Depends on what you mean - the most epic match could be "felt like the final", even if the final is not at all bad. I would say an example of this would be USO 21. The most epic match could be argued to be Zverev-Djokovic SF, whereas the final was a step below that in terms of quality. On the other hand, the final still seemed like a contest.
I'd argue you need 3 things for a prior match to "feel" like the final.
- The two strongest players (or two of the 3-4 strongest) in the tournament need to meet
- The match those two players have is of at least decent quality and goes on at least 4 sets
- The final is between the winner of that match and someone with a significantly lower chance of winning, where the winner of the "de-facto final" wins.
These criteria would be met in circumstances like RG 2013, for instance - Nadal and Djokovic were in a class of their own, played an epic 5-setter, and Nadal went on to steamroll Ferrer (we can also assume Djokovic would have easily beaten him too).
In slams like RG 2021, this also works fairly well, though the 3rd point is a maybe. Tsitsipas had a significantly lower chance, but he was easily the 3rd best in the tournament and the leadup to it. Under these criteria, USO 2014 had no match that "felt" like the final too. Cilic would need to be involved in that match, and Cilic either had 3-setters against the favorites (Federer), or had tough matches against people that had no business ever winning (Simon).
My list:
USO -
1971 SF (Okker v. Smith)
, 90 QF (Lendl v. Sampras),
93 QF (Chang v. Sampras),
98 SF (Sampras v. Rafter),
04 QF (Federer v. Agassi),
08 SF (Djokovic v. Federer) (if any),
17 SF (Nadal v. Delpo).
honorable mentions (Failed on 3rd point) 75 SF (Orantes v. Vilas), 84/5 SF (Mac v. Connors/Wilander), 92 SF (Chang v. Edberg), 01 QF (Hewitt v. Roddick), 11 SF (Djokovic v. Federer).
Wimbledon -
68 QF (Laver v. Ralston),
73 SF (Taylor v. Kodes),
74 QF (Kodes v. Connors)
, 97 QF (Sampras v. Becker),
03 SF (Roddick v. Federer),
18 SF (Djokovic v. Nadal).
honorable mentions - Failed on 2nd point - 83 SF (Lendl v. McEnroe), 96 QF (Sampras v. Krajicek). Failed on 3rd - 89 SF (Lendl v. Becker), 93 QF (Sampras v. Agassi), 95 SF (Goran v. Sampras).
Australian Open (post 82) -
01 SF (Agassi v. Rafter),
05 SF (Federer v. Safin) (on the edge since Hewitt had a chance in the final against either),
06 QF (Federer v. Davydenko),
14 QF (Wawrinka v. Djokovic).
honorable mentions - Failed on 3rd - 85 SF (Lendl v. Edberg), 00 SF (Agassi v. Sampras), 21 QF (Djokovic v. Zverev); Failed on 1st - 97 QF (Sampras v. Costa).
Roland Garros -
76 QF (Panatta v. Borg),
88 SF (Wilander v. Agassi),
arguably 94 SF (Courier v. Bruguera),
97 QF (Kafelnikov v. Kuerten),
possibly 05 SF (Federer v. Nadal),
09 SF (Delpo v. Federer),
13 SF (Djokovic v. Nadal),
arguably 21 SF (Djokovic v. Nadal).
honorable mentions - failed on 1st - 95 QF (Costa v. Muster), Failed on 3rd - (arguable on 1st) 08 SF (Djokovic v. Nadal).
Seems like about 1-2 times per decade per major, we get a match that feels like a final. That averages out to about once every 8 majors.
Recently that seems about right - 21 RG SF, 18 WIM SF, 17 USO SF, 14 AO QF, 13 RG SF. Naturally, this seems like in periods of domination we see the final when it should happen (which makes sense, since the dominant player will make it to the final and should steamroll the competition. Every once in a while a player will be met in the semis that is better than the other half's best player, but that isn't incredibly common.)
