In case you're not bored enough on your long runs to figure this kind of thing out in your head, here's a JavaScript calculator that will figure out pacing splits and training paces for different distances and performances. For the training paces, don't take any of it too seriously. The relative levels are some which I think make some sense, based on my own running and reading. If you're Bob Kennedy or racing him, these are probably wrong for you.

Last Modified 27 November 2002
Robert Grumbine bobg@radix.net

This form is for paces. Enter your time of interest for any of the distances to get the even-paced time for other distances. If I'm missing a distance of interest, please write me.
Distance Even Pacing Estimated best
400m
800m
1000m
1500m
Mile
3000m
2 Mile
5 k
10 k
15 k
10 mile
H. Mara
Mara
50 k
10k PR5k PRMile PR800m PR
LSD/Recovery Long runs pace for the benefits of going long and recovery runs which give you a chance to log some beneficial miles, while permitting you to run hard the next day.
Brisk maybe 30 seconds/mile faster. Not a good run really, too quick for easy recovery, too slow for working on strength/speed.
Tempo 10k + 15s/mile. 4-7 k, get used to running with some speed, good for some strengthening (particularly if done on hilly ground). Also to get used to running efficiently with some speed.
LT 10k pace, 2000-2400m intervals get accustomed to running at/near lactic threshold while maintaining good form.
LT+ 15 seconds/mile faster than 10k, 1600-2400m intervals work on raising the lactic limit.
VO2 5k race pace 800-1600m intervals, get used to being at the VO2 limit
VO2+ 1600m race pace: 400m intervals. Try to raise VO2 limit.
Lacticland 800m race pace, 200m intervals -- get used to running very fast with large amounts of lactic acid pooling in the system.
Note that by some methods of comparing times, the 5k time should be about 15 seconds/mile faster than the 10k, which would make the LT+ and VO2 workout paces the same. I included them separately on the grounds that 1) the predictor may not work very accurately for all runners, and 2) many of us wind up with one of the times being slower than the other even if the predictor were right.