IMO medians are always preferred for 'pars', too many incorrect timings to skew things otherwise. What I did when I used time was to think in terms of par fractions for any final time at the track/dist/surface combination. This makes track surface variants completely irrelevant for assessing pace because the 'par' sectionals (internal fractions) come from the final time. The relevant variant for the sectionals is wind, not surface speed.
Also, IMO, the order of execution is, wind variant to adjust the sectionals first, followed by a new but faster final time that accounts for inefficient energy distribution as indicated by those sectionals. Last but not least you can then bring the 'daily' track variant into the calculation.
|