Who's Your New Top 5 in D1?

[Champ]

I'm personally excited for UMD - Michigan this weekend, hopefully that will help sort some things out

Not sure about that... if UMD loses to Michigan, where does that put them?

It's probably just that I'm excited to watch the game.

[laxdad03]

Not sure about that... if UMD loses to Michigan, where does that put them?

If they're truly #5, then they are expected to lose to CSU (who they lost to already) and Michigan (assuming Michigan is ranked #1). Does a loss to Michigan drop them lower than 5?

If you're a poll voter... and the #5 team loses close games to all four teams above them, do they drop?

If they have multiple losses, it drops them below the 1 loss teams who are similarly ranked, but above the multiple loss theams that, in the judgement of the voter, that team is better than.

Catlax, do you really mean to imply that rankings should strictly reflect number of losses in the way you describe, and that the "judgement of the voter" only comes into play between teams with similar numbers of losses? I really think rankings can't be done on any particular strictly defined criteria, they need to be a synthesis of a lot of factors, taking into account, among other things, who wins and losses are with, and by how much. For example, I can't see people seeming to want to automatically rank Oregon above teams that have only close losses to the top couple teams, simply because Oregon hasn't lost and they have. Not to pick on Oregon, but they haven't really been tested at all yet this year, let's wait until they have a chance to earn better/top rankings by playing against some top teams. It also seems to me that really close losses (e.g. 1 goal, down-to-the-wire, could-easily-have-gone-either-way) should be taken to mean that the two teams involved are more likely to be very closely matched (rather than automatically leading to huge ranking statements based on "a W is a W"), and as a result those two teams should probably be ranked similarly (i.e. without a lot of teams necessarily being ranked in between them), whereas significant (several goal) losses MAY more likely indicate a bigger difference in ability. Of course, all this needs to be done with some latitude, as teams' level of play and resulting scores will naturally vary from week to week. That's why, for example, things happen like A beats B, B beats C, C beats A. But we haven't had much of that at the top this year yet, mostly we have a bunch of very good teams with few or no common top opponents so far, which makes it tough.

The next poll will be truly interesting -- for what it's worth. Let's trust that pollsters all use their best overall judgement, whatever that may be in each individual case, and that having a large number of them will tend to even things out to a reasonable overall ranking result (particularly more so by the end of the season), as is appropriate. And then, of course, in the tournament, a W really is a W, and that will be what really matters...then.

[laxdad03]

If you're a poll voter... and the #5 team loses close games to all four teams above them, do they drop?

As a non-poll voter (so take it for what it's worth) I say it depends on their scheduling. If you only scheduled Top 5 and the rest of your season is non-ranked, then you drop, and potentially if the scores are not close, you could drop significantly.

If you have a few games around #8-10 and another couple games around #12-14, then they would not drop as much, because I'd have additional chances to evaluate them. So if the loses in the Top 5 were close, then maybe 1 spot, if lopsided then 2 or 3 spots.

First off, Zamboni, I really like (and appreciate) your analyses (both for extra effort and for rationality in most cases), you deserve all our thanks.

Just one general comment, intended broadly but prompted by your last statement: I don't think that rankings should/can be usefully thought of simply in terms of "number of spots gained/dropped" (i.e. with respect to a previous poll position), but rather in terms of rankings relative to the other teams around them, however many there may be. Suppose there were 20 teams (I know, a bit of a ridiculous example, but...) who had played each other very closely and seemed for all reasons to be of similar caliber, and one or two of them had played top teams fairly closely; and then another team, who had previously been ranked above all of the 20 (perhaps due to lack of common opponents, or whatever), lost badly to one of the top teams; I wouldn't think it odd (in lack of other countermanding data) to drop that team below all 20 of the others. On the other hand, a bad loss to a team ranked way above everybody else wouldn't necessarily cause any drop whatsoever, if there were no reason (common opponents, or whatever) to think that the teams below them were any better. I think too many people here seem to place too much emphasis on "they dropped (or gained) x spots" rather than whom, on a relative position, they descended below (or rose above).

[Zamboni_Driver]

I think too many people here seem to place too much emphasis on "they dropped (or gained) x spots" rather than whom, on a relative position, they descended below (or rose above).

Completely agreed. My statements were simplified. I've stated this before that I go by a points system in evaluating teams. For example, #5-8 may all be 1 point away from each other. #9 may be 11 pts less than #8. If #8 loses, that doesn't mean they drop below 9 and 9 moves up - it could mean the spread between 8 and 9 decreases, but they stay in the same ranking. Just because 8 and 9 are close in rank doesn't mean they are close in talent. And the flip side, just because 25 and 11 are far apart in rank, doesn't mean 11 has nothing to worry about.

Suppose there were 20 teams (I know, a bit of a ridiculous example, but...) who had played each other very closely and seemed for all reasons to be of similar caliber

And as far as this - it isn't ridiculous at all - I believe it is the future. Just look at the number of team vying for #12-#16, and then again look at all the teams looking to hold on or jump into a T20-25 spot.

[CATLAX MAN]

Not sure about that... if UMD loses to Michigan, where does that put them?

If they're truly #5, then they are expected to lose to CSU (who they lost to already) and Michigan (assuming Michigan is ranked #1). Does a loss to Michigan drop them lower than 5?

If you're a poll voter... and the #5 team loses close games to all four teams above them, do they drop?

If they have multiple losses, it drops them below the 1 loss teams who are similarly ranked, but above the multiple loss theams that, in the judgement of the voter, that team is better than.

Catlax, do you really mean to imply that rankings should strictly reflect number of losses in the way you describe, and that the "judgement of the voter" only comes into play between teams with similar numbers of losses?

No, not at all. I was trying to answer the specific hypothetical question that Nate asked whereby if the #5 team loses close game to 4 teams above them, how far do they drop? I think if you have 1 loss teams that are similar in talent & body of work (like our group of the top 7-8 teams are right now), they probably should get the ranking over the 4 loss team. However, I will say that you would never hear from this corner that a voter should suspend his judgement......even in this situation.

Our group of voters have proven over and over throughout the years that they are well-informed about the relative strength of teams and that they generally "get it right." Everyone, from time to time, questions the relative placement of teams when the polls come out, but I have yet to see an undeserving team at the National Tournament. That should be evidence enough that this system works pretty well.

[laxdad03]

Good, I thought I had understood (and appreciated) you better over the years.

I think my answer to Nate's question might have been "No, not necessarily, unless there is something else to base such a drop on." I agree wholeheartedly, "body of work" is key, and there may well be something subjective in that concept (at least, there can be varying viewpoints that can all be rationally defended -- although there may be still other viewpoints that aren't always quite so rational, as we sometimes see). And yes, I like our voters/system, I think on the whole they do an excellent job.