I think the H2H diff calculation needs to be adjusted:

1) Let's say Djokovic wasn't in the picture and we're just looking at Nadal/Federer. Because the H2H is 24-16 for Nadal, that's +8 for Nadal. This is bad enough, but then the formula goes a step further to add -8 to Federer, so an initial gap of 8 turns into 16: double-counting the gap

2) Differential also doesn't do a good job of reflecting the player's relative strength. As it stands now, Djokovic has +7 and Federer -12. This the equivalent of applying a weighting to 0 for Federer and to 19 for Djokovic- a hugely exaggerated spread

To get a truer picture of the relative difference between the three and avoid double-counting, then instead of [sum of matches won] - [sum of matches lost], it should be [sum of matches won] / [sum of matches played].

Hey, thanks for responding.

I don't like this calc either. Not sure what to do with it exactly - the weighting put in there is a "placeholder" more than anything.

I believe that H2H is an important measure between these 3 though - this is an individual sport and these 3 played about 150 matches between each other, sharing more than a decade of playing careers.

So, in my mind - the H2H HAS TO play a somewhat significant role.

Again, I don't like the calc either - and was looking for some well-argumented suggestion on how to do it.

So, I agree with you, and if you can suggest the way we should deal with it - and provide some reasoning, It'll be really appreciated.