Fosco's Team Rankings

[Fosco33]

Alright everyone - we've had enough debate on TT's crazy moves, the Oline's lack of protection and #4.

For the experienced posters - this will look familiar.

I take key stats on offense, defense and rank them 1-32. I then apply a formula based on the individual stat rankings to figure out the top overall statistical teams.

Yeah, yeah - I know. Stats don't mean crap. But results do. I've won spread pools of 45+ guys with $100 buy-ins and am currently at 60% against the spread (also #5 in the pick 'em pool here). Over the last 5 years, I've been able to predict almost 80% of the playoff teams by week 10 using this system.

Here's the overall ranking:

http://www.hostpic.biz/uploads/7dd420c1ca.jpg (http://www.hostpic.biz)

Interesting items:
Colts, Niners and Bengals are overperforming (winning more than they should).

Ravens and Dolphins are underperforming (winning less than they should).

All other teams are within +/- 5 ranks of their actual winning percentage ranking (performing about as expected).

* there's hope yet for the Packers at #6 overall *


Offensive breakdown:

http://www.hostpic.biz/uploads/5bd8c49dfb.jpg (http://www.hostpic.biz)

Defensive breakdown

http://www.hostpic.biz/uploads/7001b33aef.jpg (http://www.hostpic.biz)

Time of Possession and Turnover Margin

http://www.hostpic.biz/uploads/5c3a6fa0db.jpg (http://www.hostpic.biz)

Points Scored and Allowed

http://www.hostpic.biz/uploads/0d062cece9.jpg (http://www.hostpic.biz)

[Partial]

I thoroughly enjoyed your power rankings. If I remember correctly they were quite accurate. Good to see more content.

[Guiness]

Indi seem out of place in the rankings Fosco - shouldn't they be first on the offensive chart? Maybe there are more columns you're not showing us.

[Fosco33]

Indi seem out of place in the rankings Fosco - shouldn't they be first on the offensive chart? Maybe there are more columns you're not showing us.

Indy is # 2 in a few offensive categories: (6.7 yards/play; Cowboys are 6.8), (51% on 3rd down success; Dolphins are an impressive 56%).

Per usual, Indy is low on time of possession (they score quick). In year's past, I've adjusted the formula for the teams that work in no huddle.

Indy is 5th in points scored (Ravens, Giants, Vikings and Saints all score more).

What may be confusing is that I kept the same overall ranking (Indy as #7) for all areas (just for ease of copying/pasting charts and finding teams)... ((look at the 'Rank' Columns to see where a team stacks up instead))

[Maxie the Taxi]

Thanks for posting these. They make sense to me.

I was pleased to see Dallas was put in the proper prospective! The Cowboys ain't as good as they think they are.

The Pack ain't as bad as we might have thought either.

[Fritz]

I am highly curious as to how the Pack will play on Sunday. I would dearly like to see improvement on the offensive line and am frankly hoping that Lang will start at LT. That would give Clifton another week to heal and would give Lang some experience - not much, but some.

Man, if that offensive line can start blocking - pass and run blocking - that'd make me feel a whole lot better.

[get louder at lambeau]

Good stuff, Fosco. Thanks for posting that. Do special teams play a role at all in your rankings?

[bobblehead]

You said you are 60% against the spread...how do you choose who you are going to bet? You can't mean you are 60% and you play every game??

[Fosco33]

You said you are 60% against the spread...how do you choose who you are going to bet? You can't mean you are 60% and you play every game??

Yes, we pick every game (Monday night tiebreaker, with points) using the Tuesday lines.

There's no confidence picks - just total # correct over total games against the spread.

I've used my overall summary rankings to compare against the spread, for example:

Packers are 7.5 point favorites on the road vs. the Browns

Packers statistical ranking is 8.7. Browns are a paltry 25.9.
Since the spread is typically adjusted by Vegas for home field (3 points), I compare the total (10.5) vs. the difference in the ranks (25.9-8.7 = 17.2).

Given the difference in the ranking (17) is much larger than the adjusted spread (10), I take the Packers with the spread.

A better use of the tool: Vikes (8.4) vs. Steelers (12.8) with Pitt favored by 4.5 at home (with the HFA of 3 points).

The comparison (8.4-12.8 or -4.4) is different than the spread (+4.5). Therefore you take the Vikes as a small road dog.

I will make adjustments because I know the formula isn't 100% (bad weather, big spread - take the dog; injured dog QB - take the favorite). There are also teams that are playing over/under their potential (see original post) or lines that are wacky (because of crazy fans and greedy Vegas - aka Dallas and Houston).

I won the pool 2 years ago @61%, too (and am currently at 60%). I'm thinking of creating a free site just to post this stuff... then maybe people will use it and want to pay for it in the future.

[Fosco33]

Good stuff, Fosco. Thanks for posting that. Do special teams play a role at all in your rankings?

I've played with many other metrics- penalties; more breakouts (offensive passing/running; defensive passing/running); sacks; etc.

What I found is as the formula gets adjusted, the end result (overall summary ranking) is about the same. Meaning - I'd waste my time trying to get every stat right as a % of the overall ranking and the results wouldn't change. I use the KISS method (keep it simple, stupid).

[Maxie the Taxi]

Good stuff, Fosco. Thanks for posting that. Do special teams play a role at all in your rankings?

I've played with many other metrics- penalties; more breakouts (offensive passing/running; defensive passing/running); sacks; etc.

What I found is as the formula gets adjusted, the end result (overall summary ranking) is about the same. Meaning - I'd waste my time trying to get every stat right as a % of the overall ranking and the results wouldn't change. I use the KISS method (keep it simple, stupid).

Are you familiar with BudGoodeSports.com? (http://mule.he.net/~budsport/index.php)

He claims that Yds/Pass Attempt Differential is the best stat for prediction. His weighted prediction tool is scary accurate.

[Fosco33]

Good stuff, Fosco. Thanks for posting that. Do special teams play a role at all in your rankings?

I've played with many other metrics- penalties; more breakouts (offensive passing/running; defensive passing/running); sacks; etc.

What I found is as the formula gets adjusted, the end result (overall summary ranking) is about the same. Meaning - I'd waste my time trying to get every stat right as a % of the overall ranking and the results wouldn't change. I use the KISS method (keep it simple, stupid).

Are you familiar with BudGoodeSports.com? (http://mule.he.net/~budsport/index.php)

He claims that Yds/Pass Attempt Differential is the best stat for prediction. His weighted prediction tool is scary accurate.

Nope - haven't heard of that site. Thanks!

I could do all sorts of other metrics - and I might. I love Turnover Margin and 3rd downs. If you convert (or prevent) 3rd downs, you get the ball. You get the ball (and keep it) - you get more yards/points/wins. Simple stuff.

I checked out and ran their game predictor - funny that the Pack v Browns at 75% confidence played out exactly as I discussed above... I may be onto something here :P

[bobblehead]

You said you are 60% against the spread...how do you choose who you are going to bet? You can't mean you are 60% and you play every game??

Yes, we pick every game (Monday night tiebreaker, with points) using the Tuesday lines.

There's no confidence picks - just total # correct over total games against the spread.

I've used my overall summary rankings to compare against the spread, for example:

Packers are 7.5 point favorites on the road vs. the Browns

Packers statistical ranking is 8.7. Browns are a paltry 25.9.
Since the spread is typically adjusted by Vegas for home field (3 points), I compare the total (10.5) vs. the difference in the ranks (25.9-8.7 = 17.2).

Given the difference in the ranking (17) is much larger than the adjusted spread (10), I take the Packers with the spread.

A better use of the tool: Vikes (8.4) vs. Steelers (12.8) with Pitt favored by 4.5 at home (with the HFA of 3 points).

The comparison (8.4-12.8 or -4.4) is different than the spread (+4.5). Therefore you take the Vikes as a small road dog.

I will make adjustments because I know the formula isn't 100% (bad weather, big spread - take the dog; injured dog QB - take the favorite). There are also teams that are playing over/under their potential (see original post) or lines that are wacky (because of crazy fans and greedy Vegas - aka Dallas and Houston).

I won the pool 2 years ago @61%, too (and am currently at 60%). I'm thinking of creating a free site just to post this stuff... then maybe people will use it and want to pay for it in the future.

PM me your picks for the week. I'll go lay a small amount on every game each week and progress the bets as the season goes on. If you hold 60% for the season I'll send you a check based on how much I make....You have nothing to lose. If I lose money I'll bad mouth you to no end...but I won't actually be mad since I bet NFL plenty anyway. (and I'll likely start at $30 a game so I won't take too big of a bath if it goes bad.)

I only want this so I have a passing interest on every game and you do it anyway.

[Rastak]

Great stuff indeed Fosco....very well done.

[bobblehead]

You said you are 60% against the spread...how do you choose who you are going to bet? You can't mean you are 60% and you play every game??

Yes, we pick every game (Monday night tiebreaker, with points) using the Tuesday lines.

There's no confidence picks - just total # correct over total games against the spread.

I've used my overall summary rankings to compare against the spread, for example:

Packers are 7.5 point favorites on the road vs. the Browns

Packers statistical ranking is 8.7. Browns are a paltry 25.9.
Since the spread is typically adjusted by Vegas for home field (3 points), I compare the total (10.5) vs. the difference in the ranks (25.9-8.7 = 17.2).

Given the difference in the ranking (17) is much larger than the adjusted spread (10), I take the Packers with the spread.

A better use of the tool: Vikes (8.4) vs. Steelers (12.8) with Pitt favored by 4.5 at home (with the HFA of 3 points).

The comparison (8.4-12.8 or -4.4) is different than the spread (+4.5). Therefore you take the Vikes as a small road dog.

I will make adjustments because I know the formula isn't 100% (bad weather, big spread - take the dog; injured dog QB - take the favorite). There are also teams that are playing over/under their potential (see original post) or lines that are wacky (because of crazy fans and greedy Vegas - aka Dallas and Houston).

I won the pool 2 years ago @61%, too (and am currently at 60%). I'm thinking of creating a free site just to post this stuff... then maybe people will use it and want to pay for it in the future.

OK, you confused me just a bit with your explanation. the line for GB was -7.5, and you added 3 to that?

with minn/pitt the line is pitt -4, but you said the adjusted line is pitt -4.5....shouldn't it be pitt -1.5 since they are home? I got lost somewhere, I think I understand, but something isn't clicking....I'll go check your PM and see if I can figure it out, but a better explanation for other rats might help.

edit: what I'm asking is do YOU alter for home field or is it straight numbers. ie. Eagles (10.4) AT skins (19.1)...fosco's line is eagles -8.7, Vegas line is Eagles -6.5

Would you pick the eagles, or would you then adjust your line same as vegas and take 3 off for the eagles being the road team and your line is eagles -5.7. It changes which way you bet.

[bobblehead]

Hmm...after running all the numbers if I make the 3 point adjustment for the home team these are the games that vegas has a big disparity with you.

Green Bay -14.2 actual -7
Vikings -1.4 actual + 4
Patriots (no home team) -22.3 actual -14
Bills -1.2 actual +7

Not sure what I'm going to do now :) Guess I'll swing by the casino in the morning and make my Raiders money line bet and then look at the pregame lines and decide. I'll likely end up putting $44 on each game above for fun. Hope I go 3-1 :shock:

edit: or 4-0 with the vikings losing by 3

[Fosco33]

Sorry for the confusion. I don't consistently apply hfa and should pick either adjusting vegas lines to match with the stat lines or vice versa (or better explain my thought process). Sometimes I'll apply an away advantage (teams like Carolina from last year; or opponents that suck no matter where you play). Next week I'll try and post the adjustments.

The games you outlined do have potential. I like the Packers, Colts, Vikes, Bills, Falcons (weighting higher bets). I know you felt pretty good about the Raiders and after considering Sanchez has lost his confidence and Cotchery is questionable; Kris Jenkins is out for the year and Fargas actually got touches last week - I'll go with the Raiders. They've cost me winning a week ($180) a few weeks back so I'm gun shy.

Having to pick all the games (with gut intuition on Sunday morning)
Chicago, Packers, Chargers, Colts, Texans, Vikings, Patriots, Bills, Raiders, Saints, Falcons, Giants, Redskins. Only 6 teams will cover their spread (6/13) which is where you want to be. Now just cross fingers and crack a beer.

[MJZiggy]

Sorry for the confusion. I don't consistently apply hfa and should pick either adjusting vegas lines to match with the stat lines or vice versa (or better explain my thought process). Sometimes I'll apply an away advantage (teams like Carolina from last year; or opponents that suck no matter where you play). Next week I'll try and post the adjustments.

The games you outlined do have potential. I like the Packers, Colts, Vikes, Bills, Falcons (weighting higher bets). I know you felt pretty good about the Raiders and after considering Sanchez has lost his confidence and Cotchery is questionable; Kris Jenkins is out for the year and Fargas actually got touches last week - I'll go with the Raiders. They've cost me winning a week ($180) a few weeks back so I'm gun shy.

Having to pick all the games (with gut intuition on Sunday morning)
Chicago, Packers, Chargers, Colts, Texans, Vikings, Patriots, Bills, Raiders, Saints, Falcons, Giants, Redskins. Only 6 teams will cover their spread (6/13) which is where you want to be. Now just cross fingers and crack a beer.

Now you explain it, after Bobble's piddled away his life savings.

[bobblehead]

Actually I broke even on the day, but it was my pick of Oakland on the money line that cost me money, I went 3-1 with the games of Fosco's that I played.

By his rank/point system the Pack, NE, Bills, Vikes were the strong plays. I got my wish and went 3-1 with the vikes losing....although simply losing by 3 without the pick 6 would have been better as I would have been 4-0 (getting 4 points).

Overall the Picks you PM'd me went 8-3-1 with the Redskins losing atm so likley 8-4-1...still damn good picking. That is with the vegas line at game time.

As I told Fosco, I didn't like any games this week and wasn't betting my normal bets, but was putting $100 on the Raiders on the money line on a hunch (flying cross country, sanchez been razzed, Raiders DBacks are still money....I was wrong...BIG). I like to have action on games, it makes them interesting. I'm having a good season and thanks to the 3-1 I only lost about $40 yesterday total...that is small compared to my normal stakes.

Keep posting them Fosco, I'm interested to see how the rankings adjusted.

[bobblehead]

Bump.....Any chance on seeing the updated numbers Fosco?

[Waldo]

I love all the stat analysis like this.

I've got something totally different under the hood in the system I run (in my blog). Points only, net points (O scoring minus scoring D)

However it uses an engineering method to correct for strength of schedule for each team to near infinity. Basically instead of figuring out the per game average for each team (call it X and Y for offense and defense) using simple arithmetic, each offense X and defense Y instead is a variable that creates expected scores based on the matchups played, the entire set of X and Y for the league are set equal to each other (ΣX=ΣY), then the entire set of X and Y solved for iteratively so the results of the expected scores matches the set of actual scores with the lowest error possible.

The scoring O and scoring D then has the "yeah but" stripped from it, each stat is corrected for opponents (who are corrected for opponents, who are corrected for opponents, etc...).

I use it to do power rankings and other stuff, like a weekly updated strength of schedule, projected record, best/worst units, etc...

Thus far my power rankings are 73-30 straight up. Theoretically I could do spread calculations, next year perhaps.

[HarveyWallbangers]

Thus far my power rankings are 73-30 straight up. Theoretically I could do spread calculations, next year perhaps.

73-30 is pretty ordinary for straight up.

[Waldo]

Thus far my power rankings are 73-30 straight up. Theoretically I could do spread calculations, next year perhaps.

73-30 is pretty ordinary for straight up.

Power rankings or matchup calculations?

Most add a HFA. I'm beating many of the much more complicated (number of stats) systems out there straight up.

[Fosco33]

Bump.....Any chance on seeing the updated numbers Fosco?

Yes, I added some new metrics (point differential and yards gained/att vs. yards defened/att). And I'm cleaning up the methodology on HFA (going to adjust my summaries instead of the vegas lines - I took a look at years past and this was closer).

[Fosco33]

I love all the stat analysis like this.

I've got something totally different under the hood in the system I run (in my blog). Points only, net points (O scoring minus scoring D)

However it uses an engineering method to correct for strength of schedule for each team to near infinity. Basically instead of figuring out the per game average for each team (call it X and Y for offense and defense) using simple arithmetic, each offense X and defense Y instead is a variable that creates expected scores based on the matchups played, the entire set of X and Y for the league are set equal to each other (ΣX=ΣY), then the entire set of X and Y solved for iteratively so the results of the expected scores matches the set of actual scores with the lowest error possible.

The scoring O and scoring D then has the "yeah but" stripped from it, each stat is corrected for opponents (who are corrected for opponents, who are corrected for opponents, etc...).

I use it to do power rankings and other stuff, like a weekly updated strength of schedule, projected record, best/worst units, etc...

Thus far my power rankings are 73-30 straight up. Theoretically I could do spread calculations, next year perhaps.

Interesting. Where do you find updated SOS... I can only find preseason.

I was also considering a season summary vs. recent history (last 4 games). But I'd have to track each individual teams stats (by game) and have an ongoing list. It'd take me way to much time for something that may actually show a very similar trend. It basically takes me 30 minutes or less to pull together my info now - and I adjust the stats in my formula to limit the impacts of straight arithmetic vs. more using algorithims/etc. As I use this to help me smooth out vegas spreads, I still have to account for non-stat issues (momentum, injuries, weather, etc.). If my sole job was to do this, I'd get more complicated/accurate.

[bobblehead]

Bump.....Any chance on seeing the updated numbers Fosco?

Yes, I added some new metrics (point differential and yards gained/att vs. yards defened/att). And I'm cleaning up the methodology on HFA (going to adjust my summaries instead of the vegas lines - I took a look at years past and this was closer).

Yea, I glanced at that other website and the yards/att seems pretty telling. If you get time please post your spreadsheet by friday night so I can run the numbers saturday. I get into this stuff.

Waldo, that is a pretty complex system. You must have a background in statistics as I have a lot of background in it and that was tough for me to follow. I would be interested in your results too if you feel like posting them. I am a junkie for usable betting information.

edit: If I'm reading it correctly you will predict an actual score for each game, no?? Theoretically you could run the permutation of all 32 teams and predict the score of any possible game (not just the actual ones scheduled) for that week, but specifically if you ran the scores for the week you in fact COULD make a prediction against the spread.

[Fosco33]

Bump.....Any chance on seeing the updated numbers Fosco?

Yes, I added some new metrics (point differential and yards gained/att vs. yards defened/att). And I'm cleaning up the methodology on HFA (going to adjust my summaries instead of the vegas lines - I took a look at years past and this was closer).

Yea, I glanced at that other website and the yards/att seems pretty telling. If you get time please post your spreadsheet by friday night so I can run the numbers saturday. I get into this stuff.

Waldo, that is a pretty complex system. You must have a background in statistics as I have a lot of background in it and that was tough for me to follow. I would be interested in your results too if you feel like posting them. I am a junkie for usable betting information.

edit: If I'm reading it correctly you will predict an actual score for each game, no?? Theoretically you could run the permutation of all 32 teams and predict the score of any possible game (not just the actual ones scheduled) for that week, but specifically if you ran the scores for the week you in fact COULD make a prediction against the spread.

I got them done last night - just have to print screen and post them here. I could get them in better clarity if I could find a free hosting site that was bigger than 50KB.

[Waldo]

Waldo, that is a pretty complex system. You must have a background in statistics as I have a lot of background in it and that was tough for me to follow.

Engineering. That is not statistics.

Similar concepts are used to model indeterminate structures (structures that cannot be solved by a basic linear equation, instead a circular calculation is required). Or flow systems like drainage basins and traffic networks, both of which are also circular equations and require best fit approximate solutions.

[Waldo]

Interesting. Where do you find updated SOS... I can only find preseason.

I was also considering a season summary vs. recent history (last 4 games). But I'd have to track each individual teams stats (by game) and have an ongoing list. It'd take me way to much time for something that may actually show a very similar trend. It basically takes me 30 minutes or less to pull together my info now - and I adjust the stats in my formula to limit the impacts of straight arithmetic vs. more using algorithims/etc. As I use this to help me smooth out vegas spreads, I still have to account for non-stat issues (momentum, injuries, weather, etc.). If my sole job was to do this, I'd get more complicated/accurate.

Everything is calculated from the basic data set of weekly game scores.

[Partial]

Bump.....Any chance on seeing the updated numbers Fosco?

Yes, I added some new metrics (point differential and yards gained/att vs. yards defened/att). And I'm cleaning up the methodology on HFA (going to adjust my summaries instead of the vegas lines - I took a look at years past and this was closer).

Yea, I glanced at that other website and the yards/att seems pretty telling. If you get time please post your spreadsheet by friday night so I can run the numbers saturday. I get into this stuff.

Waldo, that is a pretty complex system. You must have a background in statistics as I have a lot of background in it and that was tough for me to follow. I would be interested in your results too if you feel like posting them. I am a junkie for usable betting information.

edit: If I'm reading it correctly you will predict an actual score for each game, no?? Theoretically you could run the permutation of all 32 teams and predict the score of any possible game (not just the actual ones scheduled) for that week, but specifically if you ran the scores for the week you in fact COULD make a prediction against the spread.

I got them done last night - just have to print screen and post them here. I could get them in better clarity if I could find a free hosting site that was bigger than 50KB.

Flickr?

[Fosco33]

Waldo, that is a pretty complex system. You must have a background in statistics as I have a lot of background in it and that was tough for me to follow.

Engineering. That is not statistics.

Similar concepts are used to model indeterminate structures (structures that cannot be solved by a basic linear equation, instead a circular calculation is required). Or flow systems like drainage basins and traffic networks, both of which are also circular equations and require best fit approximate solutions.

You ever use the Erlang C modeling for traffic. I used it once to develop a gigantic call center routing system.

[Fosco33]

Interesting. Where do you find updated SOS... I can only find preseason.

I was also considering a season summary vs. recent history (last 4 games). But I'd have to track each individual teams stats (by game) and have an ongoing list. It'd take me way to much time for something that may actually show a very similar trend. It basically takes me 30 minutes or less to pull together my info now - and I adjust the stats in my formula to limit the impacts of straight arithmetic vs. more using algorithims/etc. As I use this to help me smooth out vegas spreads, I still have to account for non-stat issues (momentum, injuries, weather, etc.). If my sole job was to do this, I'd get more complicated/accurate.

Everything is calculated from the basic data set of weekly game scores.

So you have to create a separate tracker - I'm looking for info available online. Oh well.

[bobblehead]

Waldo, that is a pretty complex system. You must have a background in statistics as I have a lot of background in it and that was tough for me to follow.

Engineering. That is not statistics.

Similar concepts are used to model indeterminate structures (structures that cannot be solved by a basic linear equation, instead a circular calculation is required). Or flow systems like drainage basins and traffic networks, both of which are also circular equations and require best fit approximate solutions.

yea, that makes sense...and you did mention engineering in the OP. The beauty is that your method likely works better for considering strength of competition because it will adjust and readjust each teams "strength" as it cycles through....I like it.

[Waldo]

Here is last weeks data run.

I've been looking for a better way of presenting it, this is the first gfx try instead of typed.

http://i726.photobucket.com/albums/ww269/Waldo56/PowerRankwk7.png

Rk = Rank
Pre = Prev
Net = Off - Def, Net Scoring
Off = Offense (opponent adjusted ppg)
O Rk = Rank of offense
Def = Defense (opponent adjusted ppg)
D Rk = Rank of defense
Chg = Difference between most recent 4 weeks and full season
C Rk = Rank of the change, positive to negative

Strength of schedule:

I used the power ranks for this, instead of W-L or quality opponents, rank is used as the data to compute strength of schedule. The average rank of their opponents (16 is average). These are calculated from the power ranks, no part of the power rankings is calculated from this strength of schedule.

http://i726.photobucket.com/albums/ww269/Waldo56/SchedWk7.png

P Rk = Previous Rank
F Rk = Future Rank

Estimated Record:

The record estimate uses the power ranks, existing wins, and future schedule to calculate the outcome of all remaining games, and assumes a probability of that outcome occurring.

It is assumed:
If a team is favored by 14+ points, I assumed that they win 100% of the time.
If a team is favored between 7 and 14 points, I assumed that they win 80% of the time.
If a teams is favored between 0 and 7 points, I assumed that they win 60% of the time.
If a team is not favored between 0 and 7 points, I assumed that they win 40% of the time.
If a team is not favored between 7 and 14 points, I assumed that they win 20% of the time.

http://i726.photobucket.com/albums/ww269/Waldo56/RecordWk7.png