Tuesday, October 1, 2013

Tuesday Math

I’ve posted on a couple of Tuesdays this season in order to comment on everything that’s going on in CFB…I know that it’s hard for me to get the MMQ out on Mondays and this gives me a chance to digest and comment further on stuff going on in college football or Michigan in particular.

Today, if you that don’t like math or the rehashing of prior issues covered, you can stop reading now.  If you do, I’m going to try and shed a little light on Michigan’s offensive woes that may or may not make sense to you, but they make sense to me. 
For those that don’t know me, I have a hobby.  Poker.  I’ve been playing since I was in single digit age around the kitchen table but didn’t take the endeavor up seriously until I made my first trip to Vegas when I was 24.  Then, I couldn’t stop thinking about the game and all those probability and statistic courses I took in college really started coming in handy.  Believe it or not, math and how it relates to poker and football are somewhat parallel, especially when it comes to the execution of plays on offense.

Here’s the thing:  As poker has evolved from Fixed Limit to No Limit, good players have had to make adjustments in their game.  In Fixed Limit, calling with a set to a rivered straight or flush board at the end of the hand only cost you an additional big bet on a pot that was laying you say, 10 or 15 to 1, made sense and was correct.  No Limit, on the other hand, presents a completely different equation.  In No Limit, when the rivered flush or straight shows up and an opponent moves all in with his remaining stack, well, that call may now represent odds of only 3 or 4 to 1.  Meaning:  YOU HAVE TO BE RIGHT MOST IF NOT ALL OF THE TIME IN NO LIMIT when you make that call.  In Fixed Limit, you make the call, bitch about the other guy getting lucky, and move on.  If you make the call and your opponent was bluffing, that’s a huge payout that makes all of the other incorrect decisions okay.  In No Limit, when you make 2 incorrect decisions, your expectation for success goes negative.  So, in the end when you’re playing no limit…you better be right.
What if anything, does that have to do with offense in football at any level?  Years ago, the top tier college offenses played a lot like Fixed Limit poker. In other words, they simply ground out there game plan: Here’s our offense.  Here’s what we’re going to run.  You try and stop us.  And that worked because even though the defense would gear up to stop exactly what the offense was bringing, after 10 or 15 attempts, the better offense with better athletes would eventually get the defense to make a mistake and break a big one.   Or simply grind the opposing defense down.

Switch to today’s game.  Defenses you encounter on a week in-week-out basis are better as are the coaches coaching them.  So simply lining up and stating that we’re going to beat you man on man doesn’t work.  O-coordinators have to see what the defense is trying to stop, call the right play for that defense and then the offense has to execute that play to PERFECTION.
In other words, the O-coordinator could be doing his job perfectly.  Let’s give him a 100% on play calling.  Now, the offense gets the play call.  It’s a zone read option and the following things have to happen:  The line needs to play assignment football and block all of their positions correctly.  The QB needs to read what the line backers and safeties are doing and determine whether or not he wants to give the tailback the ball, tuck it and run it himself, or find the open receiver that the safeties are letting get open.  And everyone needs to execute the hand-off and run/pass and catch to perfection.  So, when you look at it mathematically, let’s put some variables on all those events and say that:

A – Play calling:  We hopefully have good plays for any given defense.
B– Line Execution =  This probability goes down when you have a younger offensive line that is still developing

C – QB decisions =  This can vary depending on the QB and the game/season he’s having.  I think it goes up with confidence, also, and good decisions beget good decisions and vice-versa.  Bad decisions lead to more bad decisions.
D.  Final Exectuion =  This is a result of combining all aspects of what can happen on any fundamental football basic:  Handing the ball off, hitting the correct gap that is being opened up, or passing and catching the ball, and not fumbling.

Now, anyone who’s studied probabilities knows that the outcome of an event or events with multiple probabilities is those probabilities multiplied for any given event.
When you look at the final equation:

Successful Play = Play Calling X Line Execution X QB Decisions X Overall Execution 

SP = A x B x C x D
So, let’s give those events some example numbers of what we think the current status of the offense is:

Play calling:  100% (I know, this might be a little generous in some of your eyes, but bear with me – hopefully in a minute you’ll see where this is going)
Line Execution:  90%  (Same thing…Just hang in there a little longer)
QB Decisions:  80% (I know, DG hasn’t made a right decision all season in some of your eyes)
Overall Execution:  80% (I know…Turnovers, INT’s, etc…..Hang on….)
SP = 100% X 90% X 80% X80% = 1 X 0.9 X 0.8 X 0.8=

57% success rate.  This figure I’m sure a lot of you will agree with. It could be even lower….and probably is.
Even with what appears to be pretty good probabilities for each individual execution can look awfully bad when everything is said and done at the end of any given play.  To make it even more simple, let’s say EVERYONE but the offensive line does their job and the offensive line screws up 50% of the time:

SP = 100 x 50 x 100 x 100 = 50% success.  Which isn’t bad, but what if this is the success rate on critical downs?
HOWEVER, everything I just listed above is actually not that simple and here’s where it can get tricky:  There is a SEQUENCE of these probabilities that MUST HAPPEN IN ORDER for any given play to have success.  In other words, we cannot look at the 2nd probability until the first event occurs, and so on. 
Therefore, play calling is absolutely critical followed by line execution, followed by QB decision making followed by final play execution.

If the early event is low probability, it impacts the outcome WORSE than everything else combined. 
So, if you’re not calling the correct plays and your probability of the called play being correct is 60%, then:

SP = 60% x 100% x 100% x100% = 60%  success rate.  And this gets worse if you’re players aren’t executing properly.  Which is why play calling is so critical in any game.  Followed by line execution which I think is where most of Michigan’s problems lie….but not all of them.  
So where’s the similarity to poker and why care?

Before, when you were “grinding” out plays in a running offense, the success rate didn’t have to be that high.  In today’s “No Limit” offense vs. defense environment, everybody has to do their job right and do it well.  Plain and simple.  Because now you are looking at 3 and 4 to 1 odds on any given play vs. a 15 to 1 grind it out “Breakaway” where the opposing defense screws up.  I’m not saying big plays don’t happen, I’m just saying its different now.

After re-watching parts of the Michigan U-Conn game again, I tried to figure out where things were going wrong.  In summary:  It wasn’t all on Devin.  Nor was it all on the offensive line.  Or the running backs.  And it wasn’t on the offensive coordinator calling the wrong play – but that may be a bigger part of it than what I had previously posted. 
It was proportionally on ALL of them, but you can hopefully now see that if any one part of the equation above breaks down…well, it can make for a long afternoon or evening. And it can look really bad when you add it all up….

Just my thoughts….
So how does any of this get better?  Watch what the defenses are taking away.  And then you can determine if we have the right play call going in.  I’m not sure we did at U-Conn as Michigan ran way too much on 1st down.  And I get that maybe the Michigan coaches we’re trying to do something against U-Conn based on film they saw and that just wasn’t possible.   That is, Michgan tried to revert to a, “Here we come.  Stop us”  mentality on offense.  And U-Conn did just that.  Michigan called 1st down runs way too often based on what the defense was taking away. 
Add to that some missed assignments and you can say whatever you want about Devin, but he’s suddenly in a world of “probability” hurt where even a high “Correct” decision probability on his part is going to get destroyed by the earlier outcomes. 

Hopefully, the coaches and O-Line have figured all of this out and Michigan comes out looking better on offense against Minnesota…if not…
Well, we’re in for a long season. 

No comments: