What is the conceptual, physiological basis for NP calculation?

Discussion in 'Power Training' started by doctorSpoc, Jan 10, 2014.

  1. acoggan

    acoggan Member

    Joined:
    Jul 4, 2003
    Messages:
    3,047
    Likes Received:
    9
    As the presentation attempts to illustrate, the smoothing was chosen based on the known kinetics of numerous physiological responses to changes in exercise intensity, whereas the weighting was chosen based on the known curvilinear relationship between exercise intensity and physiological (esp. metabolic) strain.
     


  2. doctorSpoc

    doctorSpoc New Member

    Joined:
    Nov 18, 2005
    Messages:
    1,488
    Likes Received:
    4
    I'm lazy.... looking for the coles notes version... thanks! I'll take a look...
     
  3. An old Guy

    An old Guy Member

    Joined:
    Feb 12, 2011
    Messages:
    1,380
    Likes Received:
    21
    I was simply repeating comments by TrainingPeaks.com about Bannister's model. I made no comments about scientific theory or models.

    ---

    As for using either Bannister's model or Coggan's model for developing training plans. TrainingPeaks.com says there is so much error in each that both are worthless for that purpose.

    And here we have acoggan saying that there is an NIH paper claiming support for his self promotion.

    ---

    You might review the math that produced that number. My math gets under 1.17. And acoggan says that has seen such numbers. I guess with that endorsement there is no need for me to supply ride data.

    ---

    I guess I can make a comment on scientific theories and models. If the theory and model disagree, one of them is wrong. Maybe both are wrong. It appears that Coggan's model disagrees with whatever the theory is.
     
  4. RapDaddyo

    RapDaddyo Active Member

    Joined:
    May 17, 2005
    Messages:
    5,088
    Likes Received:
    41
    I just ran your claim with an NP model I built a while back. After a full hour, NP settles in at 123%FTP. My earlier computation of 1.34 was based on a quick-and-dirty NP calculator designed for shorter segments. So, I guess you have set a new standard for NP busters. I'll let others decide whether they want to rely on my math or yours.
     
  5. quenya

    quenya New Member

    Joined:
    Jan 14, 2010
    Messages:
    411
    Likes Received:
    3
    Tough call, but I'll go with RDO's math.
     
  6. An old Guy

    An old Guy Member

    Joined:
    Feb 12, 2011
    Messages:
    1,380
    Likes Received:
    21
    (More likely your "quick-and-dirty NP calculator" was designed for longer intervals. At least the quick and dirty math ((0^4*30 + 160^4*30)/60)^.25 = 1.34) indicates so.)

    40/40sec intervals gives 113%
    50/50 second intervals gives 124%

    of course

    30min/30min intervals gives 134%

    Regardless of the actual number. Neither I nor those who acoggan has knowledge of really tried to hit big IF numbers. It was never a competition. Rather an observation of each of us that NP and the other things are not valid.

    ---

    I am sure if you want to hold a competition someone will show up.
     
  7. quenya

    quenya New Member

    Joined:
    Jan 14, 2010
    Messages:
    411
    Likes Received:
    3
    AOG, they already have those competitions. Bike racers call them 'crits'! Locally to myself the Santa Cruz Crit, Cats Hill, Nevada City Classic are excellent places to generate, so called, NP busters.
     
  8. dhk2

    dhk2 Active Member

    Joined:
    Aug 8, 2006
    Messages:
    2,214
    Likes Received:
    39
    Perhaps I don't understand how NP is defined and calculated. When I do the numbers for any on/off set of intervals of equal length and 160% intensity, the result will always be 134%.

    The result is independent of interval length for equal on/off times, because I'm always dividing (1.6)^4 by two. N'est pas?
     
  9. RapDaddyo

    RapDaddyo Active Member

    Joined:
    May 17, 2005
    Messages:
    5,088
    Likes Received:
    41
    That simplistic algorithm works for a two-segment ride. That's what I used initially. For a repeating segments ride, the easiest thing to do is to build an Excel model where you build NP for each period (second) based on the rolling 30sec AP as defined in the NP algorithm. It's pretty quick to build a spreadsheet for this for two repeating segments. Obviously, it gets more complicated to build a general purpose spreadsheet when you have more than two repeating segments. At that point, I abandon Excel and switch to code.
     
  10. An old Guy

    An old Guy Member

    Joined:
    Feb 12, 2011
    Messages:
    1,380
    Likes Received:
    21
    I always thought the idea was to be first across the line not to prove that NP is invalid. But you learn something everyday.

    ---

    I watched 3 crits on the internet last year.

    A fellow I knew was riding in the P/1 events. He is always a pleasure to watch. He sat off the back and relaxed until the last couple laps and then moved up to get a better position to watch the finish. After the event he would go out and do his scheduled training ride.
     
  11. dhk2

    dhk2 Active Member

    Joined:
    Aug 8, 2006
    Messages:
    2,214
    Likes Received:
    39
    RDO, wouldn't it also apply to a ride of any number on on/off segments, as long as the power is flat and consistent during the on segments, and zero during the off? Realize this doesn't apply to the real-world application to PM files from road riding. But I'm thinking about my little intervals on the gym trainer when I focus on holding the goal output reading steady and consistent for each interval.

    IAW, if I do on/off intervals of fixed equal durations, say 1 min on/1 off, at a constant 160% FTP, won't the NP factor be 1.34, regardless of the length of each interval, or the number of intervals? If I extend my rest intervals to 2 mins (1 min on/2 off), won't the factor always be 1.21?

    I'm not sampling output of course, just using my target power number as a constant on the Cybex display (despite the fact it jumps around +/- roughly 5-10 watts despite my best efforts to keep it constant. Again, realize the real-world variation riding on the road with a PM would generate even more variation.
     
  12. RapDaddyo

    RapDaddyo Active Member

    Joined:
    May 17, 2005
    Messages:
    5,088
    Likes Received:
    41
    The problem is with the 30s rolling average computation of NP.
     
  13. quenya

    quenya New Member

    Joined:
    Jan 14, 2010
    Messages:
    411
    Likes Received:
    3
    Interesting. Doesn't sound like the idea there was to be first across the line for everyone. For my part I am a good time trialist, and perhaps a decent road racer, for me crits are anaerobic workouts, fitness gauges, or races where I intend to support friends and teammates. Data collection is in fact often at least a secondary or tertiary objective.
     
  14. Alex Simmons

    Alex Simmons Member

    Joined:
    Mar 12, 2006
    Messages:
    2,471
    Likes Received:
    20
    No. A long enough session with intervals of same intensity but different durations of work and rest intervals would have different NPs, and hence Intensity Factors (IF).

    e.g.:

    30-seconds on / off @ 160% / 0% would have an IF of 1.07

    1-min on / off @ 160% / 0% would have an IF of 1.23

    2-min: IF of 1.29

    3-min: IF of 1.31

    5-min: IF of 1.33
     
  15. An old Guy

    An old Guy Member

    Joined:
    Feb 12, 2011
    Messages:
    1,380
    Likes Received:
    21
    When I show up for an event, I look around and see what is possible.

    At that event there were 100 Pros and 1's in teams of 1 or 2. Then there were 6 guys on a pro team. Tempo for 6 guys is 30% more power than anyone else can produce.

    I don't see a way to finish in the top 5. So just sit off the back. Stay out of trouble.
     
  16. dhk2

    dhk2 Active Member

    Joined:
    Aug 8, 2006
    Messages:
    2,214
    Likes Received:
    39
    Thanks for replies RDO and Alex. To compute that NP factor, I was just using the steps on Dr Coggan's slide #17 from the link above, which appeared to be independent of the length of the on/off time intervals. IE, units of time could be seconds, minutes, even hours (in theory) as long as the ratio of on/off was consistent. Was thinking of a simple case of 1-on/1-off intervals, as represented ideally by a square wave with a 50% duty cycle, where the on power is always a steady 160% of FTP.

    Appears I need to do more study of how NP and IF are defined and calculated. Didn't intend to waste your time with novice questions here.
     
Loading...
Loading...