Federer served 50% in the 3rd set in 2007 so he definitely wasn't excellent in the 3rd. If anything, he was excellent in the 2nd set when he served 70%, had 18 winners to 9 unforced errors, but still almost lost the set. This shows how good 2007 Djokovic became in a span of a couple of months and how good of a player he would become. A more mature version of Djokovic would have given Federer hell on a day like that.
Statistically, the 2008 final is the worst final that Federer won but Murray was such trash in that match that it didn't even matter. Federer served 48% in that 3rd set in 2008 and Murray lost the set 6-2.
The serve percentage isn’t super great but Fed significantly improved his placement of those serves that did go in. Started in the second half of the second set and continued throughout the third. It was a trade-off that ultimately worked as Djokovic wasn’t able to return them as effectively. Part of that was Djokovic employing a very sus second serve return strategy in the third though, I will concede. But the point is that Fed was much better with his placement which I think mitigates the difference in %.
Fed was also sharper off the baseline in set 3. He was rough in the first few games of the second set (though not as bad as the first), which is the primary reason for Djokovic’s early lead, but the third set was largely free of those errors.
Not sure where your W-UFE figures come from. TA has Fed’s numbers equal in set 2 and slightly winner-favored in set 3. That was a relatively error-strewn match from both players, I’d be surprised if any of the sets had a 2-1 winner-error ratio.
Disagree on 2008. Yeah Murray was terrible of course but Fed was much more aggressive in that match. The 2007 final was probably the most passive Fed had played in a Slam final up to that point (well, besides maybe RG 2006 but that’s clay), and his FH was weaker than usual even if it was a good BH and return day.
Where is this statistical inferiority in the 2008 final coming from? He posted a strictly better winner-UFE differential going by TA (in 2007 it was negative which you rarely saw for Fed in the late stages of a HC Slam in those days) and an only marginally worse first serve %. The first set was much better than anything in the 2007 final, and I’d credit Fed’s forehand with that.
As a side note: I think people overrate how well Djokovic was playing in the 2007 match. His forehand was money, but he was too inconsistent, chokey, and his return needed work. He played better in both the 2008 and 2009 matches with Fed imo, even though they weren’t as close. Fed pretty much had as close to an off day as he could get in that final, though he cleaned up his play as the match progressed.