I’m mostly thinking of football, but I don’t see why this wouldn’t be the same for any sport. When you have two teams with the same conference record (let’s say 7-1) you have to go to the tiebreaker which usually or always is head-to-head. Seems logical that the best way to determine which team is best is to look at what happened when they played each other, right?
Here’s the first problem: It rewards a team for losing to a crappy opponent instead of a good one. Let’s say you have a situation where Team A beats Team B, Team B beats Team C, and Team C beats Team A. You have a 3-way tie at the top (in this scenario those are their only losses) and you have to skip the head-to-head and use another method to select the best team, and you can choose the team that dominated every game and whose only loss was in 3OT to another team with only 1 loss.
However, if Team A gives Team B their only loss, but then loses to a crappy team, then Team A wins the tiebreaker due to their head-to-head.
Second: Not every game matters. Ohio State just lost to Penn State. It doesn’t matter. If OSU beats Michigan, they still win their half of the Big North, the same as it would have been if OSU had beat Penn State. This loss only causes OSU to have a lower seeding in the playoff (assuming they win out). Similarly, Michigan could drop a game to a crappy team along the way, but as long as they don’t lose more than 1 and beat OSU, they will win their half of the Big North and most likely go on to make the playoffs. Penn State’s upset of OSU has knocked a little lustre off of the big OSU/UM game at the end of the season, but it hasn’t really changed the situation, because of the head-to-head tiebreaker.
Discuss
If we assume that the overall record of a team versus the (essentially) same competition is the best determinant of who was the better team overall during a given time period,
And if we have two teams who manage identical records on this basis, and, thus, overall have given us no indication of which is the better team,
Then it would seem that the fact that one team lost to the other is a pretty good indication that, at least at one specific point in time, the winner was better than the loser. Which admittedly doesn’t mean that it was overall the better team, but does at least give the important comparison otherwise missing from the equation.
If head-to-head was ignored, some other proxy for determining the “better” team would have to be substituted. Sadly, those almost always end up being statistics that rely upon some judgment as to what the “best” performance is. Soccer uses goal differential, which makes sense, because running up the score isn’t something you can easily do in soccer. But point differential in football would be an insidious determinant, for reasons that should be obvious.
In your example, assuming the Wolverines and the Buckeyes run the table until they face each other, and Ohio St. wins, can you truly fathom some system that allows Michigan to be the champions anyway? “Yeah, Buckeyes, you beat them fair and square. But because you pissed away that game to Penn St., and they only managed to lose to you, they’re better than you.”
:dubious:
Baked into the word tiebreaker we already know that by the most obvious criterion the teams in question are equal.
I think you have to look a bit more closely at the purposes of tiebreakers. There really are two.
The one people initially think of is to craft a tiebreaker which accurately looks at things other than the tied primary criterion to judge which team’s performance was actually better.
The second one is more practical. Due to the playoff mechanics, we simply need to choose a ‘winner’, even if the two teams really were tied in all meaningful measures. This choice method needs to be clear and unarguable even in situations where overall performance is objectively equal or at least arguably so.
Head-to-head is an absolutely perfect tiebreaker of the second type. You can’t argue how to apply it, it’s clear to everyone who the ‘winner’ is. (Ignoring the edge cases such as where three teams rock-paper-scissors…) It may even slightly indicate better performance like a test of the first type, making it slightly better than tossing a coin.
I guess it comes down to head-to-head being a perfect tiebreaker once you figure out both teams are objectively pretty close in power. I would only have a problem with it if there’s a good criteria left of the first type that hasn’t been used first, something obvious and useful that isn’t win record. That’s an order of tiebreakers issue, though, I don’t think it removes head-to-head from the list and it’s obviously specific to a given league’s rules.
Baseball (historically) has had the best tiebreaker - “okay, you’re tied? play one more game.” But that’s not really practical for a game like football where a single game involves so much wear and tear.
The NFL used to have a game to break conference ties (1958 Eastern, 1965 Western). But they stopped that probably when they added another round…Baltimore and Los Angeles were tied in 1967 but Los Angeles went to the playoffs on a tie breaker.
In 1977 there was some weird tiebreaker where the Colts would be better off if they deliberately lost to the lowly Detroit Lions in the next to last week…and they did getting a punt blocked instead of taking a safety and getting a free kick.
I actually see what you are saying. The Ohio State-Michigan game is at Ohio State. If Michigan’s only loss is a 3OT thriller at Ohio State, while Ohio State lost to a lowly Penn State team, you can almost make an objective argument that Michigan is better. However, I hope you have anti-pitchfork armor, because the Ohio State faithful would come for you. While you may be objectively right, I don’t see any way you can prove it.