Feb

26

Ivy Basketball Tiebreakers Are A Nightmare (Which Is Good For Columbia)

Written by

artist's rendition of Ivy playoff scenarios
artist's rendition of Ivy playoff scenarios

artist’s rendition of Ivy playoff scenarios

If you haven’t heard, this year is the inaugural Ivy League Basketball Tournament, in which four teams will compete in a bracket to determine who will represent the Ivy League in the NCAA tournament. Since the regular season now has a responsibility to assign the top four seeds, instead of just the top one, there are a lot of new tiebreaker scenarios which the Ivy League has to keep track of, especially on the men’s side.

Some of these edge scenarios are downright terrifying.

This morning, the @Ivy_Basketball Twitter account clarified a ruling on tiebreaker rules, sending the league’s team of analysts into a frenzy. Basically, the tweet says that one parameter for breaking ties, the tied teams’ records against other Ivy teams, will refer not just to teams in the tournament, but all eight teams in the league. However, most statistical models were imagining that only the teams who made it into the tournament would matter, so those models became wrong. The league’s tiebreaker incompetence, for its devoted followers, is nothing new.

Here are the rules, per the Ivy League, on who wins a tiebreaker if, for instance, Penn and Columbia end the season tied at 6-8:

  1. Compare the head-to-head record between the tied teams. If one team swept the other, they win.
  2. If still tied, compare the record of each team against the top-ranked team in the league. If that is tied, continue down the list of teams in the league, all the way down to the very last team in the league. Whoever has the better record wins.
  3. If still tied, compare an average of the rankings given to the tied teams by several analytical systems (BPI, RPI, KenPom, Sagarin). Whoever has the better record wins.
  4. If still tied, go to a coin flip.

This new ruling opens up some outs for Columbia. Because of Columbia’s weak out-of-conference performance, they don’t stand a chance to win Tiebreaker 3. But there are scenarios in which Columbia can make the 4-seed, or even the 3-seed, thanks to the new rules. For instance, if the Lions and Penn Quakers both go 2-0 next weekend, and the Yale Bulldogs secure the 3-seed, then Columbia can make the tournament with a Brown win against Cornell. (If you don’t trust me, I encourage you to check out my horrifyingly complex (and poorly formatted) chart of how every scenario works out.)

Here are the basic playoff scenarios that matter to Columbia. For playoff odds not accounting for slightly undefined edge cases, refer to @YaleSportsGroup’s odds:

  • If Columbia does better than Penn this weekend, they’re in the tournament.
  • If Columbia goes 2-0, Penn does not go 2-0, and Yale goes 0-2, the Lions will lock in the 3 seed thanks to their Tiebreaker 2 against Harvard.
  • If Columbia goes 2-0, Penn goes 2-0, and Yale goes 1-1, Penn gets the 4 seed with a Cornell win over Brown, and Columbia gets the 4 seed with a Brown win over Cornell.
  • If Columbia goes 2-0, Penn goes 2-0, and Yale goes 0-2, both Columbia and Penn make the tournament, while Yale falls out. Penn gets the 3 seed unless Dartmouth beats Princeton and Brown beats Cornell (very unlikely).
  • If Columbia beats Brown and loses to Yale, they secure the 4 seed if Penn loses to Harvard, but do not make the tournament if Penn beats Harvard.
  • If Columbia beats Yale and loses to Brown, they secure the 4 seed if Penn loses to Harvard. If Penn defeats Harvard while losing to Dartmouth, things get complicated. Penn gets the 4 if Cornell goes 2-0, and Columbia gets the 4 if Cornell goes 0-2. If Cornell goes 1-1 in this scenario, Penn gets if Princeton beats Dartmouth (likely), and Columbia gets in if Dartmouth beats Princeton.
  • If Columbia and Penn both go 0-2, Dartmouth can secure the 4 seed by beating Princeton. Otherwise, Columbia gets the 4 seed, unless Cornell goes 2-0.

tl;dr – Neither Columbia nor Penn truly control their own destiny. Columbia should try to win, obviously. Beyond that, they want Yale and Penn to lose, they want Brown to beat Cornell, and they want Dartmouth to do well.

I now want to go into some very edge-case scenarios. While unlikely, the League must have a procedure to deal with them. According to the League, “ties will be broken in rank order beginning with the highest seed.” That is to say, a tie for first will be settled before a tie for fourth. That way, the tie for fourth can refer who won first place to settle Tiebreaker 2. However, the League has yet to clarify some muddier scenarios. Would the League break a head-to-head tie for 6th place in order to settle a 4th place tie, or would they defer to Tiebreaker 3 before doing that? According to the current language, it seems more likely that the League would do the latter. However, nothing is certain. The question between which of those two would happen could be framed as such: Which does the Ivy League care about more? Outside analytics, or “meaningless” games between last place teams? I believe that the League simply wants to use Tiebreaker 3 as a slightly better alternative to a coin flip, and that they would settle low-level ties “out of order” in order to avoid going to Tiebreaker 3. My scenarios in the previous paragraph also take this view.

That’s all very abstract, so let me put forward a scenario. Columbia and Penn both win their next two games. Yale loses their next two games. Brown beats Cornell, and Princeton beats Dartmouth. This makes a 3-way tie for 3rd place (Yale, Columbia, Penn at 7-7) and a 3-way tie for 4th place (Cornell, Dartmouth, and Brown at 4-10). Yale falls out of the 3rd place tie because they got swept by Harvard, whereas Columbia and Penn both beat the Crimson. Columbia and Penn (T-3rd) are now tied against the 1st, 2nd, and 5th place teams. Does the League go to Tiebreaker 3 right there, or do they break the 6th place tie? If they do, Brown wins 6th place thanks to a 3-1 record against the other two teams. Columbia now has the tiebreaker against 6th place Brown, giving Columbia the 3 seed and Penn the 4 seed.

I await word from the Ivy League, and will update this article when more word comes out from the Ivy League. I want to discuss one more scenario first, though. Imagine that Columbia, Princeton, and Penn go 0-2, while Harvard, Dartmouth, and Cornell go 2-0. There is now a 2-way tie for 1st place, and a 4-way tie for 4th. Assume the League breaks down the 4th place tie, in order to break the 1st place Tiebreaker 2. Columbia, Penn, Dartmouth, and Cornell are all tied at 5-9, and are all a combined 3-3 against each other. Dartmouth wins Tiebreaker 2 against Princeton, whether Princeton is the 1 or the 2 seed, securing the fourth spot. Then Harvard retroactively wins the 1st place Tiebreaker 2 against Dartmouth, knocking Princeton out of the first seed.

The fact that we can even consider these scenarios is evidence of the League’s incompetence to truly define its parameters. What’s troubling is that the League now has to “clarify” a rule, effectively making a new one a week before the season ends. That ruling will inherently benefit certain teams in some fringe scenarios more than others. Hopefully, clarifications this year will prevent the league from falling into this trap next year. But it seems that the Ivy League always finds a way.

Tags: , , , , , ,

© 2006-2015 Blue and White Publishing Inc.