1

Average pace calculation error (Read 625 times)

Trent


Good Bad & The Monkey

    Looking at my reports, for the last 52 weeks, there is a summary of the week 6/30/2008. The average pace for this week is not correct. Does RA correctly weight the average pace for runs recorded as m, mile and km? Also, is it correctly dealing with runs that have no time value recorded (i.e., counting these as missing values rather than as 0)? Thanks
      If you gave no duration for a run, that distance will be added, but pace will be lower than it should be overall.

      Ricky

      —our ability to perform up to our physiological potential in a race is determined by whether or not we truly psychologically believe that what we are attempting is realistic. Anton Krupicka

      Trent


      Good Bad & The Monkey

        If you gave no duration for a run, that distance will be added, but pace will be lower than it should be overall.
        The durations for these should not be calculated in the average pace since it is not known. Right now, the calculation assumes that I completed the run in 0:00:00. I just as well could have finished it in 0:10:13 or 5:32:32.
          The durations for these should not be calculated in the average pace since it is not known. Right now, the calculation assumes that I completed the run in 0:00:00. I just as well could have finished it in 0:10:13 or 5:32:32.
          Why not? The distance is known, but the pace is zero. Why would you not put in a time? So you log (?) 3 runs for the week. 30 miles total in a total time of 4:00:00 . That's 8:00 per mile. Then you add a fourth run of 10 miles, but do not add a time. Therefore you have told the computer, "I have done 40 miles in 4 hours, What's my average pace?" I think it would compute it as 40 miles in 4:00:00 @ 6:00. You say the it should not be calculated in the average pace since it is not known how long it took you to run it. Duh! It assumes zero. When I have a run with an unknown duration, I allot more time than usual (say 8:30 pace) and give the computer the data it needs to calculate a pace, even if it is slower than I actually ran it.

          Ricky

          —our ability to perform up to our physiological potential in a race is determined by whether or not we truly psychologically believe that what we are attempting is realistic. Anton Krupicka

          Trent


          Good Bad & The Monkey

            The pace for the 4th run is unknown. I did not run 40 miles in 4:00:00 @ 6:00. I ran 30 miles in 4:000:00 and the other 10 at some additional amount of time. Those other 10 miles were not run in 0:00:00, they were run in an unknown amount of time and it is incorrect to assume 0. In basic statistics, this is a missing value and it is not to be used used, and if you are to assume a value for it you are meant to interpolate based on the other values that you do have. You say the computer assumes. That is correct, the computer assumes. It should not.
              You say the computer assumes. That is correct, the computer assumes. It should not.
              Then do not add distance to your total miles without a duration. It's gonna calculate total distance by total time running. Duh! It's a cumulative calculation is it not? mta: You answered not why one would log a distance covered but give no duration. I say that's cheating, yourself.

              Ricky

              —our ability to perform up to our physiological potential in a race is determined by whether or not we truly psychologically believe that what we are attempting is realistic. Anton Krupicka

              Trent


              Good Bad & The Monkey

                So pRed, you are saying that I cannot log my miles unless I have a watch on? I am saying that the computer should not assume. It is obvious that it is doing what you are describing. I am simply pointing out that it should not. Missing values should either be ignored or interpolated; they cannot be assumed to be zero.
                eric :)


                  Trent, While it would be ideal to come up with an exact value, it is not feasible with the current server resources. The pace is calculated as total distance / total duration. As you pointed out, there are different ways to deal with this missing data. No matter what I choose, there will be someone complaining about it. When writing these algorithms, I have to consider all kinds of constraints and this is the best I can come up with.
                  Trent


                  Good Bad & The Monkey

                    Can you not just ignore the missing values, which is the most accepted way of doing this in statistics? In the current approach, you are calculating values that can be wildly inaccurate if just one or two runs are untimed.