My bets for the season

[KillerGoose]

(07-01-2022, 08:16 AM)SHRacerX Wrote: Right, what I meant when I said that was that when Stafford posted that 60 QBR against the Bengals in the Super Bowl, that was better than Mahomes, Tannehill, or Carr posted in the playoffs against Cincy.  

Absolutely. A 60 QBR very good. QBR is basically an altered EPA statistic. 

[*SHRacerX*]

(07-01-2022, 09:31 AM)KillerGoose Wrote: Absolutely. A 60 QBR very good. QBR is basically an altered EPA statistic. 

Yes, I don't care for it.  Puts way too much stock in things like last minute drives, etc.  I remember seeing a game where Kitna carved up a team, but the opposing team had a better QBR because of mindless stats at the end of the game in a blow out.  

[KillerGoose]

(07-02-2022, 09:35 AM)SHRacerX Wrote: Yes, I don't care for it.  Puts way too much stock in things like last minute drives, etc.  I remember seeing a game where Kitna carved up a team, but the opposing team had a better QBR because of mindless stats at the end of the game in a blow out.  

I actually don't mind it. It catches a ton of heat due to the unknown methodology, but it does a pretty good of identifying which QBs are playing the best. The overall goal of a metric like that is to try to gauge the players impact on winning and it has a significantly better correlation to team success than passer rating does, for example. I looked into it at one point this off-season, and roughly 72% of the variation in win percentages for QBs across the league could be explained by their QBR. That's pretty damn good. For reference, passer rating could explain 52% of win percentage variations. So, it does a better job of identifying QBs that are valuable to their team than passer rating does. The only other thing that I know of which will correlate that strongly is just good ole EPA, which isn't a QB statistic but a play-by-play statistic, tell you the value of the play. It can be attributed to WRs, QBs, RBs. 

It isn't perfect by any means, but I have seen enough in the past to say that it is probably better than passer rating to evaluate QBs. People don't like ESPN, though, and they find the results unintuitive to what they expect, so I highly doubt it will ever be referenced much. 

[fredtoast]

(07-02-2022, 10:19 AM)KillerGoose Wrote:  The overall goal of a metric like that is to try to gauge the players impact on winning and it has a significantly better correlation to team success than passer rating does, for example. I looked into it at one point this off-season, and roughly 72% of the variation in win percentages for QBs across the league could be explained by their QBR. That's pretty damn good. For reference, passer rating could explain 52% of win percentage variations. So, it does a better job of identifying QBs that are valuable to their team than passer rating does. 

I like the QBR tries to incorporate QB rushes, sacks, and fumbles but it just doesn't work

The best QB does not always win.  So you should not measure QB play by who wins.  For example, a QB on a good running team will have more manageable third downs to convert and QBs on the best overall teams will have more comeback wins.

The best QB does not always win.  So QB rating should not be judged on how it correlates with winning.