2-point Conversion Chances - Strategy

[Thurman#1]

15 hours ago, mike22nc said:

It’s not about chasing points, it’s about maximizing expected points. The more times you score, the more we would expect this to positively effect the outcome as long as we are more likely to be successful than not.

 

 

The problem is that assuming that you manage over 50% of them, you've maximized expected points over the season.

 

But NOT worked towards maximizing wins within the same period.

 

The problem with this is that by going for two consistently and racking up three or four, hell, say eight extra points per season you may (or may not) lose a game or two as a result. In exchange for the overall extra points, you give up on the extremely predictable nature of those points.

 

Say we played last season over again with the exact same results, except that we went for two and got five extra points, how many extra wins would that have brought us?

 

The correct answer is that there's no way of knowing. Say we got one extra point in eight games, none of which were anywhere near close. We got eight extra points and zero extra wins.

 

Then we get two fewer points in one game, scoring four TDs, going one out of four in two point conversions and losing the game 28 - 27?

 

We got six extra points on the season and lost an extra game as a result.

 

Extra points over the course of a season don't mean much. What matters is extra or fewer points in close games. You don't know in which games those points will be absolutely crucial.

 

 

 

 

[Thurman#1]

11 hours ago, Nextmanup said:

The OP's model (which admittedly is not based in reality but it just a guess) says that we would score 2 points 60 % of the time.

 

If that were the case, long term, we would get MORE points going for 2 than 1.  

 

Also, there is no "risk" in that situation and there most certainly IS a reward--getting more points (with a 60% success rate).  

 

To the OP:  I would LOVE for us to aggressively go for 2 points, every time we score.

 

The question is what would our conversion rate really be in reality.

 

Below are REAL numbers taken across all teams in the NFL, for nearly 1,000 plays.

 

The thing is, the success rate of the BILLS might very well be ahead of the league average.

 

I would sweep josh with the ball  and give him a few receiver options in the end zone.

 

If the pass is there, throw. If he can just run it in, run it in.

 

Josh is nearly unstoppable on that play.

 

Asking McDermott to  do this is like asking your dog to sing.  Probably won't happen. 

 

Play/ Success Rate / Attempts

Pass/ 43.4%    /           739

Run/    61.7%   /           258

Overall/  48.2%  /         997

 

 

 

Less like asking your dog to sing. More like putting your dog onto a giant checkerboard where 55% of the squares say YES and 45% say NO and waiting for him to poo about 70 times and asking the owner whether he'd like to bet his kid's college scholarship money on only two or three of those events and you can't predict in advance which one of those events he'll be betting on, the first, the twenty-third or the 67th.

 

It's too opaque.

 

Sounds very good at first, but unless you know which ones of those events will be important and which won't, you'd like to make sure that you don't take the riskier choice.

 

A wildly obvious example: your team scores on the last play of the game to tie. Are you going to go for two, knowing that taking that choice every time over the course of the season will be extremely likely to raise the number of total points you'll score over the course of the season? Or would you go for two with no time left and down by one? Again, doing this every time over the course of the season means you're virtually guaranteed to raise the number of overall points you'll score.

 

It depends on the circumstances each individual time, and at the beginning of games you haven't a clue which points you might miss might be wildly unimportant or completely irrelevant to your win percentage.