PDA

View Full Version : 2019 NCAA Tournament Thread - Regionals are the best weekend of hockey all year



Pages : 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 [30] 31 32 33

CLS
04-05-2019, 11:56 AM
In addition, the empty net goals are pretty much unique to hockey. There's a huge difference between a 5-2 game in which a team leads throughout and a game that 3-2 after 58 minutes and ends 5-2.

Fishman'81
04-05-2019, 02:39 PM
Predictive systems, which are interesting to the user of course, seem inappropriate for an NCAA tournament. It seems much better to reward the teams for their performance already done. Of course, this is a point given to argument.

Two sides of the same coin..? The predictions would have to based on (the season's) performance.

dxmnkd316
04-05-2019, 02:50 PM
Margin of victory is worthless in hockey. It may have some value in basketball, because of the larger sample size of points scored. I don't see how you can definitively determine anything by analyzing MOV in hockey. What? Does a 5-2 win "mean" more than a 6-1 win? Everyone knows that you can win 7-0 one night and lose the next. It's not a sport where enough "points" are scored to rely on a margin of victory stat.

And the only thing that irritates me more are those stupid "total goals" series that we used to have. The object of the game is to WIN. I don't care if it's 1-0 or 25-2. There is NO DIFFERENCE. And I don't even like total goals being used as a tiebreaker for some of these tournaments (like World Juniors, for example).

I think your hatred for total goal series, which were stupid though I was never around for them, is clouding your judgment on the use of MOV. In basketball, teams that are up get fouled and shoot free throws. How does that not disqualify it based on your criteria. Those points are essentially ENGs, hell, itís in the name, Ďfreeí.

And to answer your question, yes, if A beats B 5-0, itís objectively better than C beating B 1-0.

mookie1995
04-05-2019, 03:30 PM
In addition, the empty net goals are pretty much unique to hockey. There's a huge difference between a 5-2 game in which a team leads throughout and a game that 3-2 after 58 minutes and ends 5-2.

no.... pretty much each is a W :D

manurespreader
04-05-2019, 05:44 PM
I understand that there was A dude out in vegas who had AIC over St. Cloud and a couple of other Minnesota teams winning for that same night all of whom won. His 10 dollar bet turned into 26,000 some odd.

Nice.

Fishman'81
04-05-2019, 06:19 PM
I understand that there was A dude out in vegas who had AIC over St. Cloud and a couple of other Minnesota teams winning for that same night all of whom won. His 10 dollar bet turned into 26,000 some odd.

Nice.

Wonder what the spread was on the AIC game..?

chickod
04-05-2019, 06:20 PM
And to answer your question, yes, if A beats B 5-0, it’s objectively better than C beating B 1-0.

NO. It's not. See....that's the point. Define "better." The idea is to WIN. Period. Since when did the MOV mean anything?

dxmnkd316
04-05-2019, 06:31 PM
NO. It's not. See....that's the point. Define "better." The idea is to WIN. Period. Since when did the MOV mean anything?

lol ok

Sean Pickett
04-05-2019, 09:53 PM
However, what is interesting, but difficult to do scientifically, is something like this:
In games where one school had a significant travel advantage (mostly the 2 eastern regionals, for obvious reasons, but sometimes in the west as well)....and significant is difficult to define, the 'home teams', regardless of seeding, are running about 60% winners.

Take that for what you will in the idea of 'what's a good system'....Over on the Annual thread in which the absurdity of the current regional system is discussed (https://board.uscho.com/showthread.php?135456-Annual-thread-in-which-the-absurdity-of-the-current-regional-system-is-discussed&p=6789993) thread I posted the records (https://board.uscho.com/showthread.php?135456-Annual-thread-in-which-the-absurdity-of-the-current-regional-system-is-discussed&p=6789993&viewfull=1#post6789993) for all games played within 75 or so miles of one team, broken down into three groups: true home games, games in the same/neighboring city and within 75 miles. The combined record is 74-40 for a 0.649 winning percentage. However, that does not take into account if the winner was a higher or lower seed, so I looked at those numbers as well.
higher seeded "home" team vs lower seeded "road" team: 54-14 0.794
higher seeded "road" team vs lower seeded "home" team: 18-20 0.474
I also checked all 292 NCAA regional games played:
higher seeded team vs lower seeded team: 189-103 0.647
From this it can be seen that playing near home vs a team that had to travel does makes a significant difference.

Sean

Fishman'81
04-05-2019, 10:25 PM
100% agree.

Tried to hold off just for the sake of propriety, but is it actually possible that you don't grasp the obvious fact that "predictive" is based on results?

I would guess so, on the the basis of your previous babble.

Kepler
04-05-2019, 11:30 PM
Tried to hold off just for the sake of propriety, but is it actually possible that you don't grasp the obvious fact that "predictive" is based on results?

I would guess so, on the the basis of your previous babble.

You are punching way above your weight, son. Stay down (https://en.wikipedia.org/wiki/Dunning%E2%80%93Kruger_effect)...

Fishman'81
04-05-2019, 11:38 PM
You are punching way above your weight, son. Stay down (https://en.wikipedia.org/wiki/Dunning%E2%80%93Kruger_effect)...

Really..? Make a cogent argument to the contrary.

I'm waiting, but not expecting much, given your history.

Lump yourself into the same mass of humanity that has no idea how a statistical model works, and go tee it up, or whatever.

cetihcra
04-06-2019, 12:28 AM
I understand that there was A dude out in vegas who had AIC over St. Cloud and a couple of other Minnesota teams winning for that same night all of whom won. His 10 dollar bet turned into 26,000 some odd.

Nice.

$100 bet, but yeah. He parlayed 4 bets: AIC to win, under on the O/U (6.5), the Wild to win, the T-Wolves to win. All 4 bets were dogs.

r

Numbers
04-06-2019, 11:37 AM
Really..? Make a cogent argument to the contrary.

I'm waiting, but not expecting much, given your history.

Lump yourself into the same mass of humanity that has no idea how a statistical model works, and go tee it up, or whatever.

Fishman,
I won't answer for the other poster here, but I would like to engage this conversation a bit more.....
These are my assumptions:
1- The NCAA (coaches, ADs) desires a system which is objective rather than subjective, since subjective led, in past years, to supposition of favoritism toward name brand schools
2- Any subjective system is going to be based on math in some way
3- Since it's entirely subjective, the math needs to be somewhat easy to understand.

I would argue, actually, that the current PWR is too complicated, but that's another discussion. For now, I think it's safe to say the above are true.

Then, we have the question:
What makes a predictive system differ from a system which merely evaluates the prior results?
This is a good question. Obviously, a predictive system must have some basis in data, or it can't predict anything. In this way, of course a predictive system is based on results. It can't be based on anything else (pulling rabbits out of a hat?).

I would suggest, however, that predictive systems likely make use of a far greater collection of information....for example, although I have not looked at any other system other than the 2 most commonly discussed here, I would assume that other systems use things like:
Margin of victory
What the opponents record (or RPI or whatever) was in the day you played them
Injuries to key players
Goalie rotation
ENGs

etc.

There are 2 issues with such systems, to me.
1- There is no way of knowing exactly how to weight all of those inputs, so you are guessing, in some sort of way
2- There is no way of knowing how good the predictions are

For these reasons, I would personally prefer a system which uses only KRACH, adjusted for home ice. That particular system does have this one advantage:
Given the rankings, one can replay the entire year and get the records of all of the teams correct. That is a very powerful argument.

Of course, there are issues there as well. The nonconference season weighs too heavily in some sense, because 8-10 games on average for each team decides a range for an entire conference to fit into....

So, it's not perfect. But, to me, it's better than anything we've got going.

Would you like to flesh out a predictive system which you prefer?

Fishman'81
04-06-2019, 06:11 PM
You are punching way above your weight, son. Stay down (https://en.wikipedia.org/wiki/Dunning%E2%80%93Kruger_effect)...

Just clicked on the link, and you might have something there lol. (Me, and that guy who lives on Pennsylvania Avenue...)

Sorry to get impatient; my bad.

Fishman'81
04-06-2019, 06:35 PM
Fishman,
I won't answer for the other poster here, but I would like to engage this conversation a bit more.....
These are my assumptions:
1- The NCAA (coaches, ADs) desires a system which is objective rather than subjective, since subjective led, in past years, to supposition of favoritism toward name brand schools
2- Any subjective system is going to be based on math in some way
3- Since it's entirely subjective, the math needs to be somewhat easy to understand.

I would argue, actually, that the current PWR is too complicated, but that's another discussion. For now, I think it's safe to say the above are true.

Then, we have the question:
What makes a predictive system differ from a system which merely evaluates the prior results?
This is a good question. Obviously, a predictive system must have some basis in data, or it can't predict anything. In this way, of course a predictive system is based on results. It can't be based on anything else (pulling rabbits out of a hat?).

I would suggest, however, that predictive systems likely make use of a far greater collection of information....for example, although I have not looked at any other system other than the 2 most commonly discussed here, I would assume that other systems use things like:
Margin of victory
What the opponents record (or RPI or whatever) was in the day you played them
Injuries to key players
Goalie rotation
ENGs

etc.

There are 2 issues with such systems, to me.
1- There is no way of knowing exactly how to weight all of those inputs, so you are guessing, in some sort of way
2- There is no way of knowing how good the predictions are

For these reasons, I would personally prefer a system which uses only KRACH, adjusted for home ice. That particular system does have this one advantage:
Given the rankings, one can replay the entire year and get the records of all of the teams correct. That is a very powerful argument.

Of course, there are issues there as well. The nonconference season weighs too heavily in some sense, because 8-10 games on average for each team decides a range for an entire conference to fit into....

So, it's not perfect. But, to me, it's better than anything we've got going.

Would you like to flesh out a predictive system which you prefer?

I'm with you on this one. I've been touting the KRACH for years and years, with the RPI being my second choice, and the PWR my third... But any objective and transparent metric is better than a beauty-pageant, as you stated. (As in, the old coaches' polls way.)

This conversation brings to mind -in a tangential way- the furor that ensued in D-1 football some years ago.

I think it was TCU who was 4th in the BCS going into their final RS game... And the Horned Frogs won that game, but were leap-frogged by the team behind them, simply as a result of the #5 team having a better quality win in a statistical sense. I used to live near TCU, and had a dog in that hunt, but it passed the eye-test for me just the same. Yet, the nation was outraged!

We have a pretty good thing going on in college hockey right now. Yes, there are undeserving teams getting-in via tournament AQ's every year, but in a general sense, we're seeing the cream of the crop in the Nationals.

Numbers
04-06-2019, 08:37 PM
I'm with you on this one. I've been touting the KRACH for years and years, with the RPI being my second choice, and the PWR my third... But any objective and transparent metric is better than a beauty-pageant, as you stated. (As in, the old coaches' polls way.)

This conversation brings to mind -in a tangential way- the furor that ensued in D-1 football some years ago.

I think it was TCU who was 4th in the BCS going into their final RS game... And the Horned Frogs won that game, but were leap-frogged by the team behind them, simply as a result of the #5 team having a better quality win in a statistical sense. I used to live near TCU, and had a dog in that hunt, but it passed the eye-test for me just the same. Yet, the nation was outraged!

We have a pretty good thing going on in college hockey right now. Yes, there are undeserving teams getting-in via tournament AQ's every year, but in a general sense, we're seeing the cream of the crop in the Nationals.

Thanks. The RPI (school) website has a ranking called KASA, which means KRACH Adjusted for Site Advantage. It's the same math, with the slight change that there is a constant nation-wide advantage to playing at home. Whatever that advantage is can also be determined purely mathematically in the same way as KRACH does now. This makes it slightly more robust, in my eyes.

I agree that PWR, with its possible dependence on the very small data set of H2H games and CommOpp games, is poorer as a result of those things.

I also think that KASA or KRACH have zero chance of ever being adopted. In reality, the present PWR math is very complicated, but it seems conceptually easy. KRACH or KASA is conceptually difficult, but not as complicated. For example. I have a very simple EXCEL spreadsheet for the NHL, which, if kept updated with games played and results, can be used to calculate the Bradley-Terry Rankings for all teams in about 3 minutes, manually doing the iterations. I could write a simple VBA which would automate, that, I think. And, It might take me an hour to write the code. It's probably quicker than calculating PWR, simply because it consists of simply adding a bunch of fractions. But, it's conceptually difficult, so it won't be adopted.

I personally think that the math guys at the NCAA have micromanaged their RPI definition to try to make it mimic the KRACH results. I say that because the RPI is a very complicated beast right now. Differing weights for home and away games. 21/25/54...where does that come from? Nowhere, really. It's just an empirical number. And, the quality win bonus is very bogus. It's all clearly a magic formula. So, I'm okay with it. It's close to KRACH. And, I think that's the best we get.

But, that doesn't really answer this question: I would consider KASA or KRACH to be NOT predictive formulas. They merely postdict the season. You may say they give you odds on the next game in a way, but I don't really think they are robust enough to be considered predictive.

The discussion here led me to believe that you had something else in mind. Was I mistaken?

Thanks.

pgb-ohio
04-07-2019, 10:06 AM
I wanted to take this item seriously, and tried to look at it from the point of view of each of the 8 teams seeded #2 or #3 this year. My conclusion? Nobody's talking about it because none of those teams were treated unfairly. Including Ohio State. Given the constraints of this year's rules & sites, of course.

Do you have a specific example from a previous tournament?



Here's one (though I really don't want to go back and analyze every bracket), and it's just from 2018: Northeastern and BU were put as (#3 and #4) in Worcester (Holy Cross was the host). Believe me, the support those two teams (specifically BU) would get in Worcester far surpasses what PC gets in Providence. Of course, now the question is - was there a better solution? Swapping BU with Air Force, or even better with Tech (in Bridgeport)? Or 2015, where everyone talks about PC's placement as #4 seed, but BC's placement as #3 seed is not that much more fair, at least with respect to Denver (could have put them in Manchester and bring Minnesota in, or South Bend and swap with Harvard). I don't know, it's all history now... Sorry to be so slow getting back to this. I do appreciate your reply, and it helps to see where you're coming from. I'll respond with my perspective.

The History:
Beyond mimicking Basketball, our neutral site regionals were created with one very specific goal in mind. Eliminate scenarios where one team's fanbase has the ability to take over the venue, and "carry" it's team to victory. This is what happened at Yost in 1998, when Michigan upset an excellent North Dakota team. Now it wasn't huge upset; Michigan had a fine squad. But North Dakota had the better resume that year. I was at Yost that night, and the Wolverine crowd was awesome. Like many, I believe that crowd was a key component in Michigan's victory.

Critics make the same claim about the 2003 Regional at Mariucci, on behalf of Ferris State. I was at that regional as well. The Gopher crowd was similarly awesome. I must say that my belief is that the 2003 Gophers would have beaten the 2003 Bulldogs most anywhere due to their depth. But the game likely would have been closer & more in doubt at a neutral site.

These are the "smoking guns" people point to. The belief being that name brand schools would "buy" regionals year after year, repeatedly gaining victories they didn't deserve. Highly relevant is that fact that both the '98 Wolverines & '03 Gophers won National Titles, greatly adding fuel to the critics' fire.

The situation from 2018 you're describing doesn't fit that picture. I doubt that BU & NU fans lend cheers to each in any meaningful way, so they would tend to cancel each other out in any effort to take over the building. I wasn't there, so I won't overreach on this example. But I haven't heard anything to suggest that crowd support played a huge role last year in Worcester.

Here's another historical example, again chosen because I personally attended: 2005 in Worcester. That was the year "North Dakota Won the Beanpot." Both BC & BU were in that tournament; UND beat both on the way to the title. Needless to say, the BC & BU fans didn't join ranks. Despite the fact that both schools were well represented, there was no takeover.

Direct Reply:
You seem to be saying that all of the #2s & #3s deserve some sort of proportional representation in the stands, based on their regular season resume. And that anything less qualifies as unfairness. Well, proportional representation is a nice thought. But the current system certainly wasn't designed to produce that result. In fact, I'd argue it's a mission impossible. The fact that some teams have more than their proportional share of fans, and others less, strikes me as inevitable in any four team tournament. Geography & Fanbase Size pretty much guarantee unequal results. I believe a large majority of the college hockey community would acknowledge that much. But building takeovers are another matter. They're preventable. But at what cost?

Why We Object To PC's Home Crowd Advantage:
The current format has largely eliminated takeovers. That's one of the main reasons the status quo has significant support. But it has also destroyed the Western Regionals as events worth attending for neutral fans. Yet we're told that eliminating building takeovers is so important that we just have to take the harsh medicine & keep the peace.

That's a tough one to swallow. And it becomes impossible to keep the peace when we see an Eastern program being given the opportunity to take over its regional venue, and the obvious hypocrisy is simply denied. It's nothing personal against PC. But placing the Friars at the Dunk crosses a line we've been told is non-negotiable. And that line has now been crossed multiple times to the benefit of the same program. Precisely what the current format was supposed to prevent.

Much has been made of the fact that PC has received this benefit as a #4 Seed. No doubt that has added fuel to the fire. But I honestly believe that the same objections would have been made had PC been seeded #2 or #3.

I've already agreed that placing PC at the Dunk in 2019 was the right thing for this year's committee to do -- to make sure that this year's tournament worked out financially. And quite honestly, exposing the flaws in the current system for all to see has its positive aspects. And flawed it is.

Trouble is, there's absolutely no consensus as to what a replacement format would look like. And so we continue to muddle through with the status quo.

dxmnkd316
04-07-2019, 10:19 AM
Thanks. The RPI (school) website has a ranking called KASA, which means KRACH Adjusted for Site Advantage. It's the same math, with the slight change that there is a constant nation-wide advantage to playing at home. Whatever that advantage is can also be determined purely mathematically in the same way as KRACH does now. This makes it slightly more robust, in my eyes.

I agree that PWR, with its possible dependence on the very small data set of H2H games and CommOpp games, is poorer as a result of those things.

I also think that KASA or KRACH have zero chance of ever being adopted. In reality, the present PWR math is very complicated, but it seems conceptually easy. KRACH or KASA is conceptually difficult, but not as complicated. For example. I have a very simple EXCEL spreadsheet for the NHL, which, if kept updated with games played and results, can be used to calculate the Bradley-Terry Rankings for all teams in about 3 minutes, manually doing the iterations. I could write a simple VBA which would automate, that, I think. And, It might take me an hour to write the code. It's probably quicker than calculating PWR, simply because it consists of simply adding a bunch of fractions. But, it's conceptually difficult, so it won't be adopted.

I personally think that the math guys at the NCAA have micromanaged their RPI definition to try to make it mimic the KRACH results. I say that because the RPI is a very complicated beast right now. Differing weights for home and away games. 21/25/54...where does that come from? Nowhere, really. It's just an empirical number. And, the quality win bonus is very bogus. It's all clearly a magic formula. So, I'm okay with it. It's close to KRACH. And, I think that's the best we get.

But, that doesn't really answer this question: I would consider KASA or KRACH to be NOT predictive formulas. They merely postdict the season. You may say they give you odds on the next game in a way, but I don't really think they are robust enough to be considered predictive.

The discussion here led me to believe that you had something else in mind. Was I mistaken?

Thanks.

I have an automated spreadsheet for college hockey. That loop to automate should take about five minutes. Just have two columns, K and K0. Copy K0 into K and if you want to do it messy just loop it 30-50 times. You can also do adaptive looping so that it aggressively changes in the first few iterations and then fine tunes later.

mookie1995
04-07-2019, 10:19 AM
2 points.
Yost and mariuci and home campus sites. These were taken off the table. If the gophers or Michigan want to offer up exel or joe Louis, thatís fine.

Pc is NOT the host, but get the benefit. Hosting takes effort and commitment and thus, a reward.