Articles from around the Web
PDF VERSION NOW DISCOUNTED OVER 30%
Click here to buy PDF version.
Click here to buy PDF version
Official Account: @fboutsiders
Scott Kacsmar: @FO_ScottKacsmar
Ben Muth: @FO_WordofMuth
Aaron Schatz: @FO_ASchatz
Vince Verhei: @FO_VVerhei
-- plus --
Bill Connelly: @SBN_BillC
J.J. Cooper: @jjcoop36
Cian Fahey: @Cianaf
Brian Fremeau: @bcfremeau
Tom Gower: @ThomasGower
Andrew Healy: @AndHealy
Rivers McCown: @RiversMcCown
Chad Peltier: @CGPeltier
Matt Waldman: @MattWaldman
Rob Weintraub: @robwein
03 Nov 2011
I keep meaning to write something about the outstanding rushing numbers for LeSean McCoy this season. Eagles blogger Brian Solomon was nice enough to do it for me.
Posted by: Aaron Schatz on 03 Nov 2011
25 comments, Last at
06 Nov 2011, 3:25pm by
Just a note that the article was written right before the Washington game. Since then, Reid has indeed unleashed McCoy: 58 times for 311 yards and 3 touchdowns.
His DVOA has gone down though.
McCoy is in the ideal situation for a running back. He has the best bootleg threat in NFL history to freeze linebackers, and a very credible deep game to keep the safeties back.
McCoy is well-suited for this situation because he has elite field vision. He has that magical ability to see 1.5 seconds into the future, so he knows what cuts to make to maximize his yardage.
Hmmm... The Force is strong with this one...
-Waves hand- This is not the running back you are looking for.
Honestly, I'd like them to leash McCoy just a bit and unleash Ronnie Brown/Dion Lewis. McCoy rarely takes a big hit, but I don't want him to get injured or fatigued with overwork.
What? The Eagles have Dion Lewis as well? What an affinity for Pitt runningbacks.
I agree! Please, keep McCoy leashed. It's all a mirage, Andy, and running the football is just plain boring! Show them how smart you are this weekend with a brilliant 85:15 pass:run gameplan.
A Giants fan.
Two problems with the conclusions of this article:
- As good as McCoy has been, Vick is (on a per-play basis) almost as succesful, and that his rushes are actually more valuable since most of them are bailing out what would otherwise be unsuccesful pass plays.
- Articles like this often fail to take into account that rushing success typically goes down, not up, if you rush more (ignoring second order effects like running more plays). Strategically speaking, normally the correct strategy is to rush the ball with probability such that rushing and not rushing have equal expected return. I doubt that PHL is experiencing greater expected return on (called) rushing plays than (called) passing plays right now (other than perhaps short-yardage situations).
Strategically speaking you should mix rushes and passes to maximize total expected return not try to balance them.
The two are equivalent because value is maximized by a Nash equilibrium, for which the expected payoff of each strategy (one in the degenerate case of a pure equilibrium, multiple in a mixed) is the same.
I'm no economist, but I'm pretty sure the Nash equilibrium doesn't apply to this because of the opposing defense. You can force your opponent to change his strategy and increase your own expected return.
Nash equilibrium does apply here. For any given situation, both offence and defence choose what to do. You cannot force the defence to do anything, but they will of course modify their strategy according to your strategy. The stable solution to this is an impure strategy for each side (e.g., run 40%, pass 60% on first and 10 would be a typical component of an offensive strategy).
From Wikipedia "If each player has chosen a strategy and no player can benefit by changing his or her strategy while the other players keep theirs unchanged, then the current set of strategy choices and the corresponding payoffs constitute a Nash equilibrium."
I don't think such a thing can exist in football. Once the defense has chosen it's strategy, it is quite possible to gain a benefit by changing your own (or vice versa).
This just means that there is no Nash equilibrium in pure strategies; it doesn't inherently preclude a Nash equilibrium in mixed strategies (that is, some mix of runs and passes). The problem more lies in the fact that formations communicate some information about what plays might be run, combined with the fact that not all running plays are alike and not all passing plays are alike, and added to the fact that, well, forming accurate expectations is not a trivial task.
It's not like rock-paper-scissors. The problem is that it's not just a choice between "run" and "don't run," and furthermore your opponent can gather information from your formation and personnel package and you can gather information from his.
Football isn't quite like a repeated game, so you shouldn't necessarily have the average payoff be the same for all play selections. Only the marginal payoff should be.
For example, suppose you have a plodder back and a brilliant back. Let's call them "Thomas" and "Jamaal," to pick two names entirely at random. Jamaal's yards per carry average is 7 yards per carry for his first 10 carries, 5 yards per carry for his next ten carries, and 3 yards per carry for carries 21 and up. Thomas averages 3.5 yards per carry for any number of carries.
In equilibrium, you should expect Jamaal to average 6 yards per carry on the season, and Thomas to average 3.5, and that's completely fine, because the marginal additional carry by Jamaal is equal to the marginal additional carry by Thomas.
When these sorts of effects are happening, it can be confusing to figure out whether the game theory is actually being done right by the coaches.
Just wanted to note that this actually is not true. As Braess's Paradox shows, Nash equilibriums are not necessarily optimal.
Not all Nash equilibria are optimal, but the optimal strategy will (in this case) be an equilibrium.
I like this nerd analysis of football. But if my Dolphin's coach Tony Sparano were to read this, his head would explode. He only knows four options: run, don't run, punt or field goal fist pump.
Of course Vick is more succesfull than McCoy. The playfakes that Vick gets on his scrambles are obviously way better than the ones McCoy gets. In Vicks case the QB ACTUALLY HAS THE BALL. This forces every defender to defend the pass first- often for three full seconds. When the DBs have been drawn deep by the WRs, and the LBs have dropped into their zones, then Vick can take off.
Plus, as you mentioned, Vick is often turning broken plays into big plays. Very nice when trying to win, but from a playcalling perspective, pretty uninteresting - you can't call a broken play in the huddle.
Marginal value of a pass and rush play need to be the same for it to be optimized, not average/expected. Firstly, when the marginal values are equal, the average pass play is usually greater than the average rush play because of a combination of the distribution being positively skewed and that pass plays have greater variance. This intuitive outcome is verified when you look at the yds/attempt for a QB versus the yds/carry for the RB on the same team. Secondly, the marginal value of a type of play is extremely hard to calculate because there are an insane number of variables to consider, so it's a very tough call to say whether Vick or McCoy has been more "successful."
Really, the goal should be to give McCoy the fewest possible carries needed to pull off a win. The more carries McCoy has the more he gets banged up and the more likely he is to get injured. Therefore, if Philly is winning the goal should be to minimize McCoy's carries
Can we please start referring to LeSean as "the real McCoy" and Colt McCoy as some class of "fake McCoy"??
Now you're just playing coy.
If we did that, then we'd have to start separating the Adrian Petersons, Mike Williams and Mannings of the football world with catchy little nicknames too.
The Vikings need offensive line help, while the Bears, Lions, and Packers have significant defensive concerns.
See All XP | NFL XP | College XP
© Football Outsiders, Inc. // Site powered by Stein-Wein // Partner of USA TODAY Sports Digital Properties