D3 Men Rankings

[admin]

Anyone know where you can find D3 Strength of Schedule stats and RPI ? Unfortunately the disparity in D3 makes it hard really dial in as well as D 1

PA, I almost forgot. D3 SOS

[richard]

[quote=richard post_id=40217 time=<a href="tel:1554916729">1554916729</a> user_id=418]
Maybe not within the parameters of the fanlax rankings but the SLU Clarkson game is a big one. Clarkson is better than the ranking it has. That and the rivalry with SLU make it a game to watch. SLU needs to get out of the gate today or else things could go up in smoke.

I agree, very strong rivalry between two North Country neighbors. I hope that you are right, but maybe for an outcome different than you desire.

Don't put the cold weather gear away...Game time Temp expected in the mid 20's.
[/quote]

SLU deserved to loose this game. Got cutsie with clock management with ten and change left in the fourth and took themselves out of the game. On the last play of the game, where was the SLU goalie? Headed for the sideline? What the heck?

[palaxoff]

by admin » Wed Apr 10, 2019 6:16 pm

palaxoff wrote: ↑Wed Apr 10, 2019 7:48 am
Anyone know where you can find D3 Strength of Schedule stats and RPI ? Unfortunately the disparity in D3 makes it hard really dial in as well as D 1
PA, I almost forgot. D3 SOS

Thx Admin, just goes to show you your job is impossible ranking these teams slow morning at work so did a little comparison of Top 10.

Team Rank SOS Record Against Top 10
Salisbury 1 36 12-0 2-0
RIT 2 20 11-1 1-1
York 3 35 11-1 2-1
Weseleyan 4 2 8-2 2-1
Amherst 5 13 10-1 2-1
Tufts 6 21 10-1 1-1
Cabrini 7 14 10-2 0-2
Gettysburg 8 10 9-3 1-2
Williams 9 1 8-2 1-1
Ursinus 10 53 9-1 0-1

[DeepPocket]

Thx Admin, just goes to show you your job is impossible ranking these teams slow morning at work so did a little comparison of Top 10.

Team Rank SOS Record Against Top 10
Salisbury 1 36 12-0 2-0
RIT 2 20 11-1 1-1
York 3 35 11-1 2-1
Weseleyan 4 2 8-2 2-1
Amherst 5 13 10-1 2-1
Tufts 6 21 10-1 1-1
Cabrini 7 14 10-2 0-2
Gettysburg 8 10 9-3 1-2
Williams 9 1 8-2 1-1
Ursinus 10 53 9-1 0-1

If that’s year to date, I would expect a significant rise for York and Salisbury after they play each other.... Same for the NESCAC cluster...

Similarly, a drop off will come for those whose conference competition is weak.

[admin]

It's not impossible. it's actually easy. You say, The best team in the nation is the team with the most wins! Period! and then some team in Iowa that we've never heard of is #1. The rank-er isn't wrong. Based on his criteria, this IA team is #1. But I think we'd all agree that his criteria is wrong, his criteria is too blunt a tool for a relatively nuanced question. And, also, creating the formulas to implement the metrics isn't so hard. What is hard(er), some would say impossible, is creating a viable metric that everyone agrees with. Anyway, that's an aside.

I dunno, Deep. Rank-ers love undefeated...

[DeepPocket]

I dunno, Deep. Rank-ers love undefeated...

Yes, I’m referring strictly to the SOS rankings where Salisbury and York are 35th and 36th. (Or vice versa)

[admin]

D3 MEN: A bunch of yesterday's game scores are missing but RIT beat Ithaca, Salisbury beat Christopher Newport, Gettysburg beat Dickinson, Colby beat Bowdoin, St. Lawrence beat Clarkson, York beat Frostburg, and... FanLax Computer Rankings updated.

[admin]

D3 MEN: Tufts beat Amherst (the 3rd biggest W this season), Ursinus beat Franklin and Marshall, Union beat St. Lawrence, Hampden-Sydney beat Roanoke, Williams beat Colby, Wesleyan beat Bowdoin, and... FanLax Computer Rankings and Game Scores updated.

[palaxoff]

Sorry and I don't want to diminish the fine job done by Ursinus but how do they get a # 6 ranking. I have seen them play, their ranking games (Stevenson and Christopher Newport and now ranked by you in low 40s) Their loss is to a team ranked 18.

[DeepPocket]

Sorry and I don't want to diminish the fine job done by Ursinus but how do they get a # 6 ranking. I have seen them play, their ranking games (Stevenson and Christopher Newport and now ranked by you in low 40s) Their loss is to a team ranked 18.

I may be wrong, but it’s my understanding that the rankings on this site are a straight mathematical equation, and the reason the teams you mentioned that are now in the 40s is precisely because of those losses. The winning team takes a percentage of the losing teams points at the time of the loss. So if someone beats a ranked team, they’re going to get a lot of points, and it doesn’t become diminished if the losing team drops off after that. Those points and the value of that win remain for the winning team to have or lose

[admin]

I see the issue (saw it with W&L. The number of losses, vs. the quality of losses, is creating an error. e.g. Stevenson has a bunch of Ls but to good teams and..) and... Need to fix and will post back in a second.

Re-Posted (with changes). I think this looks pretty good.

[palaxoff]

I may be wrong, but it’s my understanding that the rankings on this site are a straight mathematical equation, and the reason the teams you mentioned that are now in the 40s is precisely because of those losses. The winning team takes a percentage of the losing teams points at the time of the loss. So if someone beats a ranked team, they’re going to get a lot of points, and it doesn’t become diminished if the losing team drops off after that. Those points and the value of that win remain for the winning team to have or lose

Not to give you more work, but it be kind of interesting to see how the rankings would shake out with the final season rankings compared to the present system.

[admin]

Not to give you more work, but it be kind of interesting to see how the rankings would shake out with the final season rankings compared to the present system.

I'm not sure what you mean.

[Boomakegga]

It’s an interesting analysis and I value that you take the time to compute it. It seems to me, if I understand it correctly, that order in which teams play impact the total points. Thus, it might be a bit more credible if there was a +/- factor to which points were then rounded. This would account for the fact that beating team A before team A loses to team B is worth more than what team B received. This would create ties ultimately which could then be seen as groupings of similarly talented teams.

Put another way if teams A,B, and C all started with 1000 points and swapped 10% for a loss, And team A (1100) beat B (900), and B (1000) beat C (900) and C (1010) beat A ( 990) , the system ranks c, b then a, but it would be different if they played in a different order and we all would agree they are evenly matched at this point and should still be at 1000 points.

[admin]

Booma, first, it gives me faith in humanity that someone cares even a little bit about this stuff. With this said...

I used to do what you're saying. And could easily switch it back but... I don't like it. A good example, Albany starts the season with 980 Points. In the first game, SU beats them and gets 98 Points. then Cornell, then UMass, then etc. By the time UVM beats them they're only getting half the Points that SU got. And we could say, Albany devolved over the course of the season and Penn evolved over the course of the season and... It doesn't ring true. Plus, there's something off about 6 teams getting different Points after beating the same team. It's neat, it seems more alive with these moving parts but... I don't think it rings true. So, currently, everyone gets the same Points for beating Albany. SU gets 98. Then Cornell beats them and both SU and Cornell get 89 points. etc.

In terms of your second paragraph... Yes. The timing of games played far too large a role. Bottomline, there was too many ways to work the system. which is a nice way of saying, the model had flaws. So... We changed that.

And, with all that said... I'd be happy to try different ideas. I find it fun, interesting.

Oh, one more thing. Last year we did a straight + or - 10%. this season, this is not the case. i.e. The 1,000 becomes 900 wouldn't be the case. The wins are still 10% but for losses, it's different. When yale beats PSU, yale should pass them but... Why should PSU drop in any real way? Anyway...

Losses – Painting with broad brushstrokes, when Team A with 800 Points loses to Team B with 1,000 Points, they lose next to 0 Points. When Team A with 800 Points loses to Team B with 800 Points, they lose some, but very few, Points. When Team A with 800 Points loses to Team B with 600 Points, they lose a decent amount of Points, to Team B with 500 Points, they lose even more Points, etc.

and more Rankings info if you care.

[palaxoff]

Your discussion with Booma kind of addressed my question. I know there is not a perfect system but your's does a good job.

I think what I'd like to see at end of season is what the numbers would be based on a time line at end of year not based on at the moment ranking. for examples what would Ursinus rank be if the ranking of Stevenson, Cortland, Christopher Newport weren't were not so high.

[admin]

Agreed. There's no perfect system because... There's no such thing as a #14 team. Life isn't so simple. There is a #14 team based on winning percentage. #14 based on number of wins. #14 based based on proximity to the equator. But to say, definitively that Elmira is the #14 team in the country is silly. BUT... I think doing something like this for D3 is worth something (which we will, very soon). And we can add more metrics and rank-ers. (There's a post coming about this. More Rankers... ) We can play with RPI, we can do most anything... as long as it's built on Ws, Ls, and game scores. Speaking of which...

PA, we can totally do that. We could even do it now. It's a long and wide spreadsheet so you'd need to ask for specific teams but... totally do-able. and, my prediction... you'll love it. you'll totally enjoy watching teams ebb and wane and... You'll think it's not accurate. Fun but fun like a fiction novel is fun. and there's nothing wrong with that.

[palaxoff]

Admin I think I can wait till the end of the season. I appreciate your efforts and no matter how dialed in you get it someone isn't going to be happy. Thx.

[admin]

seriously. it'll give us something to do during the off-season.

[admin]

D3 MEN: York beat Salisbury (the #3 Big Win this season), Tufts beat Bates, Union beat RPI, Bridgewater beat Hampden-Sydney, Lynchburg beat Randolph-Macon, and the... FanLax Computer Rankings have been updated (and reconfigured).