Re: Hockey East - Who's in, who's out, who's home: by the numbers - 2012-13 edition
Short answer: Well, yes.
Longer answer: It's confusing because we don't have a good enough sample set to be positive what the league is going to do, although we used to.
If you dig through the historical reference material I recently posted, there are two main summary points at work here.
1) In a three-way tie (to keep it simple), there are three ways to split that comparison: Top-down, bottom-up - each of which removes a team and restarts with the remainder - and all-at-once, which looks at the stack and takes it as a whole if it has an order.
2) The league had demonstrated for years that it used to use the bottom-up method, but then, in what ended up being the first year with an actual - as opposed to hypothetical - 3-way tie, they changed the interpretation, but not the letter, of the rule.
Perhaps the best example of the impact of these is the UVM/NU/UMA example listed in the links from 2010. In that case, UVM and UMA would always be in that order, but NU could be 1st, 2nd, or 3rd, depending on which method was used.
In the case that actually arose (BU/ME/UML, IIRC), it turns out that all three methods would result in the same order (BU was 2-1-0 against both, ME was 2-1-0 vs UML: BU (4-2-0)/ME (3-3-0)/UML (2-4-0) regardless, without leaving the H2H matchups). That's off the top of my head, so if I got the teams wrong, that's my fault. If I got them right, then the Alzheimer's hasn't set in yet because I can recall the details of a three-way tie-breaker (and the hypothetical UVM/UMA, where is NU?) from three years ago.
In the case above:
All-at-once would look at them as they are and see that among all three, there is a clear distinction (PC is +.500, BC is .500 and UNH is -.500). Just call it a tie-break and seed it that way.
The "old" way of bottom-up is what I detailed, but is what shouldn't happen anymore.
Top-down is what you outline, and is probably what they would do. I say that because the rule says to reduce and restart. Since they're not reducing from the bottom anymore - and a-a-o (my preference, if it's available) doesn't reduce, it just settles - then top-down is the only method that still complies.
So, yes, PC would be promoted and they would start over with BC v UNH H2H. Since that's 1-1-1, they would next go to wins, which is currently in BC's favor.
Since a-a-o and t-d gave the same result, I didn't break them out separately - only doing the *other* configuration - but perhaps I should have for clarity.
Since we're now in a four-way tie, this is literally yesterday's news, but it is still illustrative of the mechanics - and might end up being a tie again for 1st, or for the last Home Ice slot, or anywhere in between (I'm not going to do the math to see if they can't tie for 3rd).
Originally posted by jeteye1717
View Post
Longer answer: It's confusing because we don't have a good enough sample set to be positive what the league is going to do, although we used to.
If you dig through the historical reference material I recently posted, there are two main summary points at work here.
1) In a three-way tie (to keep it simple), there are three ways to split that comparison: Top-down, bottom-up - each of which removes a team and restarts with the remainder - and all-at-once, which looks at the stack and takes it as a whole if it has an order.
2) The league had demonstrated for years that it used to use the bottom-up method, but then, in what ended up being the first year with an actual - as opposed to hypothetical - 3-way tie, they changed the interpretation, but not the letter, of the rule.
Perhaps the best example of the impact of these is the UVM/NU/UMA example listed in the links from 2010. In that case, UVM and UMA would always be in that order, but NU could be 1st, 2nd, or 3rd, depending on which method was used.
In the case that actually arose (BU/ME/UML, IIRC), it turns out that all three methods would result in the same order (BU was 2-1-0 against both, ME was 2-1-0 vs UML: BU (4-2-0)/ME (3-3-0)/UML (2-4-0) regardless, without leaving the H2H matchups). That's off the top of my head, so if I got the teams wrong, that's my fault. If I got them right, then the Alzheimer's hasn't set in yet because I can recall the details of a three-way tie-breaker (and the hypothetical UVM/UMA, where is NU?) from three years ago.
In the case above:
All-at-once would look at them as they are and see that among all three, there is a clear distinction (PC is +.500, BC is .500 and UNH is -.500). Just call it a tie-break and seed it that way.
The "old" way of bottom-up is what I detailed, but is what shouldn't happen anymore.
Top-down is what you outline, and is probably what they would do. I say that because the rule says to reduce and restart. Since they're not reducing from the bottom anymore - and a-a-o (my preference, if it's available) doesn't reduce, it just settles - then top-down is the only method that still complies.
So, yes, PC would be promoted and they would start over with BC v UNH H2H. Since that's 1-1-1, they would next go to wins, which is currently in BC's favor.
Since a-a-o and t-d gave the same result, I didn't break them out separately - only doing the *other* configuration - but perhaps I should have for clarity.
Since we're now in a four-way tie, this is literally yesterday's news, but it is still illustrative of the mechanics - and might end up being a tie again for 1st, or for the last Home Ice slot, or anywhere in between (I'm not going to do the math to see if they can't tie for 3rd).
Comment