I've just read something in the parkrun newsletter about weight loss vs times. He reckoned he knocked 7secs off his 5k times for every pound lost, and that this is the same figure quoted by [some famous writing runner].
I read somewhere else that energy required to run is roughly linear with body weight. So if you change weight without gaining/losing any power/fitness, then a 14st runner taking 25min to run 5k, would hope to gain ....
yup, about 7secs per pound lost!
(Someone check my maths, but it's probably right). It can't really be a perfect linear relationship, but it's nice when these rough rules-of-thumb turn out to be pretty close to reality.