Response to Brian Burke's Washington Post article

By Mike Sommers

NB: I have also posted this on a site that I own and occasionally like to post to here.

I recently read Brian Burke’s article describing how coaches should avoid the mentality of trying to set up 3rd and short or 2nd and short and instead try to convert the first down.

It is not that I don’t agree with the article, but I wanted to consider several exceptions to that rule, and in the process came up with some interesting conclusions.

I can illustrate several instances where it’s best to NOT convert on first down. On 3rd however I can only think of 1, and it isn’t provable so much as it is intuitive, but I do have a statistical reference, unfortunately we just don’t have enough information to account for changes AS the play develops.

With 15 minutes remaining in the 1st quarter and a tie game, Brian’s EP model shows that 2nd and 3 on own 17 yard line is very close to the same as first and 10 on their own 20 yard line. 2nd and 2 on own 18 and 2nd and 1 on own 19 are BOTH vastly superior than 1st and 10 on 20. But most conversions will be a few yards past it.

If 2nd and 2 was average expectation of the following short yardage situation on first down, and average 1st down conversion went 4 yards past (1st and 10 on own 24) you would still be .01 WP better getting 2nd and short. The line in the sand is ROUGHLY “IF” you can get more than 4 yards past the first down marker, you should convert, if not, you should “settle” for 2nd and 2. If a runner is running up free and can easily get the first down and possesses the ability to go down one yard short of the first or less, the runner must be able to get over 6 yards past the first down marker for it to be the correct decision to run past the first down marker!

HOWEVER, if a team is making more optimal decisions on offense (and if they are aware of this information, they likely would be), I would argue that even 5 or 6 yards past the first down isn’t good enough to make up for the advantage of 2nd and short.

As I understand the EP model is based upon HISTORY and historically teams do not go for it on 4th down. If they did, I imagine the chances of converting and keeping the drive alive would be much higher, the EP is obviously higher in going for it on 4th and short, and that can only mean that if the team treats it as “4 down territory” that their EP would be even higher than advertised on 2nd and short. It also would be on 1st and 10 as well, but it is much more likely to be left with a 4th and short on 2nd and short than it is to be left with a 4th and short on same set of 4 downs on a 1st and 10. A gameplan to get a big play on 2nd and short, and then leave 3rd and 4th down to try to convert, or to try to convert 3 times in a row would likely boost the WP to the point that setting up 2nd and short is more favorable then perhaps converting AND getting an additional 5, 6, 7 or 8 yards past the first.

I don’t exactly know if there has been a study to determine the EP advantage the average team has if they made ALL of the correct decisions on 4th down, but I imagine over the course of the game the combination of intentionally setting up 2nd and short rather than converting at least often enough to keep defense off balance would significantly boost a team’s advantage, particularly if it was able to also go for it on 4th down more aggressively.

Plus, if one is close enough to get first, but just barely, they may resort to taking more risks trying to get the first, such as reaching the ball out, or cutting inward and staying in bounds where fumble is not only possible, but less likely to bounce out of bounds and remain the offense’s ball. That would require even more yards for a first down to be worth giving up the easy to convert 2nd and 3 or less.

Additionally, a team that is heavily positioned themselves to have an advantage over the average team in the pass game, but normally may have a slight disadvantage running would now have the opportunity to vary their play enough to prevent themselves from becoming predictable which can strengthen their passing attack and running attack over the course of a game.

The guideline is If you can’t get a big play, get 2nd and short, if you can’t get 2nd and short, get a first down, if you can’t get a first down, get what you can. If you are running towards the first down go out of bounds just short of the marker or go down in bounds unless you can get around 4-8 yards beyond the first down.

As for when 3rd down situation may dictate setting up 3rd and short, that will have to be published later as part two. There are a lot of new concepts to be introduced, and unfortunately there isn’t enough data tracking to really prove it in the game of football that I am aware of, but I am confident that there are hypothetical instances that exist, and they may exist throughout the entire game, which may drastically impact some decisions making a pass short of the 1st down on 2nd down very acceptable.

Following the EP model I believe still provides a very significant edge, however there are advancements that can be made, if only individual statistical categories are better defined as the play develops and as presnap reads are made, for now though, 2nd and short is superior to 1st and 10 occasionally, and 3rd and 1 is certainly not as good as 1st and 10, that much I am certain holds true in most situations.