THE RAIDERS WON A GAME!!!!!
a win for my boys in a terrible game but at a huge cost. dez bryant injured was bad enough but romo out for 8 weeks is a real season changer
got to try to be in with a shot for at least a wild card when the pair of them come back.
redskins for the division anyone?
With the Eagles starting so badly, and Romo going down, wouldn't be at all surprised to see 8-8 win the division. Eagles O-line looks horrific, and obviously losing Romo is huge. Not seen a single snap of the Redskins this season, so can't really comment on them, but before the preseason I would've expected all 3 QBs on their roster to start a game this year. I don't know if Cousins has done anything that might make me change that opinion. Could be last year's NFC South all over again, with a team with a losing record going to the playoffs. Especially if Philly's O-line doesn't improve very soon, and/or Romo is out for longer than expected. Didn't Rodgers have a broken collarbone a couple of seasons back, then every week they said he might be back, only for him to not come back for ages and Green Bay limped into the playoffs with Flynn, Tolzien, and someone else getting starts at QB? Rodgers came back for Week 17 then had a pass to Randall Cobb to win vs the Bears, if I remember correctly.
How on earth is a game theory coach on the sidelines presumably not a thing?
So Pittsburgh going for 2 after their first 2 touchdowns tonight made me think about this some more.
I assume a lot of people in here are familiar with being a certain number of points down determining whether you go for 1 or 2. Down by 1, score a TD, you now lead by 5, so obviously go for 2 late in the game to make it a 7 point lead, and so on. Down by 11, score TD, go for 2 so you can tie it with a field goal. Plenty of these are pretty slam dunk decisions.
Down by 14, every NFL coach in history takes the extra point to close the gap to 7 points, right? Because I think they've been doing it wrong for the last god knows how long, and if not, then the change to the PAT means they're DEFINITELY doing it wrong now, IMO.
Heard the commentators during the Seahawks v Packers game say that the
league average for 2pt conversions was 48%. The old extra points were successful 99% of the time or so, the
new 33 yard extra point I believe is expected to be somewhere around 95%.
Down 14, let's say there's 30 seconds for an onside kick and to throw the ball into the end zone 4 times - Go for 2 after TD, now down by 8 52% of the time when we fail and and down by 6 48%. Down by 6, after another touchdown, obviously we kick the extra point. Down by 8, we obviously try to tie it up, both absolute no brainer decisions. The key here though is that we can almost guarantee a win with a successful 2 point conversion, and still get to OT half the time when we fail.
---- SKIP THIS BIT IF YOU DON'T LIKE MATHS ----
For ease of calculation, I'll disregard unusual stuff like defensive scores on PATs, kick blocks recovered by offence and botched snaps that end up resulting in a successful 2 point conversion, and 1 point safeties, as the difference they make would be negligible anyway.
Down 8 (52%) - Go for 2, down 2 52%, tie 48%
Down 6 (48%) - Kick extra point, win by 1 95%, tie 5%.
Therefore, by going for 2 when down by 14, the times where we score 2 touchdowns without the opposition scoring results in the following:
Win the game 0.48 x 0.95 = 0.456 = 45.6%
Overtime (0.48 x 0.05) + (0.52 x 0.48) = 0.024 + 0.2496% = 0.2736 = 27.36%
Lose 0.52 x 0.52 = 0.2704 = 27.04%
Whereas going for 1 twice gives the following:
Win 0%
OT 0.95 x 0.95 = 90.25% we make both extra points, plus 0.05 x 0.48 = 2.4% that we miss the first extra point, then are successful on the 2nd PAT when going for 2 = 0.9265, or 92.65%
Lose 0.95 x 0.05 where we make the first PAT and miss the 2nd PAT, plus 0.05 x 0.52 where we miss the first PAT then fail the 2pt conversion = 0.0735, or 7.35%
(FWIW, Going for 1, then going for 2 after 2nd TD makes absolutely no sense, you'll only win 48% of the time and never go to OT that way. You'd need to expect to LOSE in OT something ridiculous like 92% for it to be the correct option. Never happening.)
For going for 1 to be better than going for 2 when down by 14, we need to have a pretty high chance to win in OT when we get to OT, as going for 1 will result in OT far more often.
Here, "n" represents the probability that we win in OT
P(Win game after going for 2 after first touchdown when down by 14) = 0.456 + 0.2736n
P(Win game after going for 1 after both touchdowns when down by 14) = 0 + 0.9265n
For going for 1 to be better, 0.456 + 0.2736n < 0.9265n
0.456 + 0.2736n = 0.9265n
0.456 = 0.6529n (subtract 0.2736n from both sides of the equation)
0.6984 = n (Divide both sides of the equation by 0.6529)
0.6984 < n for 1 point to be better
---- END OF MATHS BIT ----Therefore, we need to win in OT more than 69.84% of the time for going for 1 to be the correct move. How many teams in the NFL will go to OT and realistically expect to win just shy of 70% of the time? And in reality, the number will probably marginally higher than that, because a successful 2 point conversion down by 14 opens up the possibility to kick 2 x unanswered field goals in some situations where there is more time remaining on the clock (Touchdown with 10 mins left, get the ball back, 4th and long in FG range with 4 mins left and all 3 time outs to get the ball back and get in FG range a 2nd time for instance), which I haven't accounted for here.
Not going to show the Maths again, same as above with different numbers in there, but with the old PAT where we succeed 99% of the time, we "only" needed to win in OT 66.97% of the time to make going for 1 when down by 14 correct, which is still very optimistic, IMO.
Basically, NFL coaches have been getting this one wrong since forever, probably because it's one of those situations where no coach will do it because results orientated people will bitch about it and get the coach fired when they could've taken it to OT and it costs them the game. However, scoring a TD when down by 14, is an absolutely obvious spot to go for 2 IMO.
Author