I made a silly mistake in being astonished at the few negative splits in
the NYM field. Seems this is not unusual at all.
I just loaded the whole of the 2002 La Rochelle marathon results into
OpenOffice. The conditions were in the high 60s (end of November) by the
time I arrived and probably got hotter afterwards, race is flat and is
two identical loops in town (at seaside, no skyscrapers).
For 4770 finishers who had HM splits recorded :
winning time 2h17 (1'17" neg split)
last arrival 5h17
median 3h44 (av 3h45)
Under 2h30 21 (cumulative)
2h45 73
3h00 335
3h15 803
3h30 1599
4h00 3317
4h30 4327
Exactly even splits 1 (we can thus conclude that even splits is
a very poor race strategy, teehee :-D
True neg splits 414
Less than 1 mins positive 513 (includes neg splits)
Less than 2 mins positive 702
Less than 3 mins positive 908
Less than 4 mins positive 1113
av split 12:10
median split 9:44
In the first 300, 17 negative splits (cf NYM=6), some big positives too
(up to 24'😉. The absolute best pacing was done by the runner at the
201st place, in 2:55:09 official, 2:55:00 chip <bfg>
Of course none of this proves anything but it kept me occupied :-)