wikipedia.org/wiki/Reed_switch

So, Quinn, I think you'll see the benefits of the higher resolution when you get deeper into analyzing the files...you get to ENHANCE.

carleton

Quinn, what software do you use to analyze your files?

Also, I should have mentioned this earlier. Turn off any and all data smoothing. Smoothing is useful for analyzing a file from a 2-3 hour road ride. Not useful for analyzing short spurts of data as with track files (sprint or endurance).

Quinn8it

Not exactly. I'm more asking at what length interval is 2x per second not necessary?

carleton

Ah.

OK. Let's assume that we are going to use the software to section off splits. No one takes a file from a F200M, Standing Lap, 500M, 1K, 2K, 3K, or 4K and just looks at the average of the entire event.

Well, in general, the longer the event, the longer the interval can be to get a statistically significant number of samples when you start sectioning off the splits.

In my opinion, I think that sprint events (1K and shorter) should be recorded at least every 0.2 to 0.5 seconds. I think that pursuit events (2-4K) should be recorded at least every 0.5 seconds. I think that mass start events (points, scratch, madision, etc...) should be recordedat least every 0.5-1 second. Anything longer than that (The Hour) should be recorded at least every 1 second. That's just my opinion.

Storage is cheap. Batteries last longer. So, why not store all the data we can?

BTW, it should be noted that data file analysis is a thing that teams are taking more and more seriously. There is/was a guy on the British team's staff whose job it was to do just that, analyze all of the files.

Here's a great example that may illustrate why a faster sampling rate gives more insight:

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: