Jump to content


Photo
- - - - -

An Interesting Note On Cam From My QBA Research


  • Please log in to reply
10 replies to this topic

#1 fieryprophet

fieryprophet

    WARNING: Do not annoy!

  • ALL-PRO
  • 5,273 posts
  • LocationLa Grange, NC

Posted 27 December 2012 - 12:12 PM

As I've mentioned before, I've been working hard on developing my own quarterback metric that I feel is more accurate of a quarterback's performance than passer rating.

So far, I've managed to both simplify the formula and improve it's accuracy to over 5% more accurate than passer rating, averaging an 84% percentage of games correctly predicted.

What is interesting about all of this is that it shows that Cam has had 5 losses in his career despite having the higher QBA, among the highest of all quarterbacks in my dataset.

When you see which games they are it makes you sad, because we know them well:

http://50.116.44.142...nd=all&type=all

In other notes, I think I'm just about finished with the formula. It has accurately predicted 39 more games over the last 3 seasons than passer rating, which is usually the gold standard for winning metrics. I also think it does a great job delineating between quality quarterbacks and subpar ones.

And before anyone asks me about Matt Moore's ranking, he has one game in the database, and if he a 70 QBA average over a full season it would be terrific but for a single game it's merely a notch above average.

#2 cultclassiccat

cultclassiccat

    flash

  • ALL-PRO
  • 1,539 posts
  • LocationG spot

Posted 27 December 2012 - 01:45 PM

calvin and hobbes, you never dissapoint.

#3 fieryprophet

fieryprophet

    WARNING: Do not annoy!

  • ALL-PRO
  • 5,273 posts
  • LocationLa Grange, NC

Posted 27 December 2012 - 02:48 PM

calvin and hobbes, you never dissapoint.


Thanks. I just added spikes to the formula and it has pushed the overall accuracy to 84.92%. Also, I just noticed that Andrew Luck has 5 wins this year where he had the lower QBA. Luck indeed.

#4 Fiz

Fiz

    SENIOR HUDDLER

  • HUDDLER
  • PipPipPipPipPipPipPipPipPip
  • 11,057 posts

Posted 27 December 2012 - 03:45 PM

Link is broken.

#5 fieryprophet

fieryprophet

    WARNING: Do not annoy!

  • ALL-PRO
  • 5,273 posts
  • LocationLa Grange, NC

Posted 27 December 2012 - 04:20 PM

Link is broken.


Works for me. I was running an update earlier, it may have froze for a minute.

#6 2jakefansinva

2jakefansinva

    Coach Jack Bolton

  • HUDDLER
  • PipPipPipPipPipPipPipPipPip
  • 4,735 posts

Posted 27 December 2012 - 05:12 PM

As I've mentioned before, I've been working hard on developing my own quarterback metric that I feel is more accurate of a quarterback's performance than passer rating.

So far, I've managed to both simplify the formula and improve it's accuracy to over 5% more accurate than passer rating, averaging an 84% percentage of games correctly predicted.

What is interesting about all of this is that it shows that Cam has had 5 losses in his career despite having the higher QBA, among the highest of all quarterbacks in my dataset.

When you see which games they are it makes you sad, because we know them well:

http://50.116.44.142...nd=all&type=all

In other notes, I think I'm just about finished with the formula. It has accurately predicted 39 more games over the last 3 seasons than passer rating, which is usually the gold standard for winning metrics. I also think it does a great job delineating between quality quarterbacks and subpar ones.

And before anyone asks me about Matt Moore's ranking, he has one game in the database, and if he a 70 QBA average over a full season it would be terrific but for a single game it's merely a notch above average.


pretty work, Dude.

#7 weymouthst

weymouthst

    #1 zero

  • HUDDLER
  • PipPipPipPipPipPip
  • 1,041 posts
  • LocationMaine

Posted 27 December 2012 - 06:20 PM

omg....Cam has the ability to play well, that is all the metric I need to hear. What is the point of all these numbers when you can see with your own eyes whether a QB is playing well or not? You can do all the analysis you want, Cam is a good quarterback and is mobile, and capable of dynamic plays. Every passer can have a bad game, trying to predict it and what not is ridiculous.

#8 fieryprophet

fieryprophet

    WARNING: Do not annoy!

  • ALL-PRO
  • 5,273 posts
  • LocationLa Grange, NC

Posted 27 December 2012 - 06:24 PM

omg....Cam has the ability to play well, that is all the metric I need to hear. What is the point of all these numbers when you can see with your own eyes whether a QB is playing well or not? You can do all the analysis you want, Cam is a good quarterback and is mobile, and capable of dynamic plays. Every passer can have a bad game, trying to predict it and what not is ridiculous.


It's not about predicting future play, it's quantifying individual performances to try to show why a particular QB won a game. Or if they didn't, why they still played well enough to win, as in Cam's case, but the team around him didn't uphold its side of the deal.

#9 pstall

pstall

    Gazebo Effect

  • HUDDLER
  • PipPipPipPipPipPipPipPipPip
  • 20,980 posts
  • LocationMontford

Posted 28 December 2012 - 02:23 PM

OUTSTANDING work man.

#10 fieryprophet

fieryprophet

    WARNING: Do not annoy!

  • ALL-PRO
  • 5,273 posts
  • LocationLa Grange, NC

Posted 28 December 2012 - 03:22 PM

OUTSTANDING work man.


Thanks! I put a lot of time and research into this.

#11 ncsu12engr

ncsu12engr

    MEMBER

  • HUDDLER
  • PipPipPipPip
  • 234 posts

Posted 28 December 2012 - 03:34 PM

I really like what you have done with this. You have confirmed my pick for offensive rookie of the year as Russell Wilson.


0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users

Shop at Amazon Contact Us: info@carolinahuddle.com