Pause test follow-up: first gap differences against recorded running times (after the result)

If a start's first gap difference marks its first pause, it ran at (pause + 300) / 230 time units per second in its first piece.

start  first gap diff  implied 1st-piece speed  implied total  recorded total (average speed)
    1           1,185                6.4/s          979 s  with starts 1-6
    2           2,065               10.3/s          614 s  with starts 1-6
    3           1,680                8.6/s          734 s  with starts 1-6
    4           2,300               11.3/s          558 s  with starts 1-6
    5           2,440               11.9/s          530 s  with starts 1-6
    6           2,625               12.7/s          496 s  with starts 1-6
    7             320                2.7/s        2,356 s  1,289 s (4.9/s)
    8             625                4.0/s        1,575 s  1,156 s (5.5/s)
    9           1,440                7.5/s          835 s  965 s (6.5/s)
   10           1,020                5.7/s        1,102 s  900 s (7.0/s)
   11           1,765                9.0/s          703 s  739 s (8.5/s)
   12           1,985                9.9/s          636 s  667 s (9.4/s)

Starts 1 to 6 (six tasks; their file records only the sum): at the implied speeds they would have run 3,911 s in all; their file records 3,834 s (average 9.9 units/s; implied 10.2).
Starts 7 to 12: the slower a start ran overall, the earlier it parted (rank correlation of first gap difference with average speed +0.94). Starts 7 and 8 would have needed a first piece slower than their own average (2.7 and 4.0 against 4.9 and 5.5 units/s).

A consistency check worked out after the result, not a test: the pause points were not recorded.