Top 25 talk

[Hugenerd]

I'm not suggesting anyone use Massey in any way other than to suggest teams to look at. The way to answer this is not "I like Massey" or "I don't like Massey" but "Massey is wrong about team X for the following reason" or "Massey is right; team Y should be a little higher or lower in the poll."

This post I certainly agree with.  But that is not what you said (or at least strongly implied) in your original post.  You called the teams flat-out overrated or underrated, and said team x "should be" ....  To me, at least, you were directly asserting that Massey was right and the poll was wrong.

He did more than strongly imply, saying "(#X should be #Y)."  "Should be," in my opinion, means that ranking Y is correct (Massey) and ranking X (D3 poll) is wrong.  I am not sure what the intent of this wording was, but it came off very strong.

All that came after was subject to the end of the first paragraph, "according to Massey." My apologies if it wasn't clear enough that I was giving Massey's rating and not my own.

That was clear, but the usage of "should be" does imply the latter is correct.  If you were simply comparing you could have used "vs." or "compared to."

[John Gleich]

As for an example of why we shouldn't put too much stock into the polls?  DK, you mentioned Denison as one of Wooster's less than stellar opponents to date this season.  The same Dension that beat Chicago at Chicago.  That same Chicago that just beat previous #2 IWU in Bloomington.  Oh, and Wooster beat just Denison by 31...  

This is a great example of how polls can be deceiving and how match-ups can vary from team to team.  It's possible that a good team may have a particular weakness that can be exploited by a mediocre team or one kid may just have "one of those nights" that causes team #150 to beat team #3.  But that doesn't necessarily mean that team #150 should suddenly become team #3 or even that team #3 should automatically drop.  In a computer ranking, they would, but there's a human element to a poll like the D3hoops poll that will take into effect other things.

[Hugenerd]

As for an example of why we shouldn't put too much stock into the polls?  DK, you mentioned Denison as one of Wooster's less than stellar opponents to date this season.  The same Dension that beat Chicago at Chicago.  That same Chicago that just beat previous #2 IWU in Bloomington.  Oh, and Wooster beat just Denison by 31...  

This is a great example of how polls can be deceiving and how match-ups can vary from team to team.  It's possible that a good team may have a particular weakness that can be exploited by a mediocre team or one kid may just have "one of those nights" that causes team #150 to beat team #3.  But that doesn't necessarily mean that team #150 should suddenly become team #3 or even that team #3 should automatically drop.  In a computer ranking, they would, but there's a human element to a poll like the D3hoops poll that will take into effect other things.

Great point.  Conversely, team #3 could lose their best player to injury and therefore lose to team #150.  There are a lot of arguments for a human element.

The strength of computer-based polls is that they can pick up on undervalued teams that may be overlooked because they arent on the "radar" of the pollsters.

Also, someone previously stated that 6-8 games was a good sample size, which maybe true in some cases or even at the D1 level(although this is debatable). However, because of the regionality of D3, there are very little cross-over games in D3.  Therefore, comparisons between regions with the computer poll can be problematic because of the sample size of games between opponents in each region is small.  If you were selecting games randomly from all the teams in d3, 25 games would be more than enough to get good statistics and accurate rankings, but because of how regional this division is, especially in conferences where teams play 16 or even 18 or more conference games, you have only a handful of non-conference games, let alone out-of-region games.

[Hugenerd]

Wooster down by 2 to Allegheny, 1 minute into the second half.

[David Collinge]

The Wooster/Allegheny game is back-and-forth, now inside 8:00 to play with Wooster up by 2.  Audio.

[Hugenerd]

Live Stats

[KnightSlappy]

The different philosophies that people hold with respect to top 25 polls really shows through in the early going.

Some view it as "who is the better team." i.e. who would win the most games if every team could play everyone else 100 times. These pollsters tend to stick to their guns and won't move a team too much for a reasonable loss.

Some view it as a "resume poll". You get one shot each game. Simply what have you done so far? These pollsters like to move teams around a lot after each game.

By the end of the season, the resume is equal to team strength (more or less), so there is less debate between the two camps.

[David Collinge]

Ian Franks hits a buzzer-beating three to give Wooster the 76-73 victory.

[Mr. Ypsi]

The different philosophies that people hold with respect to top 25 polls really shows through in the early going.

Some view it as "who is the better team." i.e. who would win the most games if every team could play everyone else 100 times. These pollsters tend to stick to their guns and won't move a team too much for a reasonable loss.

Some view it as a "resume poll". You get one shot each game. Simply what have you done so far? These pollsters like to move teams around a lot after each game.

By the end of the season, the resume is equal to team strength (more or less), so there is less debate between the two camps.

I suspect most voters have some of each philosophy in them, though how much will obviously vary.  I don't really see two distinct 'camps'.

To me, the 'who is the better team' subjectivity is a plus of polls (if not held TOO stubbornly), even if it leads to occasional disasters like the year Oshkosh started #2 (and dropped entirely out of the poll by week 3 or 4, never to return), or the year Chicago began in the poll, though fairly low, and proceeded to begin the season something like 0-16!

Your point about Massey alerting voters to under-the-radar teams is a good one, but the poll already somewhat accomplishes that by having 3 voters from each region (plus Pat as the 25th vote).  A team may be under-the-radar nationally, but the voters within a region are likely to notice a team starting off 8-0 regardless of pedigree or quality of conference!

[wooscotsfan]

Final:  #1 Wooster 76  Allegheny 73

Wooster gets a narrow road win tonight as All American Ian Franks hit a three pointer to win it for the Scots at the buzzer.  For Wooster, Franks had 19 points, Justin Hallowell with 18 points, Matt Fegan with 15 points and Nathan Balch with 12 points.

[Dark Knight]

One thing that Massey, rightly or wrongly, does not take into consideration is projections for success as the year goes forward.  These computer models are based strictly on what has happened so far this season.  Wooster is not ranked #1 at this moment because of who they have beaten.  They are #1 because they finished last season with a very strong showing in the national tournament and returned almost everybody from that team, placing them in position to move up when the teams above them were knocked off.  This subjectivity is not something that computers can handle.  Whether they should or not, I leave to you to decide.  But a team like Wooster, toiling in a relatively weak conference, will generally suffer in computer polls when compared to a school like UW-SP with a strong conference, and there's not a whole heck of a lot they can do about it.

This is the right kind of response: that the pollsters are using some information not available to Massey, how Wooster did last year. However, it's not really true in this case -- Massey does use last year's results early in the season. By the end of the season the effects of the previous season's results are entirely gone.

Also, there are two kinds of computer ratings, differing mainly on whether they pay attention to margin of victory. The rating I was quoting does use MOV, so Wooster could be rated #1 simply by winning big against weaker teams. Other ratings, that don't use MOV, are more susceptible to the effect you mention that teams in weaker conferences suffer in the polls.

As an aside, both approaches have flaws. However, Massey has done extensive statistical analysis of the two kinds of ratings (MOV-based and not MOV-based). They are close in predictive power, but the MOV-based rating is slightly better.

But what do you all think of the Massey's implication that Whitworth's results so far are quite a bit stronger than Wooster's? (David's unanswered question about whether polls should reflect only the games played so far or also take into account the history of the programs, etc. remains, of course.)

[Hugenerd]

Milwaukee Engineering is 6-0 and hasnt received a vote yet! (too bad they lost to Aurora tonight)

[Dark Knight]

Final:  #1 Wooster 76  Allegheny 73

Wooster gets a narrow road win tonight as All American Ian Franks hit a three pointer to win it for the Scots at the buzzer.  For Wooster, Franks had 19 points, Justin Hallowell with 18 points, Matt Fegan with 15 points and Nathan Balch with 12 points.

Not a good result for Wooster's #1 ranking, as Allegheny is a below-average team in D3...

[John Gleich]

To me, the 'who is the better team' subjectivity is a plus of polls (if not held TOO stubbornly), even if it leads to occasional disasters like the year Oshkosh started #2 (and dropped entirely out of the poll by week 3 or 4, never to return), or the year Chicago began in the poll, though fairly low, and proceeded to begin the season something like 0-16!

Chicago went 0-12 I think.  And both were major anomalies... Oshkosh was really talented but underachieved a ton and Chicago lost lots of tough games and just lost their confidence and didn't believe they could do it.

[Gregory Sager]

Milwaukee Engineering is 6-0 and hasnt received a vote yet! (too bad they lost to Aurora tonight)

I was just about to post that you'd jinxed the Raiders, 'nerd.