QB Formula - predictions Week 7

[Challie]

Disclaimer: Work in progress, here to provoke thought, get thoughts on possible factors in a QB algorithm, and whatever else. Don't base too much on this. All this does is show what the past numbers would try to predict, obviously, each player won't score those numbers. When a QB goes off for 4 TDs, their scores would be huge, but there's very little way that this formula can call that kind of a shot in advance based on previous averages.

EDIT: McNabb & Roethlisburgers stats were corrected (wrong opponents entered previously).

Last Week: http://www.fantasyfootballcafe.com/foru ... 86#1216886

Scoring based on:
Pass & Rush TDs = 6 pts
50 Pass Yards = 1 pt
20 Rush Yards = 1 pt
Int & Fumble Lost = -2 pts

Factors Used:
YPG, Attempts, PPA (including yards lost via sacks), turnovers, TDs for both the QB and the opposing D. Home vs. away. Injuries for QB's off unit & the opposing D unit (based on official NFL injury list).

Here goes nothing for this week (I'll update this later in the week for latest injury report, for now, I'll stick with #s from last week's injury report).

Drew Bledsoe vs. SEA 17.71
Donovan McNabb vs. SDG 15.03
Josh McCown vs. TEN 14.37
Eli Manning vs. DEN 12.69
Carson Palmer vs. PIT 12.67
Mark Brunell vs. SFO 12.13
Gus Frerotte vs. KAN 11.25
Kurt Warner vs. TEN 10.69
Matt Hasselbeck vs. DAL 10.57
Brett Favre vs. MIN 10.10
Peyton Manning vs. HOU 9.82
Marc Bulger vs. NOR 9.81
Jake Plummer vs. NYG 9.63
Aaron Brooks vs. STL 8.99
Kelly Holcomb vs. OAK 8.41
Daunte Culpepper vs. GNB 7.26
Trent Dilfer vs. DET 6.90
Drew Brees vs. PHI 5.94
Vinny Testaverde vs. ATL 5.82
Steve McNair vs. ARI 5.51
Kerry Collins vs. BUF 5.29
Trent Green vs. MIA 5.20
Ben Roethlisberger vs. CIN 5.01
Michael Vick vs. NYJ 3.70
Anthony Wright vs. CHI 1.81
Joey Harrington vs. CLE 1.62
David Carr vs. IND 1.28
Kyle Orton vs. BAL -0.70
Alex Smith vs. WAS -5.20

I'm surprised that Bledsoe stays that high. Gus Frerotte?

QBs with fewer games played have less reliable stats (Vinnie, Alex Smith, Holcomb).

Love to hear some thoughts. I won't be online much today, but I'll be checking in through the week.

[Plindsey88]

I posted my thoughts on this, as well as my findings from trying the same thing a few years ago in your week 6 thread....

I don't mean to sound like I am discouraging you or anything, but the reason my defense formula tends to work is that I am looking at one entire team v. another entire team.... The smallest of changes in variables do not tend to affect the overall outcome... But, believe me, I've tried cranking the numbers for individuals, and you will find that you're better off looking at the fantasy numbers given up by a given D to quarterbacks, and comparing that to the total fantasy numbers of the quarterback in question..... When you start trying to factor in every possible variable for an individual you will find yourself buried in the minutia... Cause there are too many tiny variables that can affect an individual's play....

Have fun with it, but speaking as someone who's tried, I can tell you that your own common sense will be accurate more times than a formula will for an individual....

[kingmillerus]

McNabb is playing against San Diego, not Cincy in week 7.

[Challie]

EDIT: I went back and corrected the stats in the original post. Figures and opponents for Roethlisberger & McNabb were incorrect.

[Challie]

We'll see how this goes; I'm starting to look at how much to weight QB vs. Def stats; always significantly more for the QB, always equal or dependent on home or road game. We'll see what previous weeks numbers have to say about that.

[jjigglers]

You're weighting (weighing?) something wrong, because your QBs are only getting ~225 fantasy points this week, while QBs are scoring ~285 fantasy points per week (10.19 points x 28 teams in a bye week).

If you're interested, here's what I got.

1. B. Favre (MIN): 16.47
2. M. Brunell (SF): 15.84

3. C. Palmer (PIT): 14.01
4. V. Testeverde (ATL): 13.80

5. D. McNabb (SD): 12.55
6. E. Manning (DEN): 12.53
7. J. McCown (TEN): 11.68
8. G. Frerotte (KC): 11.64
9. D. Culpepper (GB): 11.54
10. D. Bledsoe (SEA): 11.19
11. J. Plummer (NYG): 11.15
12. T. Dilfer (DET): 11.10
13. D. Brees (PHI): 10.93
14. STL QB (NO): 10.83
15. M. Vick (NYJ): 10.58
16. P. Manning (HOU): 10.22
17. PIT QB (CIN): 9.69
18. A. Brooks (STL): 9.56
19. M. Hasselbeck (DAL): 9.44
20. T. Green (MIA): 9.34

21. S. McNair (ARI): 8.50
22. K. Collins (BUF): 7.09
23. J. Harrington (CLE): 6.98
24. K. Holcomb (OAK): 6.04
25. A. Smith (SF): 5.58
26. K. Boller (CHI): 5.48
27. D. Carr (IND): 5.17
28. K. Orton (BAL): 4.42

[The Rush]

These look like good projections, nice job.

[smackthefirst]

I posted my thoughts on this, as well as my findings from trying the same thing a few years ago in your week 6 thread....

I don't mean to sound like I am discouraging you or anything, but the reason my defense formula tends to work is that I am looking at one entire team v. another entire team.... The smallest of changes in variables do not tend to affect the overall outcome... But, believe me, I've tried cranking the numbers for individuals, and you will find that you're better off looking at the fantasy numbers given up by a given D to quarterbacks, and comparing that to the total fantasy numbers of the quarterback in question..... When you start trying to factor in every possible variable for an individual you will find yourself buried in the minutia... Cause there are too many tiny variables that can affect an individual's play....

Have fun with it, but speaking as someone who's tried, I can tell you that your own common sense will be accurate more times than a formula will for an individual....

Wow this could be a first if memory serves, but I have to kind of disagree Plindsey. I do agree that projecting D's is easier than all other positions but for another reason all together.

The main reason D's are easier and other positions are harder is primarily due to the scoring of those positions in FF. In the simplest sense, D's scores are primarily based on larger spectrums, whereas other postions scoring is based on smaller spectrums. If an offense drives 80 yards on a D with ease, the D is likely to lose a small amount of points depending on your scoring and where it falls in the game. However, if an offense drives 80 yards and scores a TD, the scores are much more varied. Obviously you could have 0 if your player did nothing. But if your player was involved and got the TD, you could very easily get 10 points or more.

So that same drive only varies the D's score by 3 or so where as it varies for a skill player very much more. Because of this I think you are always going to see a greater variance when trying to project position players.

But that doesn't mean I wouldn't continue to try. I would think that with the more data one incorporated and crunched would eventually allow someone to put out very good rankings. Of course everyone would have to realize that the variance for skilled positions would always exist.

[Makaveli]

i dont see how you guys both have Hasselbeck rated so low? wow, he's gonna have a big game, expect a shoot out 300 yds and 2 tds minimum

[jjigglers]

i dont see how you guys both have Hasselbeck rated so low? wow, he's gonna have a big game, expect a shoot out 300 yds and 2 tds minimum

Maybe because Dallas D is giving up less than 10 PPG to QBs against an above average set of passing offenses and Hasslebeck averaging under 12 PPG so far this season.