Redskins......by the Numbers

One of many experimental iterations ...

DixieFlatline

The Starter
Joe Gibbs Club Member
Joined
Sep 7, 2014
Messages
1,283
Reaction score
7
Points
38
Location
Chapel Hill, NC

William and Mary


Posted this originally in the Random Thoughts thread but moving it to it's own thread by the suggestion of Boone.

After yesterday's game, I was questioning the lack of a commitment to the running game especially after we had some decent runs. This is nothing new, but being a numbers guy I decided if I could find a way to get more insight into what we've actually been doing from a play call perspective.

I found this site where someone has gotten all the play by play data from publicly available sources and created files since 2013. I've got access to some cool tools at work, so I took those files, merged them, and loaded them up to run some reports to see what I see. What I have below is a first pass at breaking down our play calls from the 2014-2016 seasons. There is more I want to do with this, but thought I'd throw what I have now to see if it sparks any ideas on how to break it down further. For example, I'd love to throw in some variables of how the score may dictate the calls. Unfortunately, the files don't have that information. I did however try to make an attempt at seeing if the calls changed based on end of half/end of game.

Like I said, I just started and plan to do more. If you have any thoughts on ways to slice the data let me know and I'll see what I can do.
This is a breakdown of all our plays on 1st, 2nd, and 3rd down. The left shows pure numbers, the left normalizes the data so you can visualize percentages and then further breaks it down by Distance. When I first looked at overall play calls on first down, I was surprised to see it was almost 50/50. That is one reason why I added the graph with the distance.

I tried to get these side by side, but when doing so it seemed kind of small. If anyone has any ideas on how to play and resize images with a post, let me know.

BasebyDown.jpg BasebyDownandDistance.jpg

The next set filters out 'CrunchTime'. I defined this to be in the last 2 minutes of each half.

NotTimeCrunch_ByDown.jpg NotTimeCrunch_ByDownandDistance.jpg


This is 'CrunchTime' only.

TimeCrunch_ByDown.jpg TimeCrunch_ByDownandDistance.jpg
 

DixieFlatline

The Starter
Joe Gibbs Club Member
Joined
Sep 7, 2014
Messages
1,283
Reaction score
7
Points
38
Location
Chapel Hill, NC

William and Mary

So, I played around a little more with this NFL Play dataset last night just trying to learn what I have here and I thought I'd look at Defense. I may not have said in my last post, but since I'm still learning how to analyze this data, it may not be 100% accurate, but I expect it's pretty close. For this set, I looked at the last three seasons and wanted to look at how often we let a first down be converted at different distances. And then I compared it to the NFL numbers over the same time period. It should be noted that this Conversion percentage is for the very next play (e.g. on 2nd and 10, did they get at least 10 yards on that play? It does not look at if they eventually got 10 yards in that set of downs).

First chart is the NFL averages over the last three seasons.
NFL Defense conversion Next Play.jpg

Second chart is the Washington defense. The differences in how we played on 3rd with 11-15 yards and 6-10 yards stood out to me.
WAS Defense conversion Next Play.jpg
 

Boone

GM
Staff member
Joined
Apr 11, 2009
Messages
39,891
Reaction score
994
Points
1,944
Location
Greensboro, NC

Marine Corps Virginia

So awesome! Now we need some statistical geniuses to figure out what this data means :)
 

Neophyte

GM
Staff member
Joined
Jun 30, 2009
Messages
8,484
Reaction score
115
Points
218
Location
Dallas


The first thing that comes to mind about things I'd like to see is a breakdown of play calling against the scoreline. In broad terms, what is it when we have a lead vs when we are trailing. If the data is there to get more granular, for example what is it when we are leading by 3, by 7, by 14 and losing by those same margins, then that would be better but not knowing what is included in your source material then even the broad idea might not work.

This is some awesome stuff though. I'm going to have to look at it for a while to make anything out of it and even then I'll probably miss most of it.

Thanks for this! Great work!
 

DixieFlatline

The Starter
Joe Gibbs Club Member
Joined
Sep 7, 2014
Messages
1,283
Reaction score
7
Points
38
Location
Chapel Hill, NC

William and Mary

The first thing that comes to mind about things I'd like to see is a breakdown of play calling against the scoreline. In broad terms, what is it when we have a lead vs when we are trailing. If the data is there to get more granular, for example what is it when we are leading by 3, by 7, by 14 and losing by those same margins, then that would be better but not knowing what is included in your source material then even the broad idea might not work.

This is some awesome stuff though. I'm going to have to look at it for a while to make anything out of it and even then I'll probably miss most of it.

Thanks for this! Great work!
Calls based on score are something I'm keenly interested in figuring out. The dataset that was available didn't have anything about current score. I've emailed the guy that put it together to see if he has it or if he may know where I might be able to merge it in.

I'm also interested in figuring out patterns of play calling. For example, can we see how often we might rush for 5 yards on 1st down, but then don't run again? I think I can do that with this data, but I might need to play with it a bit.
 

JohnssBass

The Team Captain
Joined
Apr 13, 2014
Messages
2,426
Reaction score
0
Points
0

Army

Here's a link that may help you with some additional data. You'll probably have to crunch this info down to get all the details your looking for but much of it is in there.

link: https://www.pro-football-reference.com/boxscores/201709100was.htm

on 1st downs - they almost exclusively alternated by passing on the first one then running on the next 1st down. 3 straight 1st down runs (net 9 yds) at the end of the 1st qtr/beginning of the 2nd. 2 straight passes on 1st downs and that wasn't until they were down by 13 with 1:30 to go in the game. Everything else alternated. The first completion on a 1st down pass was the 5th toss.

1st downs prior to the 2:00 warning - RBs had 10 runs/32 yds compared to 4 of 12 passes/14 yds with 2 sacks.

2nd downs were somewhat problematic because they were typically in 2nd and long. But in the 2nd half they ran 12 passing plays & 0 running plays on 2nd down even on 2nd with 3 to 6 yds to go. 3 runs on 2nd down in the 1st half. 3rd down rushes - 0. For the whole game. Understandable as their best distance on 3rd was with 6 yds to go.

I'm not sure how they define a 'short' pass versus a 'deep' one but the Skins weren't awful in the short category. 22 of 33/224 yds/1 TD/1 Int. The deep was 1 of 7/16 yds which is awful. 4 of those incompletions were thrown on 1st down. 3 of those were the first 3 first down passing attempts of the game. Don't really a home run out of the gate or on 1st down - maybe build up to it with some short passes and runs. Helps avoid those 2nd & long situations that turn into 3rd & long.
 

Bulldog

The All-Time Great
Joined
Jul 19, 2009
Messages
15,227
Reaction score
191
Points
343
Location
Bethesda Md


I may be making a leap here but these numbers back up what I had suspected, namely that Jay Gruden is a HC/OC who can get an offense to put up physical numbers in terms of yardage, etc. to rank rather highly in the NFL stats - but ultimately he is NOT a good playcaller when it comes to DISTRIBUTING the calls on the play chart, and choosing WHEN to call certain plays and in WHICH GAMES a run-based attack and short passing game may be the 'call' while in others a more open downfield gameplan is more appropriate.

That is what made Gibbs, Walsh, and now Bellichick so good as coaches.

They can alter a game plan to go run-based, pass-based, or mix-match based upon the opponent they are facing.

The Redskins under Gruden don't discriminate.

We seem to come out with the same offensive philosophy irrespective of the opponent and their strengths on defense.

If we are facing the #28 rush defense to me it only makes sense to run the ball more. But that's not what we do necessarily.

In games like that we often tend to come out and throw 45 passes anyway even though that particular club may have a decent to solid secondary.

We are not taking the opponents clear WEAK SPOT and attacking that over and over until they show they can defend what we are doing.
 

DixieFlatline

The Starter
Joe Gibbs Club Member
Joined
Sep 7, 2014
Messages
1,283
Reaction score
7
Points
38
Location
Chapel Hill, NC

William and Mary

Here's a link that may help you with some additional data. You'll probably have to crunch this info down to get all the details your looking for but much of it is in there.

link: https://www.pro-football-reference.com/boxscores/201709100was.htm

on 1st downs - they almost exclusively alternated by passing on the first one then running on the next 1st down. 3 straight 1st down runs (net 9 yds) at the end of the 1st qtr/beginning of the 2nd. 2 straight passes on 1st downs and that wasn't until they were down by 13 with 1:30 to go in the game. Everything else alternated. The first completion on a 1st down pass was the 5th toss.

1st downs prior to the 2:00 warning - RBs had 10 runs/32 yds compared to 4 of 12 passes/14 yds with 2 sacks.

2nd downs were somewhat problematic because they were typically in 2nd and long. But in the 2nd half they ran 12 passing plays & 0 running plays on 2nd down even on 2nd with 3 to 6 yds to go. 3 runs on 2nd down in the 1st half. 3rd down rushes - 0. For the whole game. Understandable as their best distance on 3rd was with 6 yds to go.

I'm not sure how they define a 'short' pass versus a 'deep' one but the Skins weren't awful in the short category. 22 of 33/224 yds/1 TD/1 Int. The deep was 1 of 7/16 yds which is awful. 4 of those incompletions were thrown on 1st down. 3 of those were the first 3 first down passing attempts of the game. Don't really a home run out of the gate or on 1st down - maybe build up to it with some short passes and runs. Helps avoid those 2nd & long situations that turn into 3rd & long.
Thank you sir....I'll definitely look at that closer.
 

redskins26

The Team Captain
Joined
May 17, 2010
Messages
2,589
Reaction score
1
Points
66

Navy

Running doesn't necessarily correlate with winning but winning correlates with running. Last year I looked at pass happy teams like the patriots Atlanta greenbay yes the final tally might be 35 passes 28 rushes but in the game as they build up their leads you see wide discrepancies of 28 passes to 10 rushes till they start to run the clock out.

On a different note keep an eye out for the NE saints game this week. NE defense will probably let the saints running game have some success early on
There is a reason for that
 

Users Who Are Viewing This Thread (Total: 1, Members: 0, Guests: 1)

Private conversations
Help Users
    You haven't joined any rooms.
    Chat 0
    Top