Records start at the gun

Back in September, in my post The Intermediate PR Theorem, I discussed whether, if my 2-mile split in a 5k was 11:09, I can multiply this time by 1.864/2 to convert it to a 3k time of 10:24. So if my 3k PR had been 10:30, could I now claim a new 3k PR of 10:24? A legitimate question.

In discussing this with one of my professors, I discovered that the answer is no. Consider this:

The current world record for the 100 meters is 9.58 seconds, which Usain Bolt ran in Berlin. Bolt has only run 9.58 once in the 100-meter dash. But in fact, lots of people have run 9.58 for 100 meters before. How? In the last 100 meters of the 200-meter dash!

When you compare 100m and 200m times, it's interesting to note that individuals' 200m times are usually faster -- less than twice their 100m times. For instance, Michael Johnson's 100m PR was 10.09 and his 200m PR was 19.32. In fact, he went through the first 100m of his 200m in 10.12, and then proceeded to run 9.20 for the second 200m. Could he therefore say that his 100m PR was 9.20? No, because he had what's called a "flying start" -- he didn't have to accelerate from motionlessness in the blocks; he started at full speed.

This effect is much smaller in longer events; 200m times are much faster than 400m times. However, the fact remains that in order to be a record for a particular distance, the performance has to start at the gun. For instance, Haile Gebrselassie set the world record for 20k en route to his 1-hour world record, and this record was the first 20k of the race. Even if he ran faster between 1k and 21k than between the start and 20k, it wouldn't count as a record because he had a flying start.

So if I want to run a PR for 1500m in a mile, it had better be the first 1500m!