2016 NCAA Tournament Thread

[PSUChamps2001]

Re: 2016 NCAA Tournament Thread

You're right as far as the numbers go, but if your (Plymouth's, in this case) SOS is that feeble, you need to win nearly every game you play in order to garner the statistical benefit of the doubt...

Had Plymouth won-out, there would be no logical reason to assume that they wouldn't have done as well against better competition. But they didn't win out, and they played a very weak schedule, so there you go. Every loss against a weak team will kill you, especially when you don't have a single quality win to offset it. (And, similarly, you'll note that the KRACH didn't penalize PSU much at all for losing one game against a very good opponent, because they had a bunch of quality wins that preceded that one game.)

Look, I'm not saying that any metric is even close to perfect... But the current system makes far, far much less sense than does the KRACH, clearly.

And that's exactly my point. Will teams replace their tough SOS schedule for weaker (but yet still good win %) teams to boast up their own win %? It's obvious that your SOS (70's) doesn't hurt you, so why wouldn't teams? I know you can not determine ahead of time how a team will do that year, record wise (see Oswego). But for years, certain teams/leagues would feast on the bottom teams in other leagues, which boosted their win% and in return booster their SOS because OOOP Win % isn't weighted heavy enough IMO.

I'm just confused. You've stated that teams like Plymouth State don't belong and "never came near the Top-11", but yet when pointed out that they did, all of a sudden they must be a worthy team because the KRACH says so.

And with all due respect, you claim PSU's loss to Geneseo didn't lose them much because they had a bunch of quality wins....playing Devil's advocate here, but where? Oswego? 31st in win %? 4 ECACNE teams had higher alone. Norwich? Didn't make their conference finals and beat them when they were missing their Top scorers. Colby? Fitchburg? Canton? Middlebury? Williams (whoops oh yea you said wins), Geneseo? Oh yea see previous. Thank god for Buff State....

I just think you would find too many teams trying to play a Plymouth State type schedule. What if Plymouth State beat Salem in the last game, and was ranked #10 in KRACH. What quality wins do they have? Yet loses to Curry, Stonehill and Salem St x3 (a 22 win team...how did that work out for them?). Manhatanville a few years ago did the same thing. Even Utica had such a **** poor NC schedule a few years back, and had to adapt. Brockport, Franklin Pierce x2, Westfield St, OSU, Amherst, Buff St, OSU, Buff St, Curry, Cortland, Suffolf, Wentworth...outside of OSU and Amherst ....

We've already had conferences adapt new rules to try and gain more National Tournament teams (see SUNYAC going to single game format and not best of 3). You darn well bet they will do the same to meet any "single" formulated system to put teams in the National tournament. SUNYAC will play an 8 game SUNYAC schedule and another 10 games against ECACNE, MASCAC teams, and another 5 against DII teams so they don't count anyways. Oswego, Geneseo, Plattsburgh all have maybe 1 or 2 loses all season and walk into the National Tournament. Yet Utica goes out, plays their ECAC W schedule and schedules Norwich x2, Williams x2, Trinity, Babson, and UMB ...loses 5 or 6 games and is #13 in the KRACH next year....all three of those SUNYAC schools are in playing a crap schedule, but yet Utica who challenged themselves is out, just because the numbers say so.

Again, I totally get what you're trying to say, and I'm on the list of people getting annoyed with the overall lack of improvement of the ECACNE and MASCAC league, but until things become a tad more "equal" talent wise across the board, a mathematical equation will not work, can be beat, and will be beat. Great in theory.

 

[Fishman'81]

Re: 2016 NCAA Tournament Thread

And that's exactly my point. Will teams replace their tough SOS schedule for weaker (but yet still good win %) teams to boast up their own win %? It's obvious that your SOS (70's) doesn't hurt you, so why wouldn't teams? I know you can not determine ahead of time how a team will do that year, record wise (see Oswego). But for years, certain teams/leagues would feast on the bottom teams in other leagues, which boosted their win% and in return booster their SOS because OOOP Win % isn't weighted heavy enough IMO.

I'm just confused. You've stated that teams like Plymouth State don't belong and "never came near the Top-11", but yet when pointed out that they did, all of a sudden they must be a worthy team because the KRACH says so.

And with all due respect, you claim PSU's loss to Geneseo didn't lose them much because they had a bunch of quality wins....playing Devil's advocate here, but where? Oswego? 31st in win %? 4 ECACNE teams had higher alone. Norwich? Didn't make their conference finals and beat them when they were missing their Top scorers. Colby? Fitchburg? Canton? Middlebury? Williams (whoops oh yea you said wins), Geneseo? Oh yea see previous. Thank god for Buff State....

I just think you would find too many teams trying to play a Plymouth State type schedule. What if Plymouth State beat Salem in the last game, and was ranked #10 in KRACH. What quality wins do they have? Yet loses to Curry, Stonehill and Salem St x3 (a 22 win team...how did that work out for them?). Manhatanville a few years ago did the same thing. Even Utica had such a **** poor NC schedule a few years back, and had to adapt. Brockport, Franklin Pierce x2, Westfield St, OSU, Amherst, Buff St, OSU, Buff St, Curry, Cortland, Suffolf, Wentworth...outside of OSU and Amherst ....

We've already had conferences adapt new rules to try and gain more National Tournament teams (see SUNYAC going to single game format and not best of 3). You darn well bet they will do the same to meet any "single" formulated system to put teams in the National tournament. SUNYAC will play an 8 game SUNYAC schedule and another 10 games against ECACNE, MASCAC teams, and another 5 against DII teams so they don't count anyways. Oswego, Geneseo, Plattsburgh all have maybe 1 or 2 loses all season and walk into the National Tournament. Yet Utica goes out, plays their ECAC W schedule and schedules Norwich x2, Williams x2, Trinity, Babson, and UMB ...loses 5 or 6 games and is #13 in the KRACH next year....all three of those SUNYAC schools are in playing a crap schedule, but yet Utica who challenged themselves is out, just because the numbers say so.

Again, I totally get what you're trying to say, and I'm on the list of people getting annoyed with the overall lack of improvement of the ECACNE and MASCAC league, but until things become a tad more "equal" talent wise across the board, a mathematical equation will not work, can be beat, and will be beat. Great in theory.

Good Lord... OK, let's try this again.

A. You cannot bolster your SOS by beating bad teams. All you can hope for, if you schedule bad teams, is to win every game, because any decent statistical model will be forced to assume that being undefeated, even against a weak SOS, cannot preclude that you couldn't have done as well against a strong SOS... But even a couple of losses v. the bad teams on your schedule will kill you in that scenario. (Thus, Plymouth State having no shot at a Pool C this year.)

B. To review yet again, neither SOS nor W% stands alone. They must be considered side-by-side in order for either to possess any meaning. Please figure this out.

C. There is no way to massage the KRACH. You either need to schedule a bunch of good teams and win most of your games, or schedule cupcakes and win essentially ALL of your games. This year's KRACH is another example of that. Why do imagine that Platty was so much better in that metric than was Plymouth? Fancy math?

 

[NUProf]

Re: 2016 NCAA Tournament Thread

As I have said a number of times, KRACH (actually Bradley - Terry adapted to hockey, some guy name "Ken" chose to name it after himself) is metric based on logistic regression. As such it has a margin of error which is never included, but unless the level of saturation is high the margin of error will be high. Since there are far more D3 teams than there are games played per team, the level of connection in the competition graph (maximum distance between nodes) is way too high. There are assumptions built into the model which may or may not valid. It would be far superior to develop a tool that measures results relative to quality of competition and apply it to the teams that don't qualify for the AQ bids. Ideally this metric would allow leagues to return to a best of three or at least two game + minigame series as part of a conference tournament. If every league played multigame series in their league playoffs, the survivor would be a more likely to represent the best that each conference has to offer. The fact that the SUNYAC bailed on that format is sad - especially since it didn't even help Plattsburgh to get to the tournament anyway.

Take your league champions - take the top of the teams as determined by a reasonable metric to fill out the field- do take the committee out of it. Those are the teams that can compete. You have two routes to the tournament. 1. Win your league in a playoff format that is unlikely to include fluke results. 2. Demonstrate excellence in an objective manner.

I get really frustrated by the bloated DI basketball bracket that allow 7 teams from one conference to participate - let's see, you are the 7th best team in your league, so you should be considered for a chance at the national title - that defies logic, but a subjective committee continues to bypass teams with good records in minor (so called "mid-major") in preference for teams that were in the bottom half of major conferences. A limit on the number of teams per conference would make the regular season mean a lot more.

 

[hawk]

Re: 2016 NCAA Tournament Thread

http://www.bostonherald.com/sports/...ling_comeback_ends_with_ot_loss_in_semifinals

Article in the Boston Herald on the UMB/SNC game. I was very impressed how UMB marched through the NCAA's and just came up short against the top team in the country. The NEHC is sure going to competitive next year with UMB, Babson, NU and maybe St. A's after they file there paperwork to go D3. Who knows maybe Castleton and SM will step it up too, Is Peter Besile up for coach of the year Prez.

 

[spwood]

Re: 2016 NCAA Tournament Thread

Good Lord... OK, let's try this again.

A. You cannot bolster your SOS by beating bad teams. All you can hope for, if you schedule bad teams, is to win every game, because any decent statistical model will be forced to assume that being undefeated, even against a weak SOS, cannot preclude that you couldn't have done as well against a strong SOS... But even a couple of losses v. the bad teams on your schedule will kill you in that scenario. (Thus, Plymouth State having no shot at a Pool C this year.)

A better example was in Div III basketball this year. Lancaster Bible was undefeated going into its conference tourney. However, they knew they were going to need to win the tourney to get in the NCAAs. They were never in the NCAA Regional Rankings. Their SOS was abysmal. Worse than bad. They did win the conference tourney. And then lost a first round NCAA game by double digits.

 

[PrezdeJohnson09]

http://www.bostonherald.com/sports/...ling_comeback_ends_with_ot_loss_in_semifinals

Article in the Boston Herald on the UMB/SNC game. I was very impressed how UMB marched through the NCAA's and just came up short against the top team in the country. The NEHC is sure going to competitive next year with UMB, Babson, NU and maybe St. A's after they file there paperwork to go D3. Who knows maybe Castleton and SM will step it up too, Is Peter Besile up for coach of the year Prez.

He won National Coach of the Year along with Chris Schultz at Geneseo. Co-winners.

 

[Russell Jaslow]

Ever hear of the concept of sample-size..? Guess not, since you and many others apparently, um, "think" that a single game -which just happens to be played at the very end of the season- trumps everything that happened during 25+ RS games. (And even if the PS tournament in a particular conference doesn't include every single team under that umbrella, it certainly includes some teams that don't deserve National consideration.)

Talk about an utter dearth of logic. You really ought to work on that GED, Russ.

Geneseo finished third in the RS. Don't give a crap where they were in the SOS, KRACH, or regional rankings. All I know is they made it to Lake Placid. And they fully deserved to. So, you can take all your logic, flush it down the toilet, and stop boring the sh*t out of us.

I'll be watching a great hockey game tonight and no longer care, like I ever did, about anything you say or think.

Good night.

 

[Fishman'81]

Re: 2016 NCAA Tournament Thread

You ever hear of an a-s-s-hole? You're the definition of one.

BTW, Geneseo finished third in the RS. Don't give a crap where they were in the SOS, KRACH, or regional rankings. All I know is they made it to Lake Placid. And they fully deserved to. So, you can take all your logic, flush it down the toilet, and stop boring the sh*t out of us.

I'll be watching a great hockey game tonight and no longer care, like I ever did, about anything you say or think.

Good night.

Wow. So hostile! You drink much?

Russ, if you need a hand, I'd be happy to call 911 to summon the Jaws of Life and get those panties out of your nether-regions. My love of humanity -even for those who haven't a clue- is boundless.

Not to be an AH yet again, but it's no wonder that uscho couldn't use you anymore. You always had an axe to grind and were never even remotely objective...

Good night to you, too!

 

[Fishman'81]

Re: 2016 NCAA Tournament Thread

As I have said a number of times, KRACH (actually Bradley - Terry adapted to hockey, some guy name "Ken" chose to name it after himself) is metric based on logistic regression. As such it has a margin of error which is never included, but unless the level of saturation is high the margin of error will be high. Since there are far more D3 teams than there are games played per team, the level of connection in the competition graph (maximum distance between nodes) is way too high. There are assumptions built into the model which may or may not valid. It would be far superior to develop a tool that measures results relative to quality of competition and apply it to the teams that don't qualify for the AQ bids. Ideally this metric would allow leagues to return to a best of three or at least two game + minigame series as part of a conference tournament. If every league played multigame series in their league playoffs, the survivor would be a more likely to represent the best that each conference has to offer. The fact that the SUNYAC bailed on that format is sad - especially since it didn't even help Plattsburgh to get to the tournament anyway.

Take your league champions - take the top of the teams as determined by a reasonable metric to fill out the field- do take the committee out of it. Those are the teams that can compete. You have two routes to the tournament. 1. Win your league in a playoff format that is unlikely to include fluke results. 2. Demonstrate excellence in an objective manner.

I get really frustrated by the bloated DI basketball bracket that allow 7 teams from one conference to participate - let's see, you are the 7th best team in your league, so you should be considered for a chance at the national title - that defies logic, but a subjective committee continues to bypass teams with good records in minor (so called "mid-major") in preference for teams that were in the bottom half of major conferences. A limit on the number of teams per conference would make the regular season mean a lot more.

Really thoughtful post. A good read.

The only point you made that I might quibble with is the reference to too many teams from any given conference making the field. (That defies logic, IMO.) If you have a statistical advantage v. any given team, forget conference-affiliations.. They are essentially meaningless in a stand-alone sense.

 

[NUProf]

Re: 2016 NCAA Tournament Thread

Really thoughtful post. A good read.

The only point you made that I might quibble with is the reference to too many teams from any given conference making the field. (That defies logic, IMO.) If you have a statistical advantage v. any given team, forget conference-affiliations.. They are essentially meaningless in a stand-alone sense.

The issue is that the NCAA mandates the number of bids and the number of Pool B/C slots based on the 6.5:1 ratio. It is the same rule in every sport. They also allow any league that has 7 or more members that have played together for at least two years to have a Pool A slot (assigned at the discretion of the conference). If the ratio would change (and it won't, because the NCAA allocates 3.4% of its budget to DIII, despite the fact that more than DIII has more programs and more athletes than either of the other two divisions) there would be more Pool C slots allocated, but the financial reality is that D3 can't change the ratio, because they can just barely fund the championships as they are currently structured.

 

[Russell Jaslow]

Not to be an AH yet again, but it's no wonder that uscho couldn't use you anymore. You always had an axe to grind and were never even remotely objective...

Odd then that USCHO sent me to Lake Placid...

And me no longer doing the SUNYAC column was my choice, not USCHO's.

But, I never expected you to ever get any facts correct, the ultimate axe grinder.

Point out one article I wrote for USCHO that was not objective. Just one. I didn't think so...

 

[Fishman'81]

Re: 2016 NCAA Tournament Thread

No hard feelings. Enjoy your off-season.

 

[Tom Coughlin]

Re: 2016 NCAA Tournament Thread

Anybody know where one could buy a D3 Frozen Four shirt?