MBB: North Coast Athletic Conference

Started by WoosterFAN, January 27, 2005, 10:51:56 AM

Previous topic - Next topic

0 Members and 2 Guests are viewing this topic.

Dave 'd-mac' McHugh

But the probability is based on flawed data.
Host of Hoopsville. USBWA Executive Board member. Broadcast Director for D3sports.com. Broadcaster for NCAA.com & several colleges. PA Announcer for Gophers & Brigade. Follow me on Twitter: @davemchugh or @d3hoopsville.

ziggy

#13351
Quote from: Dave 'd-mac' McHugh on February 06, 2013, 11:44:51 PM
But the probability is based on flawed data.

So what's the correct probability? There must be an answer to this question to make a claim like they "should be undefeated". ;)

Dave 'd-mac' McHugh

Maybe there isn't one for Division III. Maybe it is based on what people can deduce when they look at games online or in person plus box scores and overall stats for teams competing against one another plus factors like where the game is being played.
Host of Hoopsville. USBWA Executive Board member. Broadcast Director for D3sports.com. Broadcaster for NCAA.com & several colleges. PA Announcer for Gophers & Brigade. Follow me on Twitter: @davemchugh or @d3hoopsville.

ziggy

Quote from: Dave 'd-mac' McHugh on February 06, 2013, 11:53:46 PM
Maybe there isn't one for Division III. Maybe it is based on what people can deduce when they look at games online or in person plus box scores and overall stats for teams competing against one another plus factors like where the game is being played.

You're confusing win probability with picking winners. If I'm going through Calvin's in-region schedule I would pick them to win every game. I would do the same for Wooster. However, win probability acknowledges that while a team may be a favorite, even a heavy favorite, there is a chance that they lose. And even with just a small chance in each individual game, over the course of a 15 or 20-game schedule that adds up to the point where a loss actually is expected.

Good night!

Dave 'd-mac' McHugh

No I understand probability as a Management major and a number cruncher in my day... I am just saying that since there really isn't a system in place, the other idea is our best bet.
Host of Hoopsville. USBWA Executive Board member. Broadcast Director for D3sports.com. Broadcaster for NCAA.com & several colleges. PA Announcer for Gophers & Brigade. Follow me on Twitter: @davemchugh or @d3hoopsville.

smedindy

#13355
Exhibition games I'll give you, but I don't understand the reluctance to include other divisional games. Massey tracks the entire UNIVERSE of college hoops and excluding them would eliminate data points not just for D-3, but for ALL teams. And I think the exhibition games are because of the NCAA's totally screwed up awful rule about what counts and what doesn't. Don't hate the playa...

Box scores, etc. are all well and good but they give you no idea how to compare Birmingham Southern to Buena Vista. How can you compare Redlands to MIT without data and some baseline? The data ISN'T that flawed. It tracks pretty darn well with my perceptions, etc. Some teams still don't have video. So there's no WAY to see that. You just gonna, "oh, well" those teams?

Also, there are more human biases. People either overrate their region, or underrate them. A solid algorithm is region neutral. And remember, Massey deals with the entire UNIVERSE of college hoops.

And I'll give you a hint - Massey and Pomeroy DO factor in offense, defense and where the game is played. I think it's solid and reasonable, and that's what you need from a ratings system.

See how well the pundits did in November who were anti-data and probability model?



Wabash Always Fights!

Dave 'd-mac' McHugh

My problem is when a team plays a D1... then that D1's schedule is taken into account and all the data that in provides. That in turn boosts a team that is simply being paid to play the game and doesn't mean they are any better for it. It also sways the data for any team that played the team that played the D1. If you can isolate the D3 data to itself you can have a fair representation of the division. Furthermore, Massey seems to be in love with Wisconsin even though we know there is plenty of good basketball teams around the country, but Massey doesn't account accordingly.

So if everyone wants a probability system... build one :).
Host of Hoopsville. USBWA Executive Board member. Broadcast Director for D3sports.com. Broadcaster for NCAA.com & several colleges. PA Announcer for Gophers & Brigade. Follow me on Twitter: @davemchugh or @d3hoopsville.

sac

I can't imagine why the WIAC would rate so high every year.  Probably because they play good basketball?

National Champions
2012  UW-Whitewater
2010  UW-Stevens Point
2005  UW-Stevens Point
2004  UW-Stevens Point
1999  UW-Platteville
1998  UW-Platteville
1995  UW-Platteville

7 out of the last 17 seems like a good ratio to me.

Pat Coleman

Quote from: sac on February 07, 2013, 04:09:05 PM
I can't imagine why the WIAC would rate so high every year.  Probably because they play good basketball?

National Champions
2012  UW-Whitewater
2010  UW-Stevens Point
2005  UW-Stevens Point
2004  UW-Stevens Point
1999  UW-Platteville
1998  UW-Platteville
1995  UW-Platteville

7 out of the last 17 seems like a good ratio to me.

I don't mind the WIAC being rated highly, but it seems other teams get a big boost for touching the WIAC and I'm not sure that always makes a lot of sense. Maybe that part of the algorithm works for D1, but I'm not sure it does for D3. I made a point about the 2011 MIAC on the Pool C (I believe) board last week about this.
Publisher. Questions? Check our FAQ for D3f, D3h.
Quote from: old 40 on September 25, 2007, 08:23:57 PMLet's discuss (sports) in a positive way, sometimes kidding each other with no disrespect.

smedindy

Not necessarily, not at all. I think you're perception is skewed.

1. Massey doesn't INTENTIONALLY have a bias pro-Wisconsin. There's nothing in his algorithm that adds X points for a team from Wisconsin, especially at this stage. That's ludicrous as an argument and points to ludditism. The WIAC schedules are pretty robust in the non-conference. Massey's lowest rated WIAC team was at 1230 overall and at #185 in D-3. Can you say that UW-Oshkosh is clearly worse than Maryville, or Heidelberg or Olivet or Bowdoin? Those are teams in that neighborhood. The idea of a clear BIAS towards the WIAC is ridiculous. The data points it to Wisconsin because they do well in non-conference.

That's born out in the other rating systems. - Wolfe had the lowest WIAC team rated at #1110 out of #1553. They're around Allegheny and East Texas Baptist in that rating system, but that doesn't use MOV. Wilson also had the WIAC as the top conference over the UAA and CCIW. Funny, that's what Massey had too. The lowest WIAC team was #186 tied with Gettysburg, Clark and Lawrence. Can you say that UW-Oshkosh would get trounced by any of those teams.

Maybe the data DOES point to the WIAC being worthy of the top mark and there is NO bias. NONE.

2. If you look at the data, then Albertus Magnus was in no way rewarded for being thumped by Yale. It's one game in a data set and their SOS is still abysmal and their power ranking is #166 with an SOS of #380. Oberlin also played Yale and the NCAC got NO benefit from that.

So this argument is ludicrous as well.

Both of those arguments are flawed and faulty, based on the evidence I've seen. Show me the Wisconsin bonus in Massey, show me how Albertus and Oberlin (and others) were boosted by playing D-1s outside of the norm. Show me data and proof, not anecdotal yammering.
Wabash Always Fights!

smedindy

Quote from: Pat Coleman on February 07, 2013, 04:18:22 PM
Quote from: sac on February 07, 2013, 04:09:05 PM
I can't imagine why the WIAC would rate so high every year.  Probably because they play good basketball?

National Champions
2012  UW-Whitewater
2010  UW-Stevens Point
2005  UW-Stevens Point
2004  UW-Stevens Point
1999  UW-Platteville
1998  UW-Platteville
1995  UW-Platteville

7 out of the last 17 seems like a good ratio to me.

I don't mind the WIAC being rated highly, but it seems other teams get a big boost for touching the WIAC and I'm not sure that always makes a lot of sense. Maybe that part of the algorithm works for D1, but I'm not sure it does for D3. I made a point about the 2011 MIAC on the Pool C (I believe) board last week about this.

Pat,

If there was a bias in Massey it wouldn't show in the other systems. Those other systems also have the WIAC up and over other conferences. Where I can see it is that there is limited non-conference games in the MIAC and the proximity to the WIAC may help IF (and a big IF) they keep the games close against the WIAC (because MOV is a factor). Otherwise, I don't think touching the WIAC matters a ton otherwise Albertus Magnus would be a lot higher on the charts because they 'touched' a D-1 team.
Wabash Always Fights!

sac

Quote from: Pat Coleman on February 07, 2013, 04:18:22 PM
Quote from: sac on February 07, 2013, 04:09:05 PM
I can't imagine why the WIAC would rate so high every year.  Probably because they play good basketball?

National Champions
2012  UW-Whitewater
2010  UW-Stevens Point
2005  UW-Stevens Point
2004  UW-Stevens Point
1999  UW-Platteville
1998  UW-Platteville
1995  UW-Platteville

7 out of the last 17 seems like a good ratio to me.

I don't mind the WIAC being rated highly, but it seems other teams get a big boost for touching the WIAC and I'm not sure that always makes a lot of sense. Maybe that part of the algorithm works for D1, but I'm not sure it does for D3. I made a point about the 2011 MIAC on the Pool C (I believe) board last week about this.

I think it could be argued strongly the MIAC would have an outstanding tournament history if wasn't in the same region the WIAC. 


Dave 'd-mac' McHugh

OK... I said Wisconsin and that region... I did not say WIAC. Don't confuse the two.

The WIAC is a very strong conference and I say it year in and year out it is one of the top five in Division III basketball. However, I did not say that Massey weighs too heavily the WIAC... I said it is centralized towards Wisconsin and that region... and Pat has pointed out why it then skews other data and other teams - check out the MIAC board for Pat's thoughts on that.

Again... I said at the beginning that I have talked about this on numerous boards and because I have... please search Pat's and other comments out... I can't just keep rehashing things as much as I would like to boost my posts :). I have a show tonight and have lost the main broadcasting computer - I am a little stressed LOL.
Host of Hoopsville. USBWA Executive Board member. Broadcast Director for D3sports.com. Broadcaster for NCAA.com & several colleges. PA Announcer for Gophers & Brigade. Follow me on Twitter: @davemchugh or @d3hoopsville.

smedindy

#13363
How can data be centralized around a region? The data flows to where it flows. See my rebuttal. I think it's a perception, not reality.

And pardon me for not seeing your other posts because I would have rebutted the heck out of them then. Just because you didn't get a response doesn't mean your arguments are golden and can't be challenged NOW. I have posted back and forth with Pat on this in the past as well and I think I see the data differently.

And to be honest, when you say Wisconsin you ARE saying the WIAC. I don't think the NathCon or Carthage or the Midwest Conference matters a whit here.
Wabash Always Fights!

Dave 'd-mac' McHugh

Quote from: smedindy on February 07, 2013, 04:34:23 PM
How can data be centralized around a region? The data flows to where it flows. See my rebuttal. I think it's a perception, not reality.

And pardon me for not seeing your other posts because I would have rebutted the heck out of them then. Just because you didn't get a response doesn't mean your arguments are golden and can't be challenged NOW. I have posted back and forth with Pat on this in the past as well and I think I see the data differently.

Just because I don't get a response??? Seriously?! I never think my arguments are golden... and it is my opinion just as it is yours! But as Pat has shown in the past, when schools in the MIAC are boosted by Massey for no reason other than they are close to other schools like those in the WIAC it is flawed... and I don't trust it.

Seriously, if you can't come up with a system that actually shows the strengths of teams across the country that isn't saturated and infected with data that means nothing to a vast majority of the other schools being ranked... it is flawed. Transylvania has a game against Kentucky now yearly... you don't think their SOS number is clearly boosted by such a game? But does that say Transylvania's schedule is stronger than Wooster's just because Kentucky is in that data flow? No! Is Randolph-Macon's schedule boosted in Massey by their game against Richmond versus Hampden-Sydney's who doesn't have a D1 team to play? Of course!

Design a system like Massey that only uses games in Division III... because we can't get a sense of a school's SOS when we are comparing games and teams that have nothing to do with Division III. And if that data is used it unfairly boosts the SOS for many schools whose SOS isn't really that high to begin with.
Host of Hoopsville. USBWA Executive Board member. Broadcast Director for D3sports.com. Broadcaster for NCAA.com & several colleges. PA Announcer for Gophers & Brigade. Follow me on Twitter: @davemchugh or @d3hoopsville.