[Ganchrow]

In the NFL it is my understanding that since the advent of the two-point conversion in 1994 the league-wide probability of making a two-point conversion has been roughly 50% and the probability of making the extra point has been close to 100%. For the below analysis I use the numbers 50% and 100% respectively. Please let me know (citing a source) if the former figure is significantly off.

The NFL game history seems to be packed to the brim with situations where coaches foolishly forgo the two-point conversion. Here’s an interesting scenario I just thought of while anguishing in the dentist chair:

4th quarter, very little time on the clock (let’s say right around 2 ½ minutes). Team A has just scored a touchdown to bring it within 8 points of Team B. Ceteris paribus, should Team A take the extra point or should it try for the two-point conversion? Well clearly, based upon the content of the preceding paragraph, I’m sure you'll realize I’m going to argue for the latter. Here’s why.

With the small amount of time remaining on the clock, it is reasonable to assume that Team A will need to do two things to win the game: the first is to prevent Team B from scoring again; and the second is to score another touchdown. (Furthermore, it is unreasonable to expect Team A (except in the most pathological of cases) to score more than one further time in the quarter.) Now if Team A is fails to accomplish these two tasks, it will lose the game. This is apparent whether it converts 0, 1, or 2 points after this last touchdown. However, if Team A IS able to accomplish these two tasks, the situation becomes more interesting.

Let’s say that Team A goes for the extra point which, as per assumption, it makes with certainty. After it scores its next TD, it is then down by 1 point. If it chooses to kick the extra point there, the game goes to OT which, given our assumption of all else being equal, it wins with probability 50% (neglecting ties). If instead of kicking the extra point after the second TD it goes for two, Team A will still win the game with 50% probability (i.e., if it converts (probability 50%) it wins, if it fails to convert (probability 50%) it loses). Hence, if Team A takes the extra point after its first TD and manages to accomplish the two tasks outlined above (preventing Team B from scoring again in the quarter AND scoring another TD), it will win the game with 50% probability.

OK, now let’s say that Team A goes for the two-point conversion which, as per assumption, it makes with 50% probability. If it converts (50% probability, all conditioned on its accomplishing the “two tasks”) it wins after taking the extra point following TD# 2, if it fails to convert (50% probability) then, down by 8, it tries for the two-point conversion after TD# 2. If it makes that (50% probability) the game goes to OT. In OT it wins with 50% probability (once again neglecting ties). If it fails to make the two-point conversion after TD# 2 (50% probability) the game is lost.

So in the case of Team A initially going for the two-point conversion, conditioned on its accomplishing the “two tasks”. The probability of a victory = prob of making first TPC + prob of not making first TPC * prob of making second TPC * prob of winning in OT = 50% + 50%*50%*50% = 62.5%.

So to sum up: down by 8 after having just scored a TD near the end of the fourth quarter, the DEFAULT action for coaches should be to go for the 2 points. A coach should only go for the extra point if team specific factors would so dictate (say perhaps if Team A is sufficiently stronger than Team B such that it’s probability of winning in OT is greater that 62.5%, or if the Team A short-yardage game and Team B’s red-zone defense are such that Team A’s chance of converting a TPC would be less than 41.42%).


QED

[rm18]

I'm not sure what the number is but it is in the id to low 40's I think.

But I think you are still right although it would be a lower percentage than 62.5

Another intersting statistical thing is it is actually advantageous to go for it on 4th down inside your own 20 yard line if you have less than 5 yards to convert, because it is easier to stop teams in the NFL on defense on a short field when they cannot stretch the field, and then factoring in that you pick up the 4ht down conversion over half the time and do not have to punt.

[Ganchrow]

I'm not sure what the number is but it is in the id to low 40's I think.

You may well be right. I think my only source for the 50% figure is something I may have heard Madden mention 3 or 4 years ag. I'd love to see some data on this any which way it points.

But I think you are still right although it would be a lower percentage than 62.5

I assume you meant to say, "a higher percentage than 50." To wit, a TPC % of 41.42% is the cut-off assuming a EP conversion % of 100%. A lower EP conversion % would obviously lower the 41.42% figure as well.

[Mudcat]

I am sold. I like that kind of outside-the-box thinking.

[rm18]

Right higher than 50%, and really the chances of two straight extra points are only about 98%

I think since the exception it is close to 50%, but has been getting lower over the years, basically every year in the history of football defenses are getting stouter on the goalline and short yardage.

[Ganchrow]

Another intersting statistical thing is it is actually advantageous to go for it on 4th down inside your own 20 yard line if you have less than 5 yards to convert, because it is easier to stop teams in the NFL on defense on a short field when they cannot stretch the field, and then factoring in that you pick up the 4ht down conversion over half the time and do not have to punt.

There's actually some great econometric literature out there suggesting that teams go for it WAY too infrequently on fourth down across a wide range of situations. I'll see if I can't find a link.

[Ganchrow]

I am sold. I like that kind of outside-the-box thinking.

Then please hire me to coach your NFL team. That's probably one of the only jobs out there for which I'd trade my own.

[Sean]

In the NFL, the national average for success on a two-point conversion is 43 percent versus 94 percent when kicking the extra point. In the NCAA, the two-point conversion is successful 43.5 percent of the time versus 93.8 percent when kicking the extra point.

source: HumanKinetics.com as excerpted from the "Complete Guide to Special Teams"

[EBone]

The problem with going for two in these coaches minds is simply: job security. As popular as NFL football is in the US, these coaches are under a mini-microscope for every move they make. If you don't meet or exceed those percentages as described by ganchrow, the mob(public) will maul you for your "stupid" decision. Really, you would have to exceed those percentages by a great bit in order to justify your actions. The mob mentality with NFL fans, sports talk radio and other venues will just crucify the coach for perceived mishandlings in going for two.

When it comes to a boiling point, the owners almost always bow to public scrutiny and someone ends up being fired (more than likely the coach). Coaches, no matter what they say, have to keep an eye on public sentiment whether they like it or not if they want to keep their jobs. I'd like to see a renegade coach try this and see if he could sway public perception. I highly doubt that would be possible.

E

[Ganchrow]

In the NFL, the national average for success on a two-point conversion is 43 percent versus 94 percent when kicking the extra point. In the NCAA, the two-point conversion is successful 43.5 percent of the time versus 93.8 percent when kicking the extra point.

source: HumanKinetics.com as excerpted from the "Complete Guide to Special Teams"

Yeah, I see that on this Human Kinetics page. I'd still like to see a primary source, however, or at least find somewhere from which I could scrape the yearly data.

[jjgold]

I wonder how that data translated to points

Say A team scores 50 TD's

Kicks extra point every time

50x7=350

or

Goes for two on every one

50x????=??

[rm18]

It is not 94% in the NFL

[Sean]

I wonder how that data translated to points

Say A team scores 50 TD's

Kicks extra point every time

50x7=350

or

Goes for two on every one

50x????=??

347.0 vs. 321.5

[Ganchrow]

The problem with going for two in these coaches minds is simply: job security. ... The mob mentality with NFL fans, sports talk radio and other venues will just crucify the coach for perceived mishandlings in going for two.
... I'd like to see a renegade coach try this and see if he could sway public perception. I highly doubt that would be possible.

The NFL coaching world is a competitive one and in order to survive in it one must constantly endeavor to better his opponents. You bring a sufficiently talented and talented coach into the NFL with enough new ideas and with enough courage to execute them, and even if the coach is wrong here or there the times will catch up with reality.

Look at Vince Lombardi. He bucked the trends. Now the guy's got a trophy named after him.

[Ganchrow]

It is not 94% in the NFL

From Yahoo Sports we see that in 2004 1,174 extra points were made out of 1,184 attempted for a conversion percentage of 99.16%. I think it's a safe assumption that this statistic has remained realatively stable over the years.

To me this would tend to call into question ALL of the Human Kinetics data quoted above.

For the entire period from 2001 - 2005-to-date the EP percentage is 98.53%