1000m splits were fine. What I really need is a description of the source of them. I'm asking because there's academic research out there that concludes you should pace a 5K on--I kid you not--a 15 sec positive split from mile 1 to miles 2 and 3. I need to point out that whatever might be happening in laboratory treadmill tests, that is now how WRs are run. (I have my guesses about what's going on in the lab tests, but I need the real world data to point out that it does not produce peak performances at the elite level.)
Track results are good for this, even though I'm talking to road racers, because there aren't any hills or long upwind/downwind stretches, etc.
I found another thread with Bekele's 10K splits (by km) , which is useful.
Letsrun is a great source for such information, but if I can track it back a little closer to an official data source that would be ideal.