I agree that most sports are watering down the playoffs by allowing too many teams in. The NFL allows 14/32, why even have a regular season at that point? Slight hyperbole, but you don't have to be very good to make it. Among pro leagues, I thought MLB had the right balance with 8/28 to 8/30 teams. You had to be pretty good to make the playoffs but there was still a chance for a team that started slow but improved considerably as the season progressed. But now they've expanded it to 12/30. NCAA I-A football with 4 teams is much more exclusive, but the talent disparity is larger than most pro leagues.
So I agree that 12 teams would be way too many (and let's not forget injuries as a factor the more games you play,
both Georgia and OSU lost their starting tight ends in-game, you don't want both teams banged up by the time you get to the championship). The three-team idea is interesting, but I think it really depends on the year, and that's the core challenge. You could have four 13-0 Power 5 champions. Technically you could have five (and possibly a 13-0 Group of 5 champion like Cincy a few years ago as well). You could have a team that lost early but looked much better and had big wins down the stretch. And I think this year showed that the "least undeserving" teams can hang right with the 13-0 ones in some years. There's no perfect solution.
EDIT: Numbers are out, twenty-two million viewers for UGA/Ohio State, best for primetime semi since the first year of the playoffs. So apparently having a good matchup trumps having a game on NYE ; p
What's weird to me is that it's the fourth highest rated semi of all time so that means there were two afternoon games that beat it . . ?
I
looked it up, and yes. One of them was Florida State-Oregon in the first year, apparently people were really interested in the playoff semis that first year. The other was Georgia-Oklahoma on 1/1/2018. I guess that was a really compelling matchup too, and it wound up going to overtime so it lived up to the hype. Also was not on New Year's Eve.
Looking at it in list format, it's clear that matchups matter, but when it's played also matters. This year had good matchups, but January 1 dominates the list (the only medium-viewership 1/1 games being in the pandemic year when they let a 6-0 team in, the rest all being at the top), and 12/31 bringing up the rear and generally faring poorly, with this year's Saturday 12/31 games with good matchups being the first time 12/31 has done above-average. The previous-highest 12/31 games, in 2016, were also on a Saturday.
So I think it still shows that ESPN was right, 12/31 is a poor choice, especially on weeknights, and 1/1 is a great choice.