FanPost

FMQB: What dog training can teach us about coaching

This week I’m too distraught to re-watch and write about the game. Instead you get my ramblings on a few topics that interest me.

Binominal Formula update

Last week I introduced everyone to the binomial formula which can tell us the chance that an event will occur exactly K times out of N. This is handy if we want to know things like what the chance is that Dallas will win at least 8 of their last 11 games (and finish 9-7). Sadly, the update isn’t very optimistic. There are two significant changes. First, last week I estimated that Dallas would have a 60% probability of winning each of their remaining games. That was too optimistic. Even if Dallas is a top 5 team in the NFL (which is what the efficiency statistics suggest), home field advantage counts for more than I understood. For example, according to the efficiency models last week Dallas only had a 48% probability of beating Minnesota even though Dallas was ranked #5 and Minnesota was ranked #16. Using this week’s efficiency rankings my rough estimate of the win probability for Dallas’s remaining games is:

Team

Win Probability

NYG

50%

Jax

80%

at GB

40%

at NYG

30%

Det

70%

NO

70%

at Indy

30%

Phi

60%

Wash

60%

at Ari

50%

at Phi

40%

Average

53%

Dropping 53% into the binomial formula (do not try this at home! The binomial formula requires a constant probability so taking the average probability is technically wrong, but I’m trying to give a sense for Dallas’s chances so here's my scientific wild a** guess) gives Dallas a 16% chance of winning at least 8 of the 11 remaining games. A 9-7 record gives you a 50% chance of making the playoffs. So it looks like Dallas has an ~8% chance of getting into the playoffs.

That leads into my next point.

Last year after the loss to Minnesota I wrote that I thought that many people were reaching the wrong conclusion from the Minnesota game. Instead of concluding that the OL needed to be revamped, my thinking was that Dallas really needed to be playing that game at home, and the key to getting a home playoff game was improved pass defense. After the loss on Sunday I feel better about that analysis for the following reasons:

-Dallas dominated the game statistically and only lost because of multiple costly errors (with at least 1 interception likely a direct result of crowd noise). This supports the claim that Dallas would have won had they played the playoff game at home

-An even greater appreciation of home field advantage. As explained above, even though Dallas has played better than Minnesota so far in 2010, Sunday was a 50/50 bet because of Minnesota’s home field advantage. That drives home for me how critical it is to win home games.

Which leads to my next point: in my opinion a bad pass defense has cost Dallas both home games. Below is a table of ANPY/A allowed (a measure of passing efficiency) for each game this year.  

Team

ANPY/A Allowed

Washington

 4.9

Chicago

 11.0

Houston

 3.4

Tennessee

 8.3

Minnesota

 4.0

 

For comparison, here are the statistics across the league from 2009.

Defense ANPY/A Allowed

Min        3.3

Median  5.4

Max       7.3

 

As you can see in both home games the passing defense was terrible. Dallas cannot lose home games. The NFL is too competitive. And bad pass defense doomed them in both home games.

 

Here’s more. The table below (via Advanced NFL Stats) summarizes Dallas’s ranking in the predictive statistics for the 2010 season. Keep in mind that offensive pass efficiency is the most important factor (about 3-4x the importance as running) and defensive pass efficiency is the second most important metric.

TEAM

Offensive Pass Efficiency

Offensive Run Efficiency

Offensive Int. Rate %

Offensive Fum. Rate %

Defensive Pass Efficiency

Defensive Run Efficiency

Defensive Int. Rate %

Penalty Rate %

DAL

4

19

19

1

16

17

30

33

 

The table tells you exactly what is wrong with Dallas. Dallas is bad in terms of penalties and forcing turnovers. The passing offense is elite but that alone is not enough to overcome the other weaknesses. Throw in some bad luck, in the form of bad special teams play (special teams plays are impossible to predict … they’re essentially random) and you’ve got 1-4.

Dallas and New England: Compare and Contrast

Other than the Cowboys, the team I pay the most attention to is New England. Don’t take that to mean I like New England, I don’t. However, New England is consistently successful so I’m interested in how they achieve their success.

First, New England is one of the most analysis driven teams in the NFL. You see this from their play calling (going for it on 4th down in their own territory against Indy last year) to their scouting and approach to the game. Here are some comments from an article on Belichick and his coaching tree.

Each one is a football dissertation, with a two- or three-month deadline. Mangini was once asked to analyze the fumble history of every free agent running back -- how often they lost the ball, from which arm, when hit by which kinds of players. As a coaching assistant in 2002, McDaniels was told to scout 35 defensive backs, compare them to free agent DBs, then recommend candidates to rebuild the secondary .…

The most successful projects -- and the rarest -- answer questions Belichick never thought to ask. When McDaniels and Caserio were tasked with scouting New England's opposing receivers and corners, they took it further, creating a formula for size, speed, ball skills, scheme and assignments (keep this in mind!!). Then they developed a ranking system based on the Patriots' personnel and playbook. Impressed, Belichick ditched his manual for scouting receivers and corners, replacing it with what he'd learned.

Here's Belichick with his coaches, analyzing each play from the previous game and comparing how many times it was practiced with how many times it was called on Sunday and how many times it was executed correctly.

Here’s more from an Eric Mangini profile.

Mangini is determined to be as numbers-driven as possible. Yes, numbers sometimes mislead, as he has learned, but not as often as gut feelings do.

This afternoon, Mangini asks Steckel to fish out the "AAR" binder, which contains After Action Reports, what the military calls feedback reviews. Last year, Mangini periodically asked every department -- coaches, personnel, equipment, video, medical, strength and conditioning -- to fill out AAR forms …

And

In formulating today's practice, Mangini wanted to focus on situational football. But because the Patriots didn't do much of that during OTAs when he worked in New England, he reviewed charts from every possible game scenario last season and checked how often each was repped. For example, the Browns faced second-down situations needing seven to 10 yards from the "high red zone" area (between the 12- and 20-yard lines) 7 percent more often than they practiced them. So Mangini shifted practice to include 7 percent more of those reps. He also analyzed how best to use those extra reps, telling Daboll to practice more "22" formations, with two backs and two tight ends, and asking defensive coordinator Rob Ryan to design specific coverages to give the offense the proper look.

Ok, so the Patriots are data driven. What are they doing? Here are a few observations.

1. The Patriots are running the ‘death by 1,000 paper cuts’ offense

The Patriots wasted no time turning back the clock in the wake of the Randy Moss trade … Namely, they instantly re-introduced the world to the famous "death by 1,000 paper cuts" offense that they rode to three Super Bowl titles from 2001 to 2004.

2. Belichick has gone against the grain with his WR personnel (remember the McDaniel study!). All the Patriots WRs fit the same mold now. Small, fast, good cutting ability in short spaces. Consider:

-The NFL loads up on big WR (Calvin Johnson, Moss, Fitzgerald, Andre Johnson, Sims-Walker, Vincent Jackson, Malcolm Floyd, Brandon Marshall, TO)

-NFL defenses respond with big CBs (Sean Smith, Antonio Cromartie, Woodson, DRC, Ashomugha)

-Belichick loads up on small quick WR (I remember the Smurfs and Anthony Carter so I've seen this before)

 

Let’s put it this way. Antonio Cromartie played Moss well but I’ll be very curious to see how he matches up on Branch or Welker the next time they play.

It’s of special interest to Dallas fans because of Dez. There’s a saying in poker ‘if you can’t spot the sucker at the table it’s probably you’. That’s often used as cautionary advice in trading as well: always remember if you’re buying there’s someone selling who thinks the price isn’t going higher (i.e. make sure you’re not the sucker). I find it interesting that Belichick would pass on Dez. I’m not in saying that I think Dez was a bad pick, it’s just interesting that Belichick would pass on Dez.

3. The Patriots RBs are a free agent cast-off and Green-Ellis. It reminds me of this comment from Brian Burke.

If I were advising a general manager, I’d tell him to largely forget about the run. Get a running back who’s good at picking up blitzes or catching the ball. Never draft a running back in the first few rounds, and whatever you do, don’t waste precious cap space (or payroll budget) on him.

4. The Patriots typically do not spend big dollars on CBs.

What are the odds on the Patriots -- a team that has refused to spend much more than the league minimum on cornerbacks since Bill Belichick arrived -- spending a first-round pick on a cornerback?

Below is table from the SI NFL season preview summarizing the franchise salaries for the various positions.

Position

Franchise Salary (Top 5)

$M

Quarterback

14.65

Cornerback

9.96

Wide Reciever

9.88

Defensive End

8.99

Offensive Line

8.45

Linebacker

8.3

Running Back

6.62

Safety

6.34

Defensive Tackle

6.06

 

My hunch is the Belichick is playing money-ball … that Belichick concluded that if you have $20M to spend on your secondary it's better to spend $12M on two elite safeties and $8M on two average CBs than to spend $18M on two elite CBs and $2M on a free agent castoff and converted 6th round CB.

 

5. The Patriots draft picks (I’ve included picks in round 1-3 only). They used their top picks on 2 safeties, an ILB, and a DB. Overall they choose 3 DB, 2 S, 2 ILB, 2 OLB, 2 WR, 1 TE, 1 DL, 1 OT, 1 QB.   

2010

Overall Pick

Player

Position

College

27

Devin McCourty

DB

Rutgers

42

Rob Gronkowski

TE

Arizona

53

Jermaine Cunningham

OLB

Florida

62

Brandon Spikes

ILB

Florida

90

Taylor Price

WR

Ohio

2009

34

Patrick Chung

S

 Oregon

40

Ron Brace

 DL

 Boston College

41

Darius Butler

 DB

 Connecticut 

58

 Sebastian Vollmer

 OT

 Houston

83

Brandon Tate

 WR

 North Carolina

2008

10

Jerod Mayo

ILB

Tennessee

62

Terrence Wheatley

DB

Colorado

78

Shawn Crable

OLB

Michigan

94

Kevin O’Connell

QB

San Diego State

2007

 

 

 

24

Brandon Meriweather

S

Miami (FL)

 

My observation would be that the Patriots have made sure they can take care of the middle of the field with good ILBs and safeties. Dallas has almost the opposite strategy with elite CBs and OLBs and relatively weak ILB and safeties.

Dog training and penalties

Normally I like to stick to numbers and hard facts. However, today I’m going to discuss a softer subject, the psychology of coaching.

In terms of psychology, dog training has many useful lessons (because it’s almost pure application of the behavioral psychology principles developed by BF Skinner). The most important lesson is that positive reinforcement is more effective than positive punishment. First, let’s define the 4 quadrants of operant conditioning

Positive Reinforcement – adding something to make a behavior happen more often.

Negative Reinforcement – taking something away to make a behavior happen more often.

Positive Punishment – adding something to make a behavior happen less often.

Negative Punishment – taking something away to make a behavior happen less often.

The casual fan is entirely focused on positive punishment (e.g. cut the player, bench the player, fines, discipline, etc.). However, dog training reminds us that there are a lot of drawbacks with positive punishment

Historically, we humans are controlled primarily through negative reinforcement. We're punished when we haven't done what is reinforcing to those who are in "charge" (such as parents, employers, or trainers). Positive reinforcement has, unfortunately, been less often used, but it is more effective than negative reinforcement and has many fewer unwanted by-products. 

 

For example, a student is punished when he doesn't study. He may study after that, but he may also stay away from school (truancy), vandalize school property, attack teachers, or do nothing. With positive reinforcement, the student would have been reinforced for studying in the first place, and most likely would have learned to love learning….

 

I highly recommend that you stay away from it. There are too many toxic side effects. The timing has to be perfect (if you punish the dog for soiling the house after she comes to you, you have just punished her for coming to you), and punishment is associated with the person doling it out.

 

Oftentimes, you only suppress behavior using positive punishment—sure, the dog stops urinating on the carpet, but now does it behind the couch. For some dogs, barking is so self-reinforcing (they like to bark) that they'll bark regardless of what you do, or bark only when the shock or citronella collar is off. The punishment has to outweigh the rewards and the motivation for it to be effective.

 

Depending on the motivation, some people (and dogs) will go to great lengths to succeed over adversity, continuing to practice their success—seeking behaviors regardless of punishment or hardships.

 

One of the reasons I like Wade is that he’s very positive. In general, I haven’t seen coaches who use a lot of punishment be successful with professional athletes. Think of the most successful coaches.

 

-Bill Walsh

-Joe Gibbs

-Bill Belichick

-Jimmy Johnson

 

The first 3 were level headed coaches. I wouldn't say they were known as player's coaches. They were liked by their players and demanded professionalism and maturity. None were known as disciplinarians.

 

Jimmy Johnson is an interesting case. Jimmy Johnson is remembered as a ‘fear-of-god’ coach, right? Well to certain players he was, but not everyone. Remember, Johnson was notorius for having favorites and not punishing his star players. Johnson was liked by his star players and you can bet it wasn’t because he was constantly punishing them. I’d guess that he used a lot of positive reinforcement with his stars (e.g. letting them break rules, stroking their egos, etc. ... Johnson did major in industrial psychology after all) and/or negative reinformement (i.e. he ran a strict team but the stars got rewarded by being excused from the rules). The guys he didn’t like? I'm not sure Johnson did get them to improve with punishment (e.g. Curvin Richards never did stop fumbling), he just cut them. And that was a luxury Johnson had because he had such a great eye for talent. Unless you have Johnson's eye for talent I'm not sure cutting players is a good strategy.

  

So when I see the DB’s doing push-ups for missing interceptions (which btw they’ve been doing for 2 years now with no increase in interceptions) , read about fines, hear that Wade is sending players to the locker room for messing up in practice (sorry I can’t find the link), I can’t help but think that they’re taking the wrong approach.

 

In fact, I feel bad for these guys. They obviously want to succeed (Jenkins is on national TV doing pushups) but they don't know how. I consider it an indictment of the coaching staff that they allow it because it suggests the coaches don't have better ideas. Its sad watching these players groping for something that works with apparently little guidance from the coaches.

 

Let’s say you want players to stop celebrating in the end zone. Here are two approaches

-Fines for all players who celebrate in the end zone, or

-Have practice officials at practice, have the players hand the ball to the official after touchdowns, and intermittently praise the players for their decorum.

 

Which approach do you think is likely to be more effective in forming the habit you want?  

 

Still not convinced? I know the penalty was on Hurd but it easily could have been on Austin. Read Austin’s description of what he was feeling at that moment.

"I was just kicking myself until 4 in the morning," said Austin, who then kicked himself for 10 more minutes in front of the media at Valley Ranch on Monday. "To do anything that would even jeopardize the well-being of the team is crazy and unexplainable … I feel terrible," Austin said. "I ran up right behind [Williams]. I was just excited for Roy. Romo checked out of the play. I saw the play he called. It worked. I was just excited. I can't do stuff like that. It was stupid."

That’s an impulse. It happens in the heat of the moment. Do you think a player could forget about fines and chest bump another player in a moment of excitement? I do. Do I think that’s likely to happen if they players are in the habit of simply handing the ball to the official? I think it’s less likely. One of the best ways to eliminate an unwanted behavior is to teach a behavior you want.

 

Now I'm not excusing Wade Phillips. I think the blame lies entirely with Wade. Via Bob Sturm.

Since the start of 2007, only Oakland has committed more penalties than the Cowboys.

Since the start of 2008, nobody has committed more penalties than the Cowboys.

Since the start of 2009, only Oakland has committed more penalties than the Cowboys.

 

And for those of you who think "this team had a penalty problem under Bill Parcells, too", let me put that to rest. From 2003-2006, "the Bill Parcells era", the Cowboys ranked 25th in the NFL in penalties during his tenure. 25th! Only 7 teams in the NFL committed FEWER penalties than the Cowboys did under Parcells.

However, I'm not convinced that Wade's problem is that he isn't metting out enough punishment. Punishment is easy. It might be that he's not structuring practices in the right way with the right rewards. Structuring and adhering to a training program is harder than punishment. Just consider that punishing players may not be the most effective approach.

 

p.s. most successful NBA coach … Phil Jackson.

 

Game Note

 

Last week I wrote that I hoped Garrett would run (specifically draws) more on third down.

 

Brian Burke also found the benefit of running increases as you approach the end zone.

Runs have a higher payoff only inside final 10 yards before the end zone, so it appears teams are passing too often there.

So my absolute favorite call of the day? The QB draw at the Minnesota 7 on 3&2. Perfect. Goes for a 1st down and sets up RW’s second TD.

 

Screen Plays

Just a quick note on screen plays. I couldn't re-watch the whole game so I don't know how the overall performance of screen plays was. However, two plays stand out and both suggest the OL struggles making blocks on screen plays. One play was a screen left to Felix where Kossier whiffed on the block in space. The other was a screen to Barber where Gurode and Davis where blocking. There was one defender who went through the midddle of the two without being touched. Barber's indecision behind his blockers didn't help. While he was deciding to go right or left the defender came up and made the tackle.  

Another user-created commentary provided by a BTB reader.

X
Log In Sign Up

forgot?
Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

Join Blogging The Boys

You must be a member of Blogging The Boys to participate.

We have our own Community Guidelines at Blogging The Boys. You should read them.

Join Blogging The Boys

You must be a member of Blogging The Boys to participate.

We have our own Community Guidelines at Blogging The Boys. You should read them.

Spinner

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9341_tracker