Let’s Get KRACH-ing

A name as goofy as KRACH deserves bad, tasteless puns in the headlines. But now we get serious …

We’ve said time and again that, if you want to adhere closely to numbers as a way to select and seed the NCAA tournament field, be sure to use a system designed to accomplish this.

For two years, we’ve advocated KRACH (can we think of a better name, please?) as the best ranking system, and we’re here to do it again.

And to illustrate more clearly than ever, let’s return to the curious case of Wisconsin.

Wisconsin is currently rated much lower in the Pairwise (14) than in KRACH (6). But more importantly to this illustration, the fact that Wisconsin defeated Alaska-Anchorage in last week’s best-of-3 WCHA playoff series, 2 games to 1, actually dropped the Badgers in the Pairwise. Had Wisconsin lost Game 3, it would be sitting around 9 in the PWR, and practically be a lock to make the NCAAs. Now, they are right on the bubble and struggling for survival.


Because Alaska-Anchorage, by virtue of losing, is no longer a “Team Under Consideration” (i.e. any team with a .500 or better RPI). That means Wisconsin’s 6-1 record against UAA is not factored into the Badgers’ “Record vs. TUCs,” which is one of the four criteria that make up the Pairwise. So Wisconsin’s Record vs. TUCs is now much weaker, which makes them now lose comparisons with various teams they might otherwise win.

This is not good, but I have a slightly different take on this than some people.

First, let’s forget what KRACH says for a second. KRACH has Wisconsin at No. 6, and so that probably does definitively tell us that Wisconsin is getting the short end here. But within the rules of the Pairwise, there’s no way to definitively state that.

Perhaps it’s correct that UAA is not a TUC, and therefore, it’s perfectly appropriate that Wisconsin doesn’t get to count its wins against it. After all, Cornell doesn’t get to piss and moan that its four wins against Clarkson are not factored in. It just so happens that it’s obvious what happened to Wisconsin, because UAA was so precariously on that RPI .500 line just at the time it was playing a team in the playoffs that was directly affected by it. So the effect is obvious, but it doesn’t make it wrong.

No … the REAL issue is not whether that’s “correct” or not. The REAL issue is that no team should ever be in a position where losing helps them get into the NCAA tournament. And that is an overall philosophical situation that is completely irrelevant as to whether the Pairwise is correct or not. It’s completely separate.

What we’re saying is, forget what ranking system is right or wrong or flawed. The fact of the matter was, in this case, for whatever reason, if Wisconsin lost, it would be better off.

That’s no good.

On top of all the obvious reasons it’s no good, it also opens the door to a team purposely tanking games. Not that Wisconsin, or anyone else would do that, but why would you want that to be a temptation?

For that reason alone, the Pairwise needs to be fixed.

One way is to introduce a sliding scale to the TUC line, so that it’s not .500 RPI or nothing (otherwiwse heretofore known as the TUCliff).

The other way is to use KRACH and stop worrying.

Say Yes to KRACH

(For an even more thorough overview of this topic, see the “Relevant Links” at the top left of this page.)

There’s a heckuva discussion going on at the USCHO Fan Forum regarding KRACH, and it’s getting quite the rhetorical challenge from some skeptics. But there’s some eloquent defenses going on there. I’ll try to summarize so you don’t have to wade through the slop.

Really, the only criticism of KRACH comes down to complaints that it too heavily weights Strength of Schedule. And while there is technically no SOS component in KRACH (it doesn’t really break it out that way), we’ll leave that statement as is for rhetorical purposes.

Some folks believe good teams in weaker conferences get hurt because, since they only play and beat up on those weaker teams, they don’t have a chance for their KRACH ranking to be better by playing some better competition. They can only beat who they play.

And while this is true, this is not really a flaw in KRACH at all. KRACH can only tell how you’ll do against the schedule you play. It cannot speculate how you’d do against a slightly better schedule. That, after all, is the whole point of these ranking systems.

I’d argue that people only perceive KRACH to be over-weighting Strength of Schedule because the Pairwise is not doing a good enough job of weighting it properly. In fact, if you took any of the numerous sound-math ranking systems out there — those used in the football BCS for example — they would all look a lot more like KRACH than PWR.

The beauty of KRACH, however, is that those good teams won’t get hurt by playing that weak schedule. If two teams go 7-4 against the Top 10, they would have the same KRACH ranking. Now, if Team A goes 20-0 against the bottom 10 KRACH teams, and Team B goes 12-8 against teams ranked 11-30 in KRACH, they would have approximately the same KRACH ranking still.

Some would argue that it’s not the fault of Team A for playing those weak sisters, and a team with a 27-4 record should be much higher. Problem is, you cannot prove that Team A would still beat those teams ranked 11-30. If it did, their ranking would be higher.

By the same token a team like Minnesota, which has a lot of losses to good teams, is perfectly appropriate to still be near the top in KRACH. Why is it hard for people to understand that, if good teams beat each other up they are still good teams, even though their records may start to creep more towards .500? Meanwhile, Cornell’s No. 7 rank in KRACH is also appropriate. The Big Red have taken care of business.

Taken to the extreme, if you play only Top 5 teams, and go .500, you still belong as a Top 5 team yourself (or at least 6). Just because you have a .500 winning percentage, doesn’t mean you are mediocre. I’m not sure why this isn’t obvious to some people.

So, a team is what it is. Its ranking is based upon what it did. Every win helps and every loss hurts, it’s just a matter of how much. KRACH decides its rankings based upon the results, and if you take the odds of winning based upon the rankings, and replayed the schedule, you’d come up with the what the team’s record actually is.

It’s pure. That’s the beauty of KRACH. There’s no way around that.

The hockey world needs to embrace straight KRACH.


Thing is, the hockey world seems to want to reward good wins, and thus place higher importance on a good win than a bad loss. I disagree with this philosophy, but if that’s what the hockey world wants, we need to deal with it.

There is also a concern that too many teams from one conference gets into the tournament, and I think this is a legitimate concern.

But there are ways to incorporate the good wins philosophy right into the KRACH, or at least use the KRACH as a basis for creating a sliding scale record vs. TUC. If you’re going to create other criteria based on certain philosophical or political considerations, at least use good numbers and sound mathematical principles as a basis point.

And if you’re worried about having too many teams from one conference, simply make a four- or five-team cutoff.

One concern you always hear about KRACH is that it’s too complicated to understand, or that a sport is not going to just change to something that is so unknown.

Well, you don’t need to do the math to understand KRACH conceptually. You can just read here: The USCHO KRACH FAQ.

As for something unknown … it should be understood that a couple of very KRACH-like ranking systems are utilized by football as part of the seven non-human-ranking components of the BCS.

More defense of KRACH

This was posted by “Rich” on the USCHO Fan Forum, and is yet another nice summation of why KRACH is better than PWR, and why it accomplishes what people would hope a rating system would do:

1. If a team wins a game, they always increase their rating in KRACH. PWR can drop a team for winning, and it can happen in RPI too.

2. A team’s rating always increases more for beating a higher-ranked team than beating a lower-ranked team. That’s not always true for PWR.

3. A team’s KRACH rating always drops if they lose, and by more if they lose to a lower-rated team. Losing to the No. 1 team cannot increase a team’s KRACH rating. If KRACH rewarded SoS somehow, then this wouldn’t be true. KRACH does not reward losses to good teams. Period.

4. Tying a better team increases a team’s KRACH rating, and tying a worse team decreases it. This makes perfect sense, but is not always true with PWR.

5. A .500 team’s KRACH rating is exactly equal to their strength of schedule.

6. If two teams have the same record, the team who has played the toughest schedule is always ranked higher. If two teams have played the same schedule, the team with the better record is ranked higher.

7. If a hypothetical team goes .500 against teams in a given place in the rankings (say, 1-5), they will be ranked within that group (assuming those are the only games that team has played). Neither RPI nor PWR guarantee this.

And most importantly:
8. Given every team’s KRACH ratings and the schedule they have played, you can calculate the team’s record exactly! If the strength of schedule component was exagerated in any way, this wouldn’t work. This is the most convincing evidence that SoS is properly weighted in KRACH. You sure can’t say that about PWR or RPI!

If you look at all of those reasons, you can see that they are exactly what one would hope a rating system would do. I find it amazing that you can take all of the given KRACH ratings and the schedule any given team has played and work out what their winning percentage is. Well, amazing until I understood how it’s calculated.

Thanks, Rich. Case closed.