FanLax Forum Poll

[Matnum PI]

1-9 team is #10 because they beat the #11 team. How did you determine the #11 team, and the 1-9 teams in order to make that team #10? This logic only works in a world where there are 11 teams, they've all played each other, and the #11 team is 0-10.

Agreed.

[FanLax Computer]

Just updated the "Anchor Rankings" with the benefit of the recent Forum Poll Rankings and these are the new Rankings:

Ranking, Team, Win Percentage, Win Rank, Loss Rank

1 Maryland 100% 1 1
2 Princeton 80% 2 4
3 Cornell 91% 3 5
4 Georgetown 90% 5 2
5 Rutgers 83% 7 3
6 Virginia 80% 7 8
7 Penn 50% 7 11
8 Ohio State 73% 6 13
9 Yale 78% 4 18
10 Harvard 78% 17 6
11 North Carolina 64% 12 12
12 Denver 55% 13 14
13 Richmond 60% 15 14
14 Notre Dame 56% 20 7
15 Army 82% 22 10
16 Jacksonville 83% 10 22
17 Brown 60% 16 24
18 Boston University 73% 28 9
19 Duke 64% 11 26
20 UMass 60% 19 21
21 Utah 67% 18 25
22 Lehigh 70% 24 19
23 Johns Hopkins 42% 13 34
24 Villanova 60% 30 20
25 Towson 45% 26 28

[ICGrad]

They do in my poll, (maybe not quite as much as wins)when comparing to other teams losses, especially if it's a close loss to a highly ranked team. That's why I dinged VA. Quite a bit after the MD. Loss as if they had even kept it in the Princeton loss to Md. Range, I would have been less penal, but to lose by that amount, even though MD. Is great, it was just too much of a disparity in my opinion.
GOBIGRED
Joewillie78

Not the same thing; you're saying you give teams more credit for close losses than for blowout losses, and for close losses to really good teams than for close losses to terrible teams. I'm on board with that.

The poll in question has a 9-5 Duke team 4 places higher than a 10-1 Cornell team, which makes no sense, especially when you consider the inputs: SOS, RPI, etc. Cornell leads Duke in every metric, has a significantly better record against a harder schedule, and they're 4 slots below them?

Now, if Cornell had a much better record but against significantly weaker competition, then maybe there's an argument to be made, but according to all the data points supplied by the linked poll, that's not the case.

I'm no statistician or quant (though I am a programmer), but if my computer model spit out something as nonsensical as that, I'd be ripping the lid off and trying to figure out what I did wrong.

[Matnum PI]

And the "anchor rankings" are last years results, right? So in your methodology, last year is more important than this year. It works as a starting point, but at this point in the season, its ludicrous.

Preseason, it's last years results. After the new season starts, the Anchor Rankings are based 100% on the current season.

[Matnum PI]

Cornell has lost one (relatively) close game; do close losses count more than wins?

Not necessarily. Close losses are treated as, so to speak, a team being just a little worse than another. If that make sense...

[CU77]

Well Dr.F's score-based power rating doesn't much like UNC, has them at #27:
http://college.laxpower2.com/menx/rating01.php

Though that ranking isn't without its issues as well.

For example, I can't for the life of me figure out how Duke is ranked 4 places higher than Cornell (7 v. 11) given Cornell's higher ranking in RPI (6/16), SOS (18/20) and QWF (3/30). Cornell is 10-1 against a stronger schedule and with a significantly higher RPI compared to a 9-5 Duke, and is 4 places lower than them...Huh?!?

The laxpower ranking is based on goal differential in all games played, with a max of 10. This is a VERY different metric than RPI (and SOS and QWF), which uses wins-and-losses only. UNC has gotten blown out a couple of times, and that pushes down their laxpower ranking.

There is of course no perfect formula.

[Gobigred]

Just updated the "Anchor Rankings" with the benefit of the recent Forum Poll Rankings and these are the new Rankings:

Ranking, Team, Win Percentage, Win Rank, Loss Rank

1 Maryland 100% 1 1
2 Princeton 80% 2 4
3 Cornell 91% 3 5

Princeton lost to Yale, ranked lower than Penn by you; Penn was Cornell's only loss. Why, then, does Princeton have a better "Loss Rank?" And don't tell me it's because they also lost to Maryland.

[Matnum PI]

There is of course no perfect formula.

There's no perfect formula because how we define the "best team" is subjective. So, yes, while this is true, there is a subjectivity to rankings, i do believe that some rankings objectively don't make sense. Said more accurately, the formula behind a ranking does not make intuitive, logical sense. Cornell at #11 is one of several excellent questions.

1 Maryland 99.90 1 10 1 10- 0- 0 67.2 100.0 40.1
2 Georgetown 96.72 3 12 4 9- 1- 0 65.6 99.5 10.7
3 Princeton 95.99 2 2 2 8- 2- 0 40.2 100.0 15.6
4 Notre Dame 95.75 13 13 15 5- 4- 0 0.0 57.2 5.8
5 Virginia 94.92 5 9 9 8- 2- 0 0.0 71.6 3.3
6 Yale 93.85 4 6 7 7- 2- 0 21.4 97.4 3.2
7 Duke 93.66 16 20 30 9- 5- 0 0.0 12.3 0.3
8 Rutgers 93.49 7 7 5 10- 2- 0 16.6 92.2 5.1
9 Penn 92.45 8 1 14 4- 4- 0 4.4 79.1 2.0
10 Jacksonville 92.37 19 29 11 10- 2- 0 48.0 48.0 0.7
11 Cornell 92.01 6 18 3 10- 1- 0 18.5 99.7 3.6
12 OSU 91.98 9 14 6 8- 3- 0 10.6 81.8 3.7
13 Richmond 91.52 15 17 19 6- 4- 0 24.8 25.0 0.4
14 Boston U 91.17 14 15 13 8- 3- 0 35.1 60.2 1.3
15 Michigan 90.87 45 51 29 7- 5- 0 1.1 1.1 0.1
16 High Point 90.38 27 23 25 6- 5- 0 24.7 24.7 0.2
17 Utah 90.32 22 32 22 6- 3- 0 43.6 43.6 0.2
18 Harvard 90.30 10 25 8 7- 2- 0 11.0 85.6 0.8
19 Denver 90.25 11 3 24 6- 5- 0 13.1 20.1 0.0
20 Army 90.18 24 46 10 9- 2- 0 23.8 32.4 0.5
21 Loyola 90.11 25 22 28 5- 5- 0 21.6 25.4 0.5
22 Villanova 89.88 26 30 27 6- 4- 0 13.5 13.5 0.4
23 Vermont 89.74 52 56 50 6- 6- 0 43.7 43.7 0.2
24 Syracuse 89.59 31 8 36 4- 7- 0 0.0 0.2 0.0
25 JHU 89.58 20 5 35 5- 7- 0 2.8 2.8 0.0

[Matnum PI]

Just updated the "Anchor Rankings" with the benefit of the recent Forum Poll Rankings and these are the new Rankings:

Ranking, Team, Win Percentage, Win Rank, Loss Rank

1 Maryland 100% 1 1
2 Princeton 80% 2 4
3 Cornell 91% 3 5

Princeton lost to Yale, ranked lower than Penn by you; Penn was Cornell's only loss. Why, then, does Princeton have a better "Loss Rank?" And don't tell me it's because they also lost to Maryland.

Three Best Wins, Three Worst Losses. And the Best Wins and Worst Losses are based on the Anchor Rankings. In this case, the Anchor Rankings are culled from the FanLax Forum Poll. With this in mind, P'ton's Three Worst losses are: Yale (4), Maryland(1), and 0. Total=5. Cornell's Three Worst Losses are: Penn (12), 0, and 0. Total=12. 5 beats 12.

[rolldodge]

Just updated the "Anchor Rankings" with the benefit of the recent Forum Poll Rankings and these are the new Rankings:

Ranking, Team, Win Percentage, Win Rank, Loss Rank

1 Maryland 100% 1 1
2 Princeton 80% 2 4
3 Cornell 91% 3 5
4 Georgetown 90% 5 2
5 Rutgers 83% 7 3
6 Virginia 80% 7 8
7 Penn 50% 7 11
8 Ohio State 73% 6 13
9 Yale 78% 4 18
10 Harvard 78% 17 6
11 North Carolina 64% 12 12
12 Denver 55% 13 14
13 Richmond 60% 15 14
14 Notre Dame 56% 20 7
15 Army 82% 22 10
16 Jacksonville 83% 10 22
17 Brown 60% 16 24
18 Boston University 73% 28 9
19 Duke 64% 11 26
20 UMass 60% 19 21
21 Utah 67% 18 25
22 Lehigh 70% 24 19
23 Johns Hopkins 42% 13 34
24 Villanova 60% 30 20
25 Towson 45% 26 28

This is a much more defensible final poll, IMHO. Some outliers, such as Yale at 9, but that might be a reasonable ranking given their loss to Penn State. Denver at 12 might seem high, but these are two examples of how hype creeps into polls: Yale is overvalued, and Denver is undervalued in most polls out there.

[rolldodge]

There is of course no perfect formula.

There's no perfect formula because how we define the "best team" is subjective. So, yes, while this is true, there is a subjectivity to rankings, i do believe that some rankings objectively don't make sense. Said more accurately, the formula behind a ranking does not make intuitive, logical sense. Cornell at #11 is one of several excellent questions.

1 Maryland 99.90 1 10 1 10- 0- 0 67.2 100.0 40.1
2 Georgetown 96.72 3 12 4 9- 1- 0 65.6 99.5 10.7
3 Princeton 95.99 2 2 2 8- 2- 0 40.2 100.0 15.6
4 Notre Dame 95.75 13 13 15 5- 4- 0 0.0 57.2 5.8
5 Virginia 94.92 5 9 9 8- 2- 0 0.0 71.6 3.3
6 Yale 93.85 4 6 7 7- 2- 0 21.4 97.4 3.2
7 Duke 93.66 16 20 30 9- 5- 0 0.0 12.3 0.3
8 Rutgers 93.49 7 7 5 10- 2- 0 16.6 92.2 5.1
9 Penn 92.45 8 1 14 4- 4- 0 4.4 79.1 2.0
10 Jacksonville 92.37 19 29 11 10- 2- 0 48.0 48.0 0.7
11 Cornell 92.01 6 18 3 10- 1- 0 18.5 99.7 3.6
12 OSU 91.98 9 14 6 8- 3- 0 10.6 81.8 3.7
13 Richmond 91.52 15 17 19 6- 4- 0 24.8 25.0 0.4
14 Boston U 91.17 14 15 13 8- 3- 0 35.1 60.2 1.3
15 Michigan 90.87 45 51 29 7- 5- 0 1.1 1.1 0.1
16 High Point 90.38 27 23 25 6- 5- 0 24.7 24.7 0.2
17 Utah 90.32 22 32 22 6- 3- 0 43.6 43.6 0.2
18 Harvard 90.30 10 25 8 7- 2- 0 11.0 85.6 0.8
19 Denver 90.25 11 3 24 6- 5- 0 13.1 20.1 0.0
20 Army 90.18 24 46 10 9- 2- 0 23.8 32.4 0.5
21 Loyola 90.11 25 22 28 5- 5- 0 21.6 25.4 0.5
22 Villanova 89.88 26 30 27 6- 4- 0 13.5 13.5 0.4
23 Vermont 89.74 52 56 50 6- 6- 0 43.7 43.7 0.2
24 Syracuse 89.59 31 8 36 4- 7- 0 0.0 0.2 0.0
25 JHU 89.58 20 5 35 5- 7- 0 2.8 2.8 0.0

Even while not perfect, what I like about the model based polls is that they often point out blind spots in the "eye test", and/or bias towards/against hype/media coverage. They are useful to look at in evaluating your own potential biases when making a "human" poll.

[Matnum PI]

Yale is overvalued, and Denver is undervalued in most polls out there.

I think most Rankings and Polls are trying to capture which teams are the best today and, for the sake of argument, the future. For example, as the season progresses, Rankings and Polls will heavily discount Yale's PSU loss. The FanLax Computer Ranking isn't trying to predict the future. FanLax Computer is assessing each team's season. In other words, there's no erasing the PSU loss. PSU may start winning games and then the loss isn't such a bad loss. But there's no discounting the loss.

[rolldodge]

Yale is overvalued, and Denver is undervalued in most polls out there.

I think most Rankings and Polls are trying to capture which teams are the best today and, for the sake of argument, the future. For example, as the season progresses, Rankings and Polls will heavily discount Yale's PSU loss. The FanLax Computer Ranking isn't trying to predict the future. FanLax Computer is assessing each team's season. In other words, there's no erasing the PSU loss. PSU may start winning games and then the loss isn't such a bad loss. But there's no discounting the loss.

Yes I agree. Yale may very well improve to the point that they are a NC contender, but the loss to Penn state and the last minute close wins do count for something. They may indicate the presence of some cracks in the armor that can be exploited. Those cracks may have been or will be patched up, but they were there. Their win against BU was impressive, but BUs recent schedule has shown them to be a tier below the top. And I say this as someone who was high on them earlier in the season.

Denver on the other hand, has very solid wins and their losses are respectable. It seems that pollsters knock them down for losing to Jacksonville while simultaneously elevating Jacksonville for winning against Duke. I find a lot of polls fall into this trap of evaluation teams inconsistently depending on opponents.

[Matnum PI]

Yale may very well improve to the point that they are a NC contender, but the loss to Penn state and the last minute close wins do count for something...

FWIW, there's a function within the FanLax Computer Ranking for play-off Ws and Ls where if a team beats a higher-ranked team in the play-offs (i.e. an upset), they automatically surpass them in the rankings. In other words, Yale's loss to PSU will most likely haunt them all season but, if they win the National C'ship or go to the Final Four or whatever, they can still be #1 or #3 or whatever in the nation.

Denver on the other hand, has very solid wins and there losses are respectable. It seems that pollster knock them down for losing to Jacksonville while simultaneously elevating Jacksonville for winning against Duke. I find a lot of polls fall into this trap of evaluation teams inconsistently depending on opponents.

What you're speaking to right now is the piece that probably interests me the most. People will see Jacksonville as really good or not-so-good depending on how J'ville helps or hurts their agenda. Same with 'cuse. But they beat 'cuse! in one sentence and then the same person will say, But 'cuse sucks this season! SU can't be both. I'm 100% fine with someone loving undefeated teams and making UVM the #2 team in the nation (theoretically) because at least the rationale makes sense. But when UVM is #2 and #12, that inconsistency shouldn't fly...

[Gobigred]

Just updated the "Anchor Rankings" with the benefit of the recent Forum Poll Rankings and these are the new Rankings:

Ranking, Team, Win Percentage, Win Rank, Loss Rank

1 Maryland 100% 1 1
2 Princeton 80% 2 4
3 Cornell 91% 3 5

Princeton lost to Yale, ranked lower than Penn by you; Penn was Cornell's only loss. Why, then, does Princeton have a better "Loss Rank?" And don't tell me it's because they also lost to Maryland.

Three Best Wins, Three Worst Losses. And the Best Wins and Worst Losses are based on the Anchor Rankings. In this case, the Anchor Rankings are culled from the FanLax Forum Poll. With this in mind, P'ton's Three Worst losses are: Yale (4), Maryland(1), and 0. Total=5. Cornell's Three Worst Losses are: Penn (12), 0, and 0. Total=12. 5 beats 12.

If you look at the rankings in your posting that I quoted, Penn is #7 and Yale is #9. Or do you have multiple rsnkings?

[rolldodge]

Yale may very well improve to the point that they are a NC contender, but the loss to Penn state and the last minute close wins do count for something...

FWIW, there's a function within the FanLax Computer Ranking for play-off Ws and Ls where if a team beats a higher-ranked team in the play-offs (i.e. an upset), they automatically surpass them in the rankings. In other words, Yale's loss to PSU will most likely haunt them all season but, if they win the National C'ship or go to the Final Four or whatever, they can still be #1 or #3 or whatever in the nation.

Denver on the other hand, has very solid wins and there losses are respectable. It seems that pollster knock them down for losing to Jacksonville while simultaneously elevating Jacksonville for winning against Duke. I find a lot of polls fall into this trap of evaluation teams inconsistently depending on opponents.

What you're speaking to right now is the piece that probably interests me the most. People will see Jacksonville as really good or not-so-good depending on how J'ville helps or hurts their agenda. Same with 'cuse. But they beat 'cuse! in one sentence and then the same person will say, But 'cuse sucks this season! SU can't be both. I'm 100% fine with someone loving undefeated teams and making UVM the #2 team in the nation (theoretically) because at least the rationale makes sense. But when UVM is #2 and #12, that inconsistency shouldn't fly...

Watch what happens if Duke wins big today over Virginia. It won't be "maybe I should reevaluate my view of Virginia by looking more closely at their resume, which maybe isn't as strong as I thought". Its going to be "Will Duke win the National Championship?!!!"

[Matnum PI]

If you look at the rankings in your posting that I quoted, Penn is #7 and Yale is #9. Or do you have multiple rsnkings?

There needs to be an Anchor Ranking or the cat will chase it's tail endlessly. Based on Penn being #7 and Yale being #9, Penn becomes #whatever and Yale become #whatever and... It never stops. There needs to be an anchor. For D1 Men's FanLax Computer Ranking, the Anchor Ranking is the FanLax Forum Poll.

[Matnum PI]

Watch what happens if Duke wins big today over Virginia. It won't be "maybe I should reevaluate my view of Virginia by looking more closely at their resume, which maybe isn't as strong as I thought". Its going to be "Will Duke win the National Championship?!!!"

8 minutes left, up by 9.

If Duke wins, according to the FanLax Computer, UVA moves from #6 to #9 and Duke moves from #19 to #17. This is UVA's worst loss and Duke's best win.

[rolldodge]

Watch what happens if Duke wins big today over Virginia. It won't be "maybe I should reevaluate my view of Virginia by looking more closely at their resume, which maybe isn't as strong as I thought". Its going to be "Will Duke win the National Championship?!!!"

8 minutes left, up by 9.

If Duke wins, according to the FanLax Computer, UVA moves from #6 to #9 and Duke moves from #19 to #17. This is UVA's worst loss and Duke's best win.

Reasonable.

[Matnum PI]