What's new
Fantasy Football - Footballguys Forums

This is a sample guest message. Register a free account today to become a member! Once signed in, you'll be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox!

Article of the Year! (1 Viewer)

agree 100%

can we expect similar articles for other positions? RB, WR, etc.....

Best FF read in years for me.

 
Great article, but note that the bottom of that last chart is as condemning as the top is confirming.

Jake Plummer is dead last with Collins a few slots above him.

Still, the methodology seems sound and you can't really argue with the correlation coefficient. I would like to see an analysis with much more data than just 1 year.

 
Nice work Chase. :thumbup:

I hope you're right about your conclusions on Brooks because I've targeted him in numerous 2006 drafts already ( have him in 2 WCOFF satellites, No Mercy, SSLIV H2H, SSL 5 IDP, and Masters leagues). That said, I hope Bledsoe proves you wrong because I have him in a handful of other ones. :bag: Great analysis though.

 
Last edited by a moderator:
And where can I find a QB ranking list from Mr. Stuart.

I'm sold.
Click to expand...
His Undervalued QB'sA.Brooks

P.Rivers ( a name i :crossfingers: think may figure prominently in QBBC w/ J.Kitna?)

Overvalued

E.Manning

M.Hasselbeck

 
Last edited by a moderator:
More love for Michael Vick, I see. Good, I say! He's very underrated right now.

With that said, I think a reason why Michael Vick does so well above the expected FP/G is because the vast majority of QBs score with their arms, so a defense with a low ExpFP/G will be one with a great passing defense. However, a great passing defense is no way to stop Vick- you actually need a better running defense. So a tough slate of defenses for Peyton Manning, say, might be a really easy slate for Michael Vick.

He's still very underrated, though. :)

 
I really liked it too, but is there a reason why Smith's draft slots article link also takes you to this article? It's an excellent read, but wouldn't one link be enough? Plus, I kind of want to read Smith's article too. :)

 
Nice read. However, any data produced that tells me that P Manning had a 0 point difference rating in 2004, when he throws for 4,557 yards 49 TD's , and only 10 picks strikes me as being very flawed! Also, from my take on the article it seems like it wants to level the playing field into a utopian league where everyone is playing on a level field. It adjusts for an easy schedule, but the truth is for a QB Like M Hasselbech it is going to be an easier task playing against teams like; SF, STL, and ARZ twice each every year, than someone like C Frye who plays against Balt, Pitt, and Cinncy twice.

Also, it doesn't really take into account relative talent around the QB, coaching style, injury risk, or competition at the QB position. Guys like A Brooks excite me, but the threat of having A Walter breathing down his neck, A Davis leaving Art Shell on a short leash, and the fact that their DF sucks makes him a big question mark in my opinion. The Raiders will probably be bad this year, and it is entirely likely that Brooks could be benched just for the fact that they have a loosing record and Al Davis wants to see if he has the QB of the future in Walter.

All in all, I found it to be good, but would like to hear what Kathy Fazio thinks about the subject before I take it as the solemn gospel.

 
Cool article. :thumbup:

I already thought Vick & Brooks were underrated and E.Manning and Bledsoe were overrated without considering last year's SOS. This makes me feel more comfortable about those judgments.

 
any data produced that tells me that P Manning had a 0 point difference rating in 2004, when he throws for 4,557 yards 49 TD's , and only 10 picks strikes me as being very flawed!
Click to expand...
What you're seeing is a zero RANK differential. Manning's rank was #1 via traditional numbers and also #1 via Chase's adjusted numbers. I'd say you should be suspicious of any system that does NOT give Manning a rank differential of zero (or real close to it) in 2004.
 
any data produced that tells me that P Manning had a 0 point difference rating in 2004, when he throws for 4,557 yards 49 TD's , and only 10 picks strikes me as being very flawed!
Click to expand...
What you're seeing is a zero RANK differential. Manning's rank was #1 via traditional numbers and also #1 via Chase's adjusted numbers. I'd say you should be suspicious of any system that does NOT give Manning a rank differential of zero (or real close to it) in 2004.
Click to expand...
That makes sense. Thanks for the clarification.
 
Nice read, even if it does makes me feel worse about losing Marc Bulger last year... :wall:

 
Last edited by a moderator:
:bag:

I thought I could provide a link to the draft slots article. But no dice. I'll ping Dodds.

 
Last edited by a moderator:
Good article! Makes you think. However, as a 2006 projections tool, one more variable should be considered, which is the schedule DIFFERENTIAL between 2005 and 2006. In other words, adjusting for the norm only tells us how the QBs would do if everyone played an average schedule. But all schedules aren't equal.

Some of the QBs who played an easy 2005 schedule have an easy expected 2006 schedule, while some have a hard expected 2006 schedule. So the ultimate tool would be to further adjust the value of these QBs based on their expected 2006 schedule.

 
And the other part of this article which should not be ignored is that the rank differential between actual points per adjusted game and value added was at most three spots except for four QBs:

Favre 9

Brooks 6

Delhomme 5

Collins -5

Favre and Brooks faced two of the tougher schedules in expected fantasy ppg and their teams combined for a record of 7-25. What probably led to their improved performance is the tendency for teams to pass more when trailing in games and get those second half "pity TDs" that defenses will allow when playing prevent.

Still an interesting read, but I'm not sure it's the "article of the year" and it has limited utility given the unknown concerning 2006 schedule. I mean, if Favre plays another tough schedule, then it doesn't really matter that he was the rank differential leader, because his numbers will be suppressed by another tough schedule...

 
I had already targetted Brooks and Favre as my mid-round QB pickups this year. This article just confirms what I was already thinking, and makes me more comfortable in my decision. GREAT VALUE.

Thanks for the nice read.

 
I had a similar article half written last year, and then saw Chase post his :D

One thing I would say is that some teams will lean on their passing game and others on their running game. The Rams passed a lot under Martz and it suited their talent. They also trailed quite a bit. Therefore I would have expected Bulger to do better than the normal level of production.

The Colts are another team that can seemingly pass at will most weeks.

It's still something we can use of course.

 
Good article! Makes you think. However, as a 2006 projections tool, one more variable should be considered, which is the schedule DIFFERENTIAL between 2005 and 2006. In other words, adjusting for the norm only tells us how the QBs would do if everyone played an average schedule. But all schedules aren't equal.

Some of the QBs who played an easy 2005 schedule have an easy expected 2006 schedule, while some have a hard expected 2006 schedule. So the ultimate tool would be to further adjust the value of these QBs based on their expected 2006 schedule.
Click to expand...
I coded something very close to what you mentioned and use it to make my own projections. However, it's only as good as I project the changes in pass defenses for all 32 teams to be. I'll use it for my rankings but I do recognize the flaws.
 
Chase,

I'll mirror what the others are saying. Good Article and Good Read. I do have a question though.

If you took the 2005 FP/game or the expected FP/G and applied that to the anticipated 2006 SOS (I don't have those numbers at hand) , is there a correlation to your conclusions? How would this compare to current rankings or does it support the thinking of undervalue or overvalue for a position?

Thanks

 
Last edited by a moderator:
First off great read really gets you thinking

I had a question though. FP/AdgG Rk is essentially suppose to mimick ADP correct? What in the case of say Marc Bulger whos ADP isnt 1 its 7 so the difference would be 6 not 0 and say Brett Favre whos ADP is 18 so the difference would be 3.

Would you trust using ADP more to compare the difference or FP/AdjG Rk and why?

 
Hmmm...note the pattern Chase???

BassNBrew Aaron Brooks Oak/3 QB9 9.08 104 107 BassNBrew Brett Favre GB/6 QB17 11.08 128 110 No wonder the MB will soon be 3-0 v. the Staff.

I'll save you investing time to do the RBs...

Code: 
BassNBrew Brian Westbrook Phi/9 RB8 1.08 8 16 BassNBrew Willie Parker Pit/4 RB20 3.08 32 26 BassNBrew Jamal Lewis Bal/7 RB23 4.05 41 28
 
What probably led to their improved performance is the tendency for teams to pass more when trailing in games and get those second half "pity TDs" that defenses will allow when playing prevent.
Click to expand...
Yeah, very likely true for these guys. Like I said before, I'd still like to see way more data than this but I'm bumping Brooks regardless.
 
Lots of kind words here. Thanks a bunch. Doug Drinen did most all of the heavy computer work and came up with the adjusted games idea, so he deserves the real credit. I really believe in the utility of this stuff, and I'm glad the final product came out to be something we both enjoyed.

Last year the numbers were more revealing. The numbers are what they are -- some years there's a big spread, and some years (like this one) there just aren't. But what's most important is that I believe 100% in the process. The biggest advantage this has is that it's entirely objective -- we run a few numbers, but there's no human element involved that might prejudice the rankings. The bad thing, of course, is that there's no subjectivity. And if you ran every single statistical test ever designed on the 2004 results, you never would have predicted Eli Manning to be a fantasy stud last year. That's a drawback of the system, but one I'm perfectly fine with.

I was out of commission yesterday, so I'm going to go through this thread now and answer all the great questions posed.

 
can we expect similar articles for other positions? RB, WR, etc.....
Click to expand...
I hope so. I know Chase was concerned that you can't do FP/AdjG calculations for other positions (since multiple RB/WR/TE will score every game), but I think if he just uses FP/G it will still be useful.
 
Great Stuff!

ETA on By Committee articles?
Click to expand...
Hey TeamDingo,Thanks. The Committee articles are coming along well. A few pieces of good news:

DTBC

The numbers this year are really strong. That's lucky, and very good. I expect the committee's value to be better than ever. The staff has thrown around some good suggestions as well on possible tweaks. More minds are always better than one.

Additionally, lots of groups are really strong. Just another weird coincidence. In the end I suspect the people who read the article on DTBC article will be in much better shape than your average fantasy player when it comes to selecting Ds. Lots of hidden value around this season.

QBBC

Drinen's given me some additional historical data this year, so that should figure in prominently. I haven't run any of the numbers yet as I'm currently working on tweaking the formula. So we'll definitely see the DTBC article come out first.

 
Great article, but note that the bottom of that last chart is as condemning as the top is confirming.

Jake Plummer is dead last with Collins a few slots above him.

Still, the methodology seems sound and you can't really argue with the correlation coefficient. I would like to see an analysis with much more data than just 1 year.
Click to expand...
Just want to be sure we're on the same page here. Jake Plummer is dead last on the last chart, i.e., the 2004 chart. This means in 2004 Plummer benefitted from a super easy schedule. In 2004, he ranked 5th. In 2005 he played in all 16 games but still ranked just 11th. I think this is a check in the "rearview SOS is good" column, because it explains why a healthy Plummer "regressed" in 2005; his 2004 numbers were merely inflated by a cupcake schedule.(And to really toot my own horn, that 2004 cupcake schedule was predicted in that year's QBBC article.)

 
Last edited by a moderator:
Nice work Chase. :thumbup:

I hope you're right about your conclusions on Brooks because I've targeted him in numerous 2006 drafts already ( have him in 2 WCOFF satellites, No Mercy, SSLIV H2H, SSL 5 IDP, and Masters leagues). That said, I hope Bledsoe proves you wrong because I have him in a handful of other ones. :bag: Great analysis though.
Click to expand...
I don't know what to make out of the fantasy community, outside of its very smart. Bledsoe ranked 6th last year, and added TO. Yet his ADP is QB10. This never would have happenned five years ago. Not too long ago Bledsoe's ADP would have been 6 without adding TO; now the fact that Owens is a Cowboy and Bledsoe's ADP dropped, means Bledsoe shouldn't really be a bust this year. QB10 is still too high, but it could have been much worse.Very difficult to consistently beat the fantasy market.

 
More love for Michael Vick, I see. Good, I say! He's very underrated right now.

With that said, I think a reason why Michael Vick does so well above the expected FP/G is because the vast majority of QBs score with their arms, so a defense with a low ExpFP/G will be one with a great passing defense. However, a great passing defense is no way to stop Vick- you actually need a better running defense. So a tough slate of defenses for Peyton Manning, say, might be a really easy slate for Michael Vick.

He's still very underrated, though. :)
Click to expand...
I'm not a huge Vick fan, so I was shocked to see him perform so well here. But I'd still be a little careful about your conclusions. For starters, Vick ranked 11th in Adjusted FP/G in 2004 (which doesn't count SOS at all), had an easy schedule, and then only ranked 9th when you factored in SOS. So essentially Vick acted like a normal above average QB in 2004, and not a normal above average QB that is the best running QB ever.Your point is a good one, but the reason I'd be careful is you're forgetting about the flip side. A defense with a hgih ExpFP/G will be one with a bad passing defense. And a weak passing defense won't help Vick much.

Of course, if Vick plays a lot more great passing Ds than bad passing Ds, and assuming, without agreeing, that the type of D Vick plays is irrelevant, then yes -- Vick will "cheat" the rearviewSOS system and look better than he is.

It's an interesting theory, but I don't believe that plays a significant role. I'd have to look into it a bit more, but right now I've got a few other things on my plate.

 
Cool article. :thumbup:

I already thought Vick & Brooks were underrated and E.Manning and Bledsoe were overrated without considering last year's SOS. This makes me feel more comfortable about those judgments.
Click to expand...
Thanks...I knew this article would be right up your alley.As I've said before, the SOS numbers just aren't as revealing as last year's version. The spread on expected FP/G was much wider in 2004.

But interestingly, the FP/AdjG spread was bigger this year, I think. And that's what really made you think Manning and Bledsoe were overrated. Manning and Bledsoe ranked 3rd and 6th in FPs; but they ranked 14th and 17th in FP/AdjG because they played every snap. They ended up ranking 16th in 20th in the value added ranking (once you factor in SOS), but the bigger downgrades were obviously the ones by 11.

Brooks and Vick definitely had hard schedules, and should have felt underrated. A guy like Brad Johnson has been ignored mostly, and he wasn't very good when he played. BUT, he did have a very, very hard schedule last year. Daunte Culpepper had an almost equally hard schedule. When you figure that the Vikings QBs had probably the toughest schedule in the league in 2005, and one of the very easiest in 2004, that at least begins to explain why everyone on Minnesota's passing game was a bust last year. And maybe Johnson, Culpepper, Burleson and any of the Min WRs should get a small uptick in their 2006 rankings.

 
If you took the  2005 FP/game or the expected FP/G and applied that to the anticipated 2006 SOS (I don't have those numbers at hand) , is there a correlation to your conclusions?  How would this compare to current rankings or does it support the thinking of undervalue or overvalue for a position?
Click to expand...
I'm not exactly sure what your question is here. I *think* you're asking me if the SOS for each QB remains constant, or whether the FPallowed to QBs by each team's defense remains constant. If what you're asking me is the first, I'd imagine the correlation is very low. If what you're asking me is the second, I'd also imagine the correlation is pretty low.People say "defenses are hard to predict" quite a lot, and I think that's mostly* true. But there are two things in that statement:

1) Defenses are hard to predict when you're choosing which fantasy defense you want to draft. This is mostly true, and the reason why I do the QBBC article. On the other hand, I did some research the other night and noticed that (adding to a very large list of things) Dodds' defensive predictions were pretty darn good. Hopefully he won't put me out of business here. ;)

2) Defenses are hard to predict when you're choosing which RB to play against which Defense. When you say defenses are hard to predict in terms of judging SOS for your offensive players, I think this is mostly true. But Clayton Gray does a great job with his SOS analysis. But he doesn't just copy over last year's numbers -- which is what I *think* you're asking me to correlate; and in that case, yes it's pretty unpredictable.

*I don't really believe this. What I think is that most people spend 90% of their time studying offense and not defense, so we're better at predicting offense. Shooting a free throw from 15 feet out isn't any more difficult than shooting a 15 foot shot at a 45% angle, but we practice that much less often, so we're worse at it. I used to be very confident about my predictions on defense, but unfortunately I wasn't able to watch as many games as I would have liked to last season. Fortunately for me, I can rely on our great IDP staff for support. :thumbup:

 
Last edited by a moderator:
[M]y take on the article it seems like it wants to level the playing field into a utopian league where everyone is playing on a level field. It adjusts for an easy schedule, but the truth is for a QB Like M Hasselbech it is going to be an easier task playing against teams like; SF, STL, and ARZ twice each every year, than someone like C Frye who plays against Balt, Pitt, and Cinncy twice.
Click to expand...
However, as a 2006 projections tool, one more variable should be considered, which is the schedule DIFFERENTIAL between 2005 and 2006. In other words, adjusting for the norm only tells us how the QBs would do if everyone played an average schedule. But all schedules aren't equal.

Some of the QBs who played an easy 2005 schedule have an easy expected 2006 schedule, while some have a hard expected 2006 schedule. So the ultimate tool would be to further adjust the value of these QBs based on their expected 2006 schedule.
Click to expand...
Hey guys,I understand what you're getting at, and I hope I can explain my reasoning here. You've correctly determined that I'm trying to get the QBs on a level playing field. I want this to be entirely objective; that way when I start my QB rankings, I don't have to say "well XXX was good but he had an easy schedule." I can just look to my value added column.

2006 SOS, conversely, is subjective. But while objective isn't always good and subjective isn't always bad, here's the problem.

With rearview SOS analysis, we know exactly what happenned. Let's pretend we have this:

The league average QB scores 15 FP/G. Team Defense Z played 16 games last year, and its first 15 games were against 15 QBs that all averaged 15 FP/G. So we know that Team Z's schedule (through 15 games) was exactly the league average.

In those 15 games, Team Z allowed 10 FP/G. So we should feel pretty confident that Team Z cuts off a QB's FP/G by 5 FPs. Now, QB A is going to play Team Defense Z in the 16th game. QBA averages 20 FP/G. We can confidently reduce QBA's points by 5 (you might argue for a 33% reduction; the practical difference is minimal, and I'd prefer using -5), and project for QBA 15 points against Team Z.

So that's all well and good. Nice, objective analysis.

Now, it's Year N+1. And we project Team Z to be 5 points per game stingier than average. But here's the problem: our projections aren't very good. We need to discount any QB playing Team Z not by 5 points, but by a much smaller number. Team Z might have lost two LBs, could have gotten cocky in the off-season, or changed their scheme. Maybe injuries will hit Team Z. Maybe their offense will change and the team dynamic will hurt the defense. Whatever it is, we need to discount that "-5" you want to give all QBs that play Team Z. It's unclear to me what it is; some astute people think we might want to only hurt a QB playing Team Z by 10%.

So in the end, I'm not going to add in 2006 schedule to my article, because it's not worth it to subtract 0.5 FP/g to each QB, or whatever the number is. Clayton Gray does a bangup job with his 2006 SOS analysis, and he even compares it to 2005 numbers (which was partly the inspiration for my article).

Hope that helps.

 
Also, it doesn't really take into account relative talent around the QB, coaching style, injury risk, or competition at the QB position. Guys like A Brooks excite me, but the threat of having A Walter breathing down his neck, A Davis leaving Art Shell on a short leash, and the fact that their DF sucks makes him a big question mark in my opinion. The Raiders will probably be bad this year, and it is entirely likely that Brooks could be benched just for the fact that they have a loosing record and Al Davis wants to see if he has the QB of the future in Walter.
Click to expand...
There's a reason i don't take that into account; it's because you do. ;) This article should be a starting point. When you look at Donovan McNabb's numbers, you'll say wow he aded a lot of value. But you'll also know that Andy Reid loves passing the ball. And he may or may not pass a ton again this year -- but that's up to you. Obviously the opposite was true of Big Ben.

All the other stuff (number of attempts, surrounding talent, etc.) you guys take into account. You also take into account changes -- so Kitna gets a boost because he adds Martz, and Culpepper takes a hit when he loses Randy Moss. And my article isn't saying DRAFT AARON BROOKS!!!111. It's basically saying "Fantasy players use tons of tools to create the fantasy market. There's one tool, though, that they don't use. And that's a tool they should use. And if they did use it, Aaron Brooks and Mike Vick would be ranked a lot higher. Proceed with caution."

 
You must log in or register to reply here.

Similar threads

Yah_Shoor_Yoobetchya
Staff Article: 2025 NFL Rookie Rankings: Top 14 Quarterbacks
Replies
2
Views
211
Gatorade
Gatorade
TheWinz
  • Sticky
2025 FBG Subscriber Contest
2
Replies
80
Views
2K
BassNBrew
BassNBrew
Kuz
Game predictor article
Replies
6
Views
252
Dizzy
Dizzy

Users who are viewing this thread

Total: 1 (members: 0, guests: 1)
Back
Top