tag:blogger.com,1999:blog-38600807.post9079219868762569868..comments2009-02-13T15:20:13.725-05:00Brian Burkenoreply@blogger.comBlogger10125tag:blogger.com,1999:blog-38600807.post-60789280749241053702009-02-13T15:20:00.000-05:002009-02-13T15:20:00.000-05:00Brad Childress's and Jim Zorn's offenses are perfect examples for this sort of study.<BR/><BR/><I>...good offensive teams and bad offensive teams turn the ball over on roughly the same percentage of their non-scoring possessions.</I><BR/><BR/>In other words, the difference between Rex Grossman or Tavaris Jackson vs. Kurt Warner or Jason Campbell is that the good QBs produce more scoring drives (which never include turnovers or points). Thanks, but that much was obvious to me before. Why ignore the primary area of difference, though? I would have thought these articles more worthy of my time had I learned how much more often non-Grossmans/TJs succeeded compared to the Grossmans and other underachievers and what sorts of plays generated TOs for various teams. (Maybe a "safe", but predictable, 6-yard quick pass was not so safe after all, given the possibilities for deflections and fumbles in traffic?)<BR/><BR/>NSDs which don't end in a TO or punt are ended by the clock (which happens to all teams) or going for it on 4th down (usually because the team is behind). Taking those factors into account may have shown that the better teams do, in fact, commit fewer turnovers overall and on NSDs.<BR/><BR/>To really evaluate risk vs. reward in passing, one would need to account for the quality and frequency of the running game and for the defense. A QB with a running game and D of certain threshholds can afford to throw more risky passes. I think the types of passes attempted would be important, too. 3% INT's on passes of 10 yards or less would be atrocious, but perhaps not so bad on attempts of 40+ yds.<BR/><BR/>I would rather have learned if Jason Campbell had performed at or above the league average on scoring drives or not than something which tried to blame him for being too good.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-35105699120632497642009-02-13T13:46:00.000-05:002009-02-13T13:46:00.000-05:00One problem is that every team will be making a different trade-off on risk-reward. One way of checking it will be to see if when placed in a position where more risk is called for, such as being 10 points down (and controlling for the quality of the opponent), we see better outcomes and not just riskier outcomes.Steve Hnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-23010884168948412272009-02-13T09:19:00.000-05:002009-02-13T09:19:00.000-05:00Back onto the main article, it is an interesting question. Certainly there is reason to believe a 'Laffer curve' like effect with regards to risk-reward. Not enough risk and you won't score points, too much risk and you won't either.<BR/><BR/>The question of 'how do we measure risk?' is the important part. If we can accurately measure offensive risk, we should be easily able to determine an optimum risk level. <BR/><BR/>This could eat up a lot of your offseason research time.Ianhttp://www.blogger.com/profile/01518825067469269377noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-59024174889255463872009-02-13T08:48:00.000-05:002009-02-13T08:48:00.000-05:00This comment has been removed by the author.Ianhttp://www.blogger.com/profile/01518825067469269377noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-15191568252581213422009-02-13T04:16:00.000-05:002009-02-13T04:16:00.000-05:00JMM<BR/><BR/>I agree. I can't think of the amount of times I've seen defenders intercept third and long passes inside their 5 yard line and then celebrate like they've just scored a touchdown. In terms of field position they'd be far better off tapping it down and taking the result of the punt. <BR/><BR/>As the stats show, more often than not the team that's just been intercepted will be the next team to score.Ianhttp://www.blogger.com/profile/01518825067469269377noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-25505831660190785402009-02-12T19:49:00.000-05:002009-02-12T19:49:00.000-05:00Let's say a team has a very good defense, which hold the opposition to many short drives whether by preventing first downs or getting turnovers. This team will get more offensive possession per game, and per season. Assuming they play average on offense (in every way) they they should ahve more points and more interceptions than average. Perhpas the correlation between interceptions and points is just reflecting the incr. number of possessions?Tommy_Grandhttp://www.hotmail.comnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-42547139919517973092009-02-12T16:49:00.000-05:002009-02-12T16:49:00.000-05:00(OFF TOPIC but kinda related) <BR/>While I accept the conclusion that turnovers are bad and are correlated to losing, there is a part of me that thinks; "teams alternate possession of the ball. Why should it matter if it is a fumble, interception, punt or on downs? Field position is the only true impact." Recently when these thoughts begin, I think of James Harrison. <BR/><BR/>This leads me to the idea of a delta probability involved with a change of possession. If the JH 100 yd gallop was delta 13 pts and more importantly a change in the win probability. Then can't we look at all changes in possession in a similar way? An intercepted Hail Mary at game end might yield a delta of 0.<BR/><BR/>If we knew the importance of a change in possession to winning probability, then we can make a judgement on when and which incremental passes are justified?JMMnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-45873043854521983562009-02-12T12:17:00.000-05:002009-02-12T12:17:00.000-05:00Both good ideas. I forgot to mention--data is from all teams 2002-2007 regular seasons (n=192).Brian Burkehttp://www.blogger.com/profile/12371470711365236987noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-43768637380126315482009-02-12T11:29:00.000-05:002009-02-12T11:29:00.000-05:00It would be nice to see where each team falls.Greetingshttp://www.blogger.com/profile/18068932180257767964noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-88720663921488166662009-02-11T23:17:00.000-05:002009-02-11T23:17:00.000-05:00a ratio of punts to INT might be worth looking intoTom Ghttp://www.us.army.milnoreply@blogger.com