1/28/13 Poll Speculation

[Fishman'81]

Re: 1/28/13 Poll Speculation

If you take a little closer look you will notice that SUNYAC has played 34 of 77 games against ECAC West which greatly affects the overall record vs. other conferences. In turn, ECAC East has played 24 of 63 games against MASCAC, Northeast 10 and ECAC NE....going 19-4-1 against inferior conferences. Shouldn't you look at WHO you are playing as well? ECAC East has gotten fat on weaker conferences...period.
So yes, the stats are the stats and you can take what you want from them, but I'd be willing to bet these numbers would be much different if you had played 34 games against ECAC West (none by Plattsburgh) instead of the 4 they have played.

I agree completely... SOS is a key stat, especially in D-3, where there is very little parity among the various conferences.

 

[CARDS_rule_the_Burgh]

My apologies; no disrespect intended. I took a depiction of your putative rankings (posted by someone else) as accurate, but when I made a rare trip to d3hockey, I couldn't find your current rankings.

I will be interested to see them, when you make then available.

http://www.d3hockey.com/rankings/wiitala/2012-13/Men/feb-5

Published yesterday.

 

[Fishman'81]

Re: 1/28/13 Poll Speculation

Thanks for the link.

While it wasn't completely unreasonable, I still see no logical reason whatsoever to rank Norwich above either Oswego or Hobart.

The former's winning % is slightly better than OSU and HC, yes, but its SOS is far, far weaker than the latters'... Furthermore, the H2H thing doesn't favor NU in any way, shape or form: the lone common-opponent Norwich has in the mix is Neumann, and Norwich takes a back-seat in that comparison,too.

I have to wonder what mysterious factor pushes NU past the others, week after week after week in this just-for-fun metric... It's definitely the least intuitive of the three computed rankings we all look at, even when considering the TUC cliff of the PWR (which, while being stupid and arbitrary, is at least transparent.)

 

[NUProf]

Re: 1/28/13 Poll Speculation

Thanks for the link.

While it wasn't completely unreasonable, I still see no logical reason whatsoever to rank Norwich above either Oswego or Hobart.

The former's winning % is slightly better than OSU and HC, yes, but its SOS is far, far weaker than the latters'... Furthermore, the H2H thing doesn't favor NU in any way, shape or form: the lone common-opponent Norwich has in the mix is Neumann, and Norwich takes a back-seat in that comparison,too.

I have to wonder what mysterious factor pushes NU past the others, week after week after week in this just-for-fun metric... It's definitely the least intuitive of the three computed rankings we all look at, even when considering the TUC cliff of the PWR (which, while being stupid and arbitrary, is at least transparent.)

The algorithm ranks Norwich above Oswego is because almost all of Norwich's metrics are ahead of Oswego. In OWP Norwich Ranks 33rd and and Oswego ranks 41st. Norwich's OOWP ranks 7th, and Oswego's ranks 16th. The one metric in which Oswego beats Norwich is in OOOWP (25th to 33rd). Norwich's winning percentage is 1st in the nation and Oswego's is 4th. In terms of the winning percentage of the teams that they have defeated, Norwich ranks 12th, and Oswego ranks 26th.

As for Hobart, if you look at the bottom of the rankings, you will see a table that indicates a margin of error. If the difference in ratings of two teams is less than the margin of error, there is not a statistically significant difference in the rankings. This means that there is not enough evidence based on the data to rank either team ahead of the other. We have to put the teams in some order, but fact of the matter is that the measurements that we use don't clearly place Norwich ahead of Hobart.

Most computer ranking systems have a margin of error, but that margin of error is not usually acknowledged. The data clearly puts Norwich's accomplishments this season ahead of Oswego, but not significantly ahead of Eau Claire or Hobart. Both comparisons are within the margin of error.

Added: Probably the only thing that will satisfy you would have been for me to say that I just make the numbers up to pump Norwich's tires and I only put Utica first because of pressure from you - It isn't true, but it is clearly what you think

 

[Fishman'81]

Re: 1/28/13 Poll Speculation

Not going to argue stats with you; it's as irrelevant for me to do that as it is for you to cite your mysterious "metrics" all the time, without ever clarifying how you weight them... We are clearly not speaking in the same terms.

Referring to my above post, and considering once again to the SOS involved and the meager common-opponent data, your calculus just doesn't hold water, logically.

I'm not explicitly saying that it's rigged, but if that's not the case, it needs some adjustment. It just doesn't make much objective sense in its current form.

Norwich might well be the best team in all of D-3 this year, but there's no coherent argument for them being "ranked" where they are right now. There are clearly three teams with better cases to make, and you have two of them behind NU. So, I would be very interested to know what process you use to place Norwich ahead of OSU and Hobart. (That's a very fair question.)

And I certainly don't worry about where you put Utica; that doesn't matter a bit.

 

[NUProf]

Re: 1/28/13 Poll Speculation

Not going to argue stats with you; it's as irrelevant for me to do that as it is for you to cite your mysterious "metrics" all the time, without ever clarifying how you weight them... We are clearly not speaking in the same terms.

Referring to my above post, and considering once again to the SOS involved and the meager common-opponent data, your calculus just doesn't hold water, logically.

I'm not explicitly saying that it's rigged, but if that's not the case, it needs some adjustment. It just doesn't make much objective sense in its current form.

Norwich might well be the best team in all of D-3 this year, but there's no coherent argument for them being "ranked" where they are right now. There are clearly three teams with better cases to make, and you have two of them behind NU. So, I would be very interested to know what process you use to place Norwich ahead of OSU and Hobart. (That's a very fair question.)

And I certainly don't worry about where you put Utica; that doesn't matter a bit.

Did you even read the part about margin of error?
You claim that Oswego's SOS is better than Norwich, and when I advocate differently, you dismiss the use of SOS metrics.
I can/will discuss the issues of how the metrics are weighted and how they are/have been revised in the "home" of the rankings.

 

[jerrynu26]

Re: 1/28/13 Poll Speculation

Did you even read the part about margin of error?

Doubtful it was acknowledged. It's the "any ranking that doesn't put teams where I want them is trash" syndrome rearing it's ugly head again.

 

[Ozz]

Re: 1/28/13 Poll Speculation

Not going to argue stats with you; it's as irrelevant for me to do that as it is for you to cite your mysterious "metrics" all the time, without ever clarifying how you weight them... We are clearly not speaking in the same terms.

Referring to my above post, and considering once again to the SOS involved and the meager common-opponent data, your calculus just doesn't hold water, logically.

Hmmmm while it is close and Fishman's knowledge seems to far surpass any of our's, I am going to have side with NUProf. There's just something about having a Ph.D from Dartmouth and being a retired Mathematics professor that just seems to force me to side with him when it comes to numbers. Call me crazy!

 

[polarbear2]

Re: 1/28/13 Poll Speculation

Call me crazy!

You're Crazy

 

[Hammer]

You've lost your *ing mind.

 

[LakerEagleLover]

Re: 1/28/13 Poll Speculation

Hmmmm while it is close and Fishman's knowledge seems to far surpass any of our's, I am going to have side with NUProf. There's just something about having a Ph.D from Dartmouth and being a retired Mathematics professor that just seems to force me to side with him when it comes to numbers. Call me crazy!

That state school tuition wasn't wasted on you...

 

[97osu]

Re: 1/28/13 Poll Speculation

UMASS-Dartmouth?

 

[NUProf]

Re: 1/28/13 Poll Speculation

UMASS-Dartmouth?

...er no

And I shall exit the discussion altogether. I do get annoyed when my professional competence is challenged. I'm not saying my rankings are totally accurate, because that would be impossible. Any more discussion about them is welcome on the other web site. I promise that after the season is over, I will provide details and statistical analysis for anyone who is interested in learning more.

 

[Fishman'81]

Re: 1/28/13 Poll Speculation

I wasn't challenging your professional competence; I don't know the first thing about that.

My point was that OSU and NU have a nearly identical winning %, while OSU has beaten Hobart, a team your own metric ranks very highly, and NU has no win of similar quality... Furthermore, two of three OSU losses have come against the team your metric ranks first. (NU hasn't even faced any team that you rank nearly as highly, let alone defeated one of that ilk.) Pretty-much the same argument can be made for Hobart over Norwich, and clearly so.

I don't know what I ever said that makes you think that I discount SOS; that's the single most salient factor in D-3 rankings, IMO, because there's such a huge disparity among the various conferences in that regard. In fact, my essential problem with your rankings is that I don't regard NU's SOS as warranting where you rank them relative to HC and OSU. The latter two have both have played obviously more difficult schedules than the former, and have similar results in terms of W/L... May I suggest that you have under-weighted the SOS in your model?

(And sure, there's always a margin of error in these things, but yours looks way too big to me... That's what separates credible stats from less-credible ones, IMO. )

In closing, I'm no fan of either of either Oswego or Hobart, and I realize that your rankings are for entertainment-purposes only... But it's just nice to see credit given where it's due.

 

[CARDS_rule_the_Burgh]

Re: 1/28/13 Poll Speculation

I wasn't challenging your professional competence; I don't know the first thing about that...

My point was that OSU and NU have a nearly identical winning %, while OSU had beaten Hobart, a team your own metric ranks very highly, and NU has no win of similar quality... Furthermore, two of three OSU losses have come against the team your metric ranks first. (NU hasn't even faced any team that you rank nearly as highly, let alone beaten one.) Pretty-much the same argument can be made for Hobart over Norwich, and clearly so.

I don't know what I ever said that makes you think that I discount SOS; that's the single most salient factor in D-3 rankings, IMO, because there's such a huge disparity among the various conferences in that regard. In fact, my essential problem with your rankings is that I don't regard NU's SOS as warranting where you rank them relative to HC and OSU. The latter two have both have played obviously more difficult schedules than the former, and have similar results in terms of W/L... May I suggest that you have under-weighted the SOS in your model?

And what he had explained to you is that, while Oswego may have some nice big teams on their schedule, overall Norwich comes out ahead of Oswego in both OWP and OOWP... aka the components of SOS. In other words, in terms of raw number, Norwich WINS the SOS comparison with Oswego. Im sure he'll correct me if I'm wrong, but I do believe he has repeatedly stated that he calculates the OWP and OOWP by adding the total number of wins/losses by all opponents (excluding the games played with the team in question) into one large W-L-T record and determining the % of that record.

(And sure, there's always a margin of error in these things, but yours looks way too big to me... That's what separates credible stats from less-credible ones, IMO. )

I'm sure he didn't arbitrarily pick the margin of error. My guess is that the margin of error is a product of the low sample size (25 games isn't much, after all) and the inadequate crossover (i.e. not everybody plays everybody). Want a smaller margin of error? Play an 82 game schedule in which everybody plays everybody... oh wait... that's not DIII.

 

[NUProf]

Re: 1/28/13 Poll Speculation

And what he had explained to you is that, while Oswego may have some nice big teams on their schedule, overall Norwich comes out ahead of Oswego in both OWP and OOWP... aka the components of SOS. In other words, in terms of raw number, Norwich WINS the SOS comparison with Oswego. Im sure he'll correct me if I'm wrong, but I do believe he has repeatedly stated that he calculates the OWP and OOWP by adding the total number of wins/losses by all opponents (excluding the games played with the team in question) into one large W-L-T record and determining the % of that record.

I'm sure he didn't arbitrarily pick the margin of error. My guess is that the margin of error is a product of the low sample size (25 games isn't much, after all) and the inadequate crossover (i.e. not everybody plays everybody). Want a smaller margin of error? Play an 82 game schedule in which everybody plays everybody... oh wait... that's not DIII.

Thank you. There is a margin of error in the Bradley-Terry metric as well. They just don't tell you what it is. To me it is not intellectually honest to fail to report it.

 

[norm1909]

Re: 1/28/13 Poll Speculation

A public school teacher was arrested today at John F. Kennedy International airport this morning as he attempted to board a flight while in possession of a ruler, a protractor, a compass, a slide-rule and a calculator. At a press conference just before noon today, Attorney General Eric Holder said he believes the man is a member of the notorious Al-Gebra movement. Although he did not identify the man, he confirmed the man has been charged by the FBI with carrying weapons of math instruction.

‘Al-Gebra is a problem for us’, the Attorney General said. ‘They derive solutions by means and extremes, and sometimes go off on tangents in search of absolute values.’ They use secret code names like “X” and “Y” and refer to themselves as “unknowns” but we have determined that they belong to a common denominator of the axis of medieval with coordinates in every country. As the Greek philosopher Isosceles used to say, “There are 3 sides to every triangle.” The Attorney General went on to say “Teaching our children sentient thought processes and equipping them to solve problems is dangerous and puts our government at risk.”

 

[Birdwatcher]

Re: 1/28/13 Poll Speculation

A public school teacher was arrested today at John F. Kennedy International airport this morning as he attempted to board a flight while in possession of a ruler, a protractor, a compass, a slide-rule and a calculator. At a press conference just before noon today, Attorney General Eric Holder said he believes the man is a member of the notorious Al-Gebra movement. Although he did not identify the man, he confirmed the man has been charged by the FBI with carrying weapons of math instruction.

‘Al-Gebra is a problem for us’, the Attorney General said. ‘They derive solutions by means and extremes, and sometimes go off on tangents in search of absolute values.’ They use secret code names like “X” and “Y” and refer to themselves as “unknowns” but we have determined that they belong to a common denominator of the axis of medieval with coordinates in every country. As the Greek philosopher Isosceles used to say, “There are 3 sides to every triangle.” The Attorney General went on to say “Teaching our children sentient thought processes and equipping them to solve problems is dangerous and puts our government at risk.”

Aha.....That's the problem!

 

[joecct]

Re: 1/28/13 Poll Speculation

A public school teacher was arrested today at John F. Kennedy International airport this morning as he attempted to board a flight while in possession of a ruler, a protractor, a compass, a slide-rule and a calculator. At a press conference just before noon today, Attorney General Eric Holder said he believes the man is a member of the notorious Al-Gebra movement. Although he did not identify the man, he confirmed the man has been charged by the FBI with carrying weapons of math instruction.

‘Al-Gebra is a problem for us’, the Attorney General said. ‘They derive solutions by means and extremes, and sometimes go off on tangents in search of absolute values.’ They use secret code names like “X” and “Y” and refer to themselves as “unknowns” but we have determined that they belong to a common denominator of the axis of medieval with coordinates in every country. As the Greek philosopher Isosceles used to say, “There are 3 sides to every triangle.” The Attorney General went on to say “Teaching our children sentient thought processes and equipping them to solve problems is dangerous and puts our government at risk.”

Something does not add up in this story. I will check with a reporter I know, Cal Culus, who can root out the facts.

 

[Laker28]

Re: 1/28/13 Poll Speculation

Wow, did you come up with that all by yourself? It was completely lame and I like it.