Correlation isn't causation, but when it corroborates almost every subjective experience as well as the physiology (full use of the creatine-phosphate system in first seconds of the race, coupled with high lactate tolerance developed in training leading to the advantage of positive split), statistical reference is valid.
ididk wrote:
Stateezee wrote:I challenge you to find as many top 100 800m times ever that are run with a negative split.
That is completely irrelevant. The most common way to run it doesn't mean it's the best way to run your fastest. This is basic statistics.