Mikie2times Posted November 30, 2005 Share Posted November 30, 2005 A poster mentioned in an earlier thread just how out of whack the QB rating system is. I was thinking the same thing and was working on developing a more accurate system. That post got me to finish it off tonight, and I just wanted to see how people feel about the approach. The following is how I arrived at each statistic for my rating, and how it compares to the NFL rating system. My ratings are in bold, and feel free to ask questions as I'm sure I left a few things out. % Positive plays= Rushes+Completitions/Attempts+ Sacks+Rushes NFL passer Rating system= Completitions/Attempts TD%= Rushing TD's + Passing TD's/Attempts+Sacks+Rushes NFL passer Rating System= TD passes/Attempts Yards Per Attempt= (Passing Yards- Yards After Catch)+Rushing Yards/Attempts+Sacks+Rushes NFL passer Rating= Yards/Attempts Turnover%= INT+FUMBLES/Attempts+Sacks+Rushes NFL Passer Rating= INT/Attempts TD's a Game= Rushing TD's+Passing TD's/Games Turnovers a Game= INT's+FUM/ Games Played Running Ability- Rushing Yards/Games To come up with how I weighted each stat I measured how much each stat was correlated with points above the league average and below league average based on 4 years of data. The order of importance went TD%, TURN%, TD/GAME, TURNOVER/GAME, adjusted yards per attempt, % Positive plays, and Running ability. Here is how the 2005 QB's shape up so far. Name----------------RATE Carson Palmer------102.8 Peyton Manning ----102.0 Ben Roethlisberger--97.2 Jake Plummer------- 89.3 Matt Hasselbeck-----84.4 Michael Vick----------82.1 Drew Brees-----------81.9 Byron Leftwich-------80.3 Tom Brady---------- 80.2 Marc Bulger----------79.7 Steve McNair---------78.3 Drew Bledsoe--------76.6 Mark Brunell---------73.6 Trent Green----------72.3 Eli Manning------------72.1 Donovan McNabb----72.0 Brett Favre-----------70.2 Jake Delhomme------70.2 Aaron Brooks---------70.0 Brian Griese----------69.3 Kurt Warner-----------64.9 Kerry Collins----------64.4 Chris Simms----------62.5 Josh McCown----------61.5 David Carr------------60.4 Anthony Wright-------58.6 Trent Dilfer------------57.4 Daunte Culpepper----56.5 Gus Frerotte----------52.9 J.P. Losman-----------51.1 Joey Harrington-------51.1 Kyle Orton-------------45.4 Link to comment Share on other sites More sharing options...
Mikie2times Posted November 30, 2005 Author Share Posted November 30, 2005 Here are some numbers that I thought might surprise, or confirm opinions. Some explanation of the stats is included below, but a full explanation is in the first post. System QB? -Tom Brady had the biggest adjustment to his yardage total when subtracting his YAC yards. His average per attempt with YAC places him 6th in the league, without YAC he ranks 13th. Same Old Drew -Bledsoe ranks 25th in turnover %. Running for his life - JP Losman is second in the NFL in Rushing Attempts per drop back. He runs 12.5% of the time, while Vick runs 21.7% of the time. Big Ben, Deceptively good - One common trend I found in teams that score a lot of points, and don't score a lot of points, is turnover%, and TD%. While big ben might not have the gaudiest statistics he simply dominates in those two categories, ranking 2nd in turnover%, and 1st in TD%. Not exactly Trent Dilfer - Kyle Orton's 10 Fumbles, and 12 INT's rank him dead last in Turnover%. He puts the ball at risk on 6.6% of his drop backs. Eli could be more efficient. - % Positive plays takes into account the number of drop backs a QB has taken, and the number of times a positive result was achieved. Positive results include pass completions, or rushing plays (to register as a rush the play must be positive). Eli Manning registers a positive play only 52% of the time, ranking 31st in the league. Also of note is out of Eli's 7 fumbles, only 1 has been lost, a clear indicator of the breaks that can happen in football. Link to comment Share on other sites More sharing options...
Johnny Coli Posted November 30, 2005 Share Posted November 30, 2005 That is some awesome work, Mike. The ranking your system puts out for this season is dead-on with how one would rank the QBs based on observation. Really nice work. Especially this little nugget: System QB?-Tom Brady had the biggest adjustment to his yardage total when subtracting his YAC yards. His average per attempt with YAC places him 6th in the league, without YAC he ranks 13th. 516172[/snapback] Not surprising, because it's what we've all known for some time now. I'd like to see the individual rankings per catagory as compared to the way the league calculates them. How do they shake out using your system, i.e. % positive plays, etc. Link to comment Share on other sites More sharing options...
34-78-83 Posted November 30, 2005 Share Posted November 30, 2005 Nice work Kzoo! Very interesting... There's probably not a way to do it based on statistics provided but average time in the pocket would be an interesting factor, considering our guys get about 1.5 to 2.5 seconds before there is a defender in their face Link to comment Share on other sites More sharing options...
RuntheDamnBall Posted November 30, 2005 Share Posted November 30, 2005 Interesting, but if you were going this far, you'd have to go further and look at individual plays such as situations where the receiver bobbled the ball, and it was intercepted. Is that one on the passer? Not really. It's happened to both Losman and Harrington at least once (maybe more?). Link to comment Share on other sites More sharing options...
AKC Posted November 30, 2005 Share Posted November 30, 2005 % Positive plays= Rushes+Completitions/Attempts+ Sacks+RushesNFL passer Rating system= Completitions/Attempts TD%= Rushing TD's + Passing TD's/Attempts+Sacks+Rushes NFL passer Rating System= TD passes/Attempts 516147[/snapback] You've expanded from the simpler "passer rating" to offer an overall "quarterback rating" and your findings are interesting. I can see my original primary concern regarding uncontested passes behind the Line Of Scrimmage is in a very small way addressed, but I'd think the results would be far more accurate on the passing side if there was more weight given to the opportunity for a pass to be defensed in any of the ways it might be as the difficulty of the pass/time to develop increases. As it is, a few QBs are throwing 10%+ of their passes behind the line with absolutely no chance of the pass being defensed. To allow these as whole "completions", or in other words to suggest they are in any way equal to a 45 yard pass completion skews the results DRAMATICALLY in favor of the few QBs in the league who play in the behind the line type of passing attacks. For instance a guy who throws a lot of downfield balls yet is only putting 2 or 3% of his passes up in the indefensible areas behind the LOS suffers mightily in the ratings versus a Tom Brady type who is throwing 10% of his behind the line. You point out that simply removing RAC yardage from Brady's stats drops him well down the QB rating ladder, if you additionally eliminate or reduce the factor for all the indefensible passes he throws behind the line he ends up in the bottom half of the league among all starting QBs at about 18th or 19th in the NFL. And to be fair to a typically very good game manager like Brady, that intangible would allow him to gain 4 or 5 spots in the ratings against the whole of the league and put him at about 14th league-wide, probably a fair spot for a guy whose team has been hovering around .500 all season. Link to comment Share on other sites More sharing options...
MDH Posted November 30, 2005 Share Posted November 30, 2005 Yards Per Attempt= (Passing Yards- Yards After Catch)+Rushing Yards/Attempts+Sacks+Rushes [/b] 516147[/snapback] Personally I think it's a big mistake taking away YAC when figuring out the ratings. By taking away YAC you're forgetting the importance of accuracy. You're saying it's the same thing to throw a pass 2 yards behind a WR on a slant where he has to stop and catch the ball as it is to throw it perfectly while the WR is in stride thereby allowing him to continue on and gain extra yards. Link to comment Share on other sites More sharing options...
AKC Posted November 30, 2005 Share Posted November 30, 2005 Personally I think it's a big mistake taking away YAC when figuring out the ratings. By taking away YAC you're forgetting the importance of accuracy. You're saying it's the same thing to throw a pass 2 yards behind a WR on a slant where he has to stop and catch the ball as it is to throw it perfectly while the WR is in stride thereby allowing him to continue on and gain extra yards. 516528[/snapback] But isn't accuracy better represented in completion % rather than in skewing the marks for a passer by crediting him for the good feet and blocking of his receiving corps and the quality and philosophy of his OC and the level of defensive talent they face? Link to comment Share on other sites More sharing options...
Mikie2times Posted November 30, 2005 Author Share Posted November 30, 2005 Run the Damn Ball/AKC- I would include defensible passes, or tipped INT's, the problem is I don't have access to that information as NFL stats are still fairly limited in depth. This is just an attempt at a more accurate system using every variable I have access to, still a much more accurate system could be made with more detailed stat keeping. MDH- 50% of the normal NFL QB rating is made up of yards per attempt, and Completition %. Now I'm not going to argue that YAC success isn't influenced by accuracy, we all know QB's that deliver a well placed ball will allow receivers to gain additional yards. The problem is once a pass is either caught, or incomplete, the QB has no additional impact on the outcome of the play. With half the standard passer rating being made up of completition% and yards per attempt it leaves short system passers a much easier path to boost the ratings. With what we have available we can't do much to change the higher accuracy numbers, and accuracy still needs to play a roll in any system, but we can still take away the credit a QB receives for having a 5 yard slant turn into a 30 yards play. In terms of measuring performance, a QB that actually completed a 30 yard pass threw the air probably had a much more difficult completition then a 5 yards slant, so I give that QB a small boost in his overal ratings. Link to comment Share on other sites More sharing options...
Mikie2times Posted November 30, 2005 Author Share Posted November 30, 2005 That is some awesome work, Mike. The ranking your system puts out for this season is dead-on with how one would rank the QBs based on observation. Really nice work. Especially this little nugget: Not surprising, because it's what we've all known for some time now. I'd like to see the individual rankings per catagory as compared to the way the league calculates them. How do they shake out using your system, i.e. % positive plays, etc. 516268[/snapback] Thanks for the kind words, here is the category %Positive Plays as it compares to completition %, If you would like to see others let me know. These aren't the actual values of the stat, just how each player ranked overall in both areas. In the NFL QB rating completition % makes up 25% of the overall value. Name----------------------------%Positive Plays--------------------COMP % Carson Palmer--------------------------1--------------------------------1 Peyton Manning-------------------------2--------------------------------2 Marc Bulger-----------------------------9-------------------------------- 3 Drew Brees------------------------------7--------------------------------4 Brian Griese-----------------------------5--------------------------------5 Daunte Culpepper----------------------15-------------------------------6 Drew Bledsoe----------------------------8--------------------------------7 Kurt Warner-----------------------------11-------------------------------8 Brett Favre-------------------------------6--------------------------------9 Matt Hasselbeck-------------------------10-------------------------------10 Trent Green------------------------------14------------------------------11 Steve McNair----------------------------12--------------------------------12 Anthony Wright--------------------------16-------------------------------13 David Carr-------------------------------20-------------------------------14 Jake Plummer----------------------------4--------------------------------15 Ben Roethlisberger----------------------13-------------------------------16 Tom Brady-------------------------------17-------------------------------17 Jake Delhomme-------------------------19--------------------------------18 Trent Dilfer-------------------------------21-------------------------------19 Donovan McNabb------------------------18--------------------------------20 Chris Simms-----------------------------24--------------------------------21 Byron Leftwich---------------------------22--------------------------------22 Mark Brunell-----------------------------23--------------------------------23 Michael Vick------------------------------3---------------------------------24 Joey Harrington-------------------------26---------------------------------25 Josh McCown----------------------------25---------------------------------26 Aaron Brooks----------------------------27---------------------------------27 Kerry Collins-----------------------------29--------------------------------28 Kyle Orton--------------------------------28--------------------------------29 Eli Manning-------------------------------31---------------------------------30 Gus Frerotte------------------------------32--------------------------------31 J.P. Losman------------------------------30---------------------------------32 Link to comment Share on other sites More sharing options...
Tortured Soul Posted December 1, 2005 Share Posted December 1, 2005 MDH- 50% of the normal NFL QB rating is made up of yards per attempt, and Completition %. Now I'm not going to argue that YAC success isn't influenced by accuracy, we all know QB's that deliver a well placed ball will allow receivers to gain additional yards. The problem is once a pass is either caught, or incomplete, the QB has no additional impact on the outcome of the play. 516599[/snapback] A couple of issues. First, this makes JP look even worse than the system already in place. Second, it's not that once a pass is either caught or not a QB has no bearing on it, it's once it leaves his hands. And how do you take into account a quarterback that's throwing to a Chad Johnson or a Hines Ward vs. a QB who throws whoever the Baltimore receivers are? I think there is only so much statistics can do in football before you need to say that the only way to compare how good players are is by watching them. Link to comment Share on other sites More sharing options...
Buffalo_Stampede Posted December 1, 2005 Share Posted December 1, 2005 A couple of issues. First, this makes JP look even worse than the system already in place. Second, it's not that once a pass is either caught or not a QB has no bearing on it, it's once it leaves his hands. And how do you take into account a quarterback that's throwing to a Chad Johnson or a Hines Ward vs. a QB who throws whoever the Baltimore receivers are? I think there is only so much statistics can do in football before you need to say that the only way to compare how good players are is by watching them. 516982[/snapback] Even watching a guy play really doesnt give you a perfect way of ranking QB's. Coaching and Play calling has alot to do with how a QB performs. For example, The Cowboys. Watch a Cowboy game and count how many times a WR or TE is running Wide open when he makes a catch. Then, watch a Bills game and count how many times a WR or TE is running Wide open when they catch the football. I just think the NFL is made up of a bunch of System QB's today, there arent many playmakers under center anymore. Link to comment Share on other sites More sharing options...
.thumb.png.1b7fb4086290c3afc001a7900b31fab0.png)
Recommended Posts