Most NFL Coaches Have No Idea How to “Go For Two”

by Warren Sharp

Instead of the NFL owners debating a new 2 point conversion rule, they should have been railing on their head coaches for dumb decision making which might have made the perceived need to change the rule completely irrelevant.

Commissioner Goodell wanted to change “the most boring play” (the PAT) to something more exciting, unpredictable, and keep fans engaged after a TD before the next kickoff.  The proposals the league considered all revolved around the 2 point location, because all PATs (point after touchdown, worth 1 point) were getting moved from the 2 to the from the 15 yd line in the proposed modifications.  The other suggestions were either to keep the 2 point conversion at the 2 yard line, or move it in, to the 1 yard line.  That line of scrimmage for the 2 point attempt was the biggest point of contention, because owners viewed attempts from the 2 yard line as 50-50 propositions, which meant more coaches are simply going to kick the PAT, making the rule change inconsequential.  At the end, the owners voted 30-2 to keep the 2 point conversions at the 2 yard line, but move the PAT to the 15 yard line.

There are many issues with this entire debate, but at its core, it revolves around dumb coaching. Before we get there, let’s back up to look at the numbers:

I did a 10 year study on 2 point attempts, and found that despite many rules changes to help offenses in the last decade, teams are no better at converting these plays than they were years ago.

For instance, since the strongest rules changes came out to protect the passer and the receivers (in 2010), the last 4 years have not seen the level of success on 2 point attempts that any year from 2005-2007 saw.  Teams are going for it slightly more frequently, but they converted 48% or worse in 3 of the last 4 years.  From 2005-2007, the worst rate in any year was 54.2%, which was a better rate than in any year since 2010.

Obviously, current coaches are pretty bad when it comes to obtaining positive results from the plays they call when attempting a two point conversion.  But another key is the pass-to-run ratio of play calls and successful conversions.  The last 10 years, 59% of run plays converted into two points, but only 46% of pass plays produced the same result.  Below, you can see the number of 2 point attempts by year, and by play type over the last 10 seasons:

(click to enlarge)

The owners are correct that, for the large sample size of the last decade, the conversion average on these plays is 50%.  Meaning the expected points for an avg team “going for two” is 1 point.  The chart shows the total two point attempts (far right) are on the rise, but apparently not enough for the NFL’s liking.  But what clearly caught my eye was the significant success edge for running plays as opposed to passing plays.  Yet running plays were only attempted 27% of the time!

If you want to argue small sample size as the reason for the disparity, I’ll counter by sharing data on ANY play from the last few years (since that 2010 rule change which made passing more attractive) where 2 yards were needed for a first down:

The 61% rate on runs is only slightly higher than the 59% on 2 point conversion runs the last decade.  Meanwhile, the conversion rate for passes the last 4 years of 51% is slightly higher than the 46% for 2 point conversion passes.  You can see that the total plays called for both types is very similar, but running is far more effective.  The bottom line is that, while this is a passing league, short yardage is best battled on the ground and the numbers speak for themselves.

Yet coaches still desire to call 3 times more passes than runs when signaling in a two point conversion play.

The next piece to the puzzle is to examine WHEN coaches are going for two.  The fact is, over 80% of the time, its in the 4th quarter or OT.  The table below indicates the number of 2 point attempts in the last decade, and when they occurred:

Typically, those attempts in the 2nd half are only occurring in certain situations based on a cheat sheet the coaches pull out to look at (again, because most are dumb) to help them know when they should make this critical decision to go for two.  This next table takes the first table, which looked at two point plays by year, and breaks it down into two point plays by margin.  Negative means the team is trailing, and positive means they are winning.  The margin is taken AFTER the touchdown, but before the two point conversion.  (So for example, a margin of -1 indicates the team was losing by a TD, and just scored, putting them down by just 1 point (TD=6 points).  The PAT would tie the game.)

As is evident, there are a few key times when a team goes for two.  Most often, its in the 4th quarter, and in a game that is somewhat close, the most often times a two point conversion is attempted is when a team is down by 2, 5 or 10 (to cut the margin to 0, 3 or 8), or when they are up by 1, 5 or 12 (to grow the margin to 3, 7 or 14).  Sometimes teams go for it when up by 4, as leading by 5 is not helpful but leading by 6 is viewed as more helpful.

While much can be gleaned from this table, note the run vs pass rates for most of the game (except when a team is leading by 12+ points and may have subbed out personnel or is playing far less aggressively):

Two point runs convert 64% of the time (74 of 115 att) while two point passes convert just 47% (150 of 320 att) when not blowing out your opponent.

After looking at the numbers, this study clearly shows that the NFL is trying to see more two point conversions throughout the game, rather only at the very end, and only in situations when the cheat sheet tells them to “go for it”.  They want teams to “go for it” throughout the game because its in their best interest from an “expected points” proposition, not just out of necessity.

The problems, as the numbers clearly indicate, are that:

  1. teams should have been going for it more often all along, and
  2. teams should have been attempting more runs as opposed to pass plays.

If the average running two point attempt converts 64% (except in cases of a blowout) over the last decade, and the average run on “XX and 2″ converts a first down on 61% of attempts, than a called run on a two point play nets 1.22 expected points (I used the lower of the two – 61% – and multiplied by 2 points, the points rewarded on a successful conversion).  Even if teams made every single PAT they attempted (a few random tries are missed) running on 2 point conversions clearly gain more expected points:  1.22 > 1.

Now the talk shifts to moving the PAT to the 15 yard line, thus forcing a kick from the 32/33 yard line.  Over the last 10 years, 93% of 314 attempts were converted from that distance.  Over the last 10 years, 90.8% of FGs from 30-35 yds were converted.  So assuming even the upper end, the expected points from the new distance is 0.93, which is certainly lower than 1.0 (avg expected points from all 2 point conversions) and definitely lower than 1.22 (avg expected points from all 2 point conversion runs plays).

Keep in mind, all of these are averages.  Some teams are better converting on short yardage runs, and some defenses are worse at allowing them.  Last year, for example, the Seattle Seahawks were a top 5 short yardage run team, converting 73% first downs when needing 1-2 yards “to go”.  The second worst defense last year in this role was the Patriots, who allowed 78% of runs needing 1-2 yards to be converted into a first down.

Examine this scenario in the Super Bowl: Seattle scored 3 TDs but went for the PAT each time.  Thus, they trailed by 4 (28-24) as time ran down, and they had the ball at the NE 5 with 1 minute to play, needing a TD to win.  But, what if Seattle and their top 5 run offense tried to score on two point conversions after their TDs vs the 2nd worst short yardage run defense in the NFL?

To score 4 total points after their 3 TDs (instead of PATs each time), Seattle would have needed to convert 2 of their 3 attempts, or 67%, a number just north of the NFL average for runs on 2 point conversions in non-blowouts over the last 10 years (64%).  But we know Seattle’s short yardage run offense was one of the best (converting 73%) and the Patriots had one of the worst defenses (allowing 78%), so converting 67% seems very achievable.

Thus, after converting on 2 of the 3 attempts for two, they would have trailed NE by only 3 points rather than 4, and a lot more options would have been open for Seattle at the end of the game.  (Of course, this sidesteps completely the decision to pass on the 1 yard line rather than run, which ultimately cost Seattle the game.)

The bottom line with this analysis is the NFL owners should be look closer at the data during these meetings to see that their own coaches were costing them games the last few years, both by not going for two enough, as well as passing too often on these attempts.  Running would have led to more success, which would have led the coaches to go for it more often, which would have perhaps avoided this discussion at each of the owner’s meetings the last few years.

[Naturally, the more teams run with success on these plays, the defense eventually will adjust, which will make the run rate dip but the pass rate should then improve.  Also, of course variance creeps into things, as does in-game strategy for individual games.]

Regardless of the rule change, coaches should “go for two” far more often than they do, particularly against teams with poor goal line/short yardage run defenses.  Success rates obviously vary, based on particular offense vs defense matchups.  But historically, we’ve seen teams waiting for the 4th quarter to whip out a cheat sheet to tell them what to do, based on time and margin (and then calling a pass play 73% of the time), which absolutely speaks volumes about how dumb most coaches are in the NFL.  Of course, that’s nothing new.  As the saying goes, “you can lead a horse to water…” and perhaps we will see more teams attempt two point conversions earlier in the game in 2015, but it remains to be seen if 70%+ attempts are still of the low-rate-passing variety.

My initial guess is, though we see a few more attempts (the trend has been pointing up in recent years even before the change) it won’t amount to nearly the waste of time spent discussing it.  Most coaches, elementally, are scared of making poorly perceived decisions, and would rather see criticism pile on a FG kicker who misses a 32 yard “gimmie” PAT as opposed to themselves after making the call to “go for 2″ at any time other than late in the game and in a situation that’s on “on the chart”, stats be damned.