tag:blogger.com,1999:blog-38600807.post1538873373483547120..comments2023-11-05T04:16:44.937-05:00Comments on Advanced Football Analytics (formerly Advanced NFL Stats): Single-Point-Failure Model of the Passing GameUnknownnoreply@blogger.comBlogger11125tag:blogger.com,1999:blog-38600807.post-12701176018228867832010-09-06T13:32:13.990-04:002010-09-06T13:32:13.990-04:00This corroborates Parcells' team building appr...This corroborates Parcells' team building approach. Whenever Parcells inherited another terrible team, he would immediately sign a bunch of veteran players who were known as "hold the fort" players. These signings weren't popular with fans, as it often seemed like Parcells was overpaying mediocre players just because he was comfortable with them, but the idea was to remove matchup weaknesses as quickly as possible. It allowed each of his teams to be competitive almost immediately.<br /><br />It should be noted, however, that while the basic theory here is perfectly sound, it doesn't follow that you are always better off with average players across the board, as you need to match up on a position-by-position basis. A defense is most likely to put its biggest matchup problem at the blind side pass rushing position, which means that you are more likely to need an above-average player there to cancel out the advantage.Seannoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-85118102338855689912010-01-23T14:40:35.167-05:002010-01-23T14:40:35.167-05:00The tail end of your article reminds me of usually...The tail end of your article reminds me of usually successful running offenses failing more than one would expect when they go to jumbo packages. In essence, the offense is creating a larger chain that needs to hold and prevent defenders from getting to the ball carrier before the vital one or two yards are obtained. Recently, we've seen offenses adjust by putting stronger links in the backfield as lead blockers, such as Dockett and Branch for Arizona against Green Bay and Cody for Alabama.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-43773544644953544202010-01-11T18:31:08.280-05:002010-01-11T18:31:08.280-05:00Your premise would lead coaches to think that a st...Your premise would lead coaches to think that a strong running game should lead to more wins than a strong passing game and yet in the NFL it is the opposite. The reason is that while having many more links in the chain leads to a higher rate of failure, it is also harder to defend. If you know who has the ball and there can only be one, for the most part the defense can focus all of its attention on that one player. The passing game mitigates this advantage by shrouding who will get the ball and when at the cost of a higher risk of failure. Now the question becomes whether the risk of failure outweighs the chance of success. The answer is of course, "It depends".<br /><br />Further, unless you are playing single wing, a significant portion of your players are not involved in a running play. For example, the receivers on the opposite side of the ball, the QB after they hand off (if they hand-off), the opposite side tackle and guard (unless they pull) etc. In a pass play, far more players are involved in the play.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-44533392158371430822009-01-13T13:06:00.000-05:002009-01-13T13:06:00.000-05:00you've just opened up a serious rabbit hole. God b...you've just opened up a serious rabbit hole. God bless anyone who tries to model this properly (including all the unknowns others have mentioned). I'd speculate that although increasing the number of WRs adds additional failure opportunities for the defensive secondary it may also have a negative impact on the success rate for the WRs (the field is more cluttered, WRs have less room to maneuver / QB has more information to process before making a decision). It is highly unlikely that these effects would cancel out, and I have no idea which would be weighted more heavily. <BR/>Play calling is also huge here, a play with a go-to target decreases the pr a QB will recognize a 2nd, 3rd,...,nth receiver even when they've won their X% of point matchups.<BR/><BR/>Totally meaningless aside: in madden i rarely run more than 2 WR sets because I tend to throw more INTs when there are more DBs on the field...Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-88321125886994198872009-01-08T17:55:00.000-05:002009-01-08T17:55:00.000-05:00It gets even more complicated regarding max protec...It gets even more complicated regarding max protection schemes. Remember that when an eligible receiver blocks in pass pro, he remains a threat to catch a pass because he could block and release. Defenses still have to account for him. [In some ways, this is a tactical counter to the zone blitz by the defense.]<BR/><BR/>The chip and the delayed release provide blocking help while still requiring defensive coverage.<BR/><BR/>s.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-34716661341431884942009-01-07T14:27:00.000-05:002009-01-07T14:27:00.000-05:00I would also think that the large number of player...I would also think that the large number of players involved in determining the success/failure of each play would be a major contributor to the perception of "parity" in the NFL.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-36679483317160436172009-01-07T14:20:00.000-05:002009-01-07T14:20:00.000-05:00Can you imagine a football team with a starting pl...<EM>Can you imagine a football team with a starting player who is a point-failure in nearly every play? He'd be a lineman who always gets beat by a pass-rusher or a defensive back who always gets beat by a receiver. It would be ugly...</EM><BR/><BR/>It sure as heck <EM>was</EM> ugly -- last year at LG for the Jets as they plugged various nobodies into the void after they got rid of Kendall for having a bad attitude, without bothering to first get a replacement for him.<BR/><BR/>This is not a thought experiment any more.<BR/><BR/>Look at Thomas Jones's stats last year and this.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-46739639348710325492009-01-07T12:33:00.000-05:002009-01-07T12:33:00.000-05:00that post sat in the drawer?I thought that was one...that post sat in the drawer?<BR/><BR/>I thought that was one of the best post of the year<BR/><BR/>I think if you knew what plays were being called you could come up with very accurate effectiveness rating for players, obviously we don't have that info thoughPhilhttps://www.blogger.com/profile/15837333926742421707noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-78317441921199369812009-01-07T11:09:00.000-05:002009-01-07T11:09:00.000-05:00you guys are way smarter than me but I enjoy the w...you guys are way smarter than me but I enjoy the work that is done on this site for more than just the unbelievable edge it provides me in my office pool :) Keep up the good work!Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-38600807.post-62184848368019901052009-01-06T23:55:00.000-05:002009-01-06T23:55:00.000-05:00Chris-Thanks. I saw your post earlier and it motiv...Chris-Thanks. I saw your post earlier and it motivated me to dust this off. I wrote this a few weeks ago, and thought it wasn't terribly interesting once I got it down on paper. <BR/><BR/>But you're right. I imagine 90% of football tactics is breaking free of the 1-on-1 match-up somehow and getting an advantage--saturating a zone, an overloaded pass rush, or a double-team on a run block.<BR/><BR/>Incidentally, here is a <A HREF="http://www.fangraphs.com/blogs/index.php/bonds-and-a-bat/" REL="nofollow">case in point</A> from a Fangraphs post about Barry Bonds. His '01/'02 seasons were so good, you could literally put a papaya in the lineup to hit behind him, and you'd come out with a replacement-level duo. In other words, you could bat <I>me</I> behind Bonds, and the Giants would have won 70+ games each year.<BR/><BR/>But if you put me (or a papaya) in at right guard for New England last year, they'd go 0-16 instead of 16-0.Brian Burkehttps://www.blogger.com/profile/12371470711365236987noreply@blogger.comtag:blogger.com,1999:blog-38600807.post-16508870467457085082009-01-06T21:39:00.000-05:002009-01-06T21:39:00.000-05:00This is a good article, and it's funny you posted ...This is a good article, and it's funny you posted this since just today I posted an article about the difficulties with modeling football (and the resulting easy-ways-out certain statisticians seem to take). <BR/><BR/>In any event, though I agree that a great deal of the game could be modeled in this way, as a series of single-point models, some of that gets more complicated for plays like options/reads and certain pass against zones. Good offenses look for 2 on 1s and 3 on 2s. It's not always a blown assignment that leads to a completion: the correct pass combination (or receiver read to get into it) should put a defender in an untenable bind -- do I cover the guy in the flat or the curl? <BR/><BR/>And option-run game football is designed to avoid single-point matchups, since you look to get (a) double team blocks at the point of attack, which then frees up your runner and QB to (b) get a two on one read to option off a defender. <BR/><BR/>Of course, your point is still valid and interesting, and in many ways it just reinforces what a team sport football is. Keep up the great work.Anonymousnoreply@blogger.com