Thread: Extra Point/2-Point Conversion Tracking

Originally Posted by NightTrainLayne

If you have a positive EV, you go for it every time. Kenny Rogers knows this very well.

From a game theory standpoint, this isn't always the case. The best example I can think of is tournament poker vs. cash game poker. If you have an edge in cash, even if it's thin, you're always going to press it. But in tournament poker, the value of preserving your tournament life sometimes outweighs the potential gain of pressing your thin edge.

If I'm trying to apply that here - perhaps some of the other variables outweighed - like the desire to not expose their 2 point plays, a potential injury to a key player, etc.

I would be interested to hear Tomlin's logic.

Originally Posted by Buff

From a game theory standpoint, this isn't always the case. The best example I can think of is tournament poker vs. cash game poker. If you have an edge in cash, even if it's thin, you're always going to press it. But in tournament poker, the value of preserving your tournament life sometimes outweighs the potential gain of pressing your thin edge.

If I'm trying to apply that here - perhaps some of the other variables outweighed - like the desire to not expose their 2 point plays, a potential injury to a key player, etc.

I would be interested to hear Tomlin's logic.

Up by 19, then up by 26? I'm not trying to be facetious here. I think if you're strategy is to go for 2 early and increase your lead, once your up by 19 and then 26, why have your starters out there for an extra play?

I guess that would fall under your tournament theory.

Oh game theory! You dirty little whore!

Originally Posted by chazoe60

Didn't watch much of that game but I believe it was a bit of a blowout, could that be why they quit going for two?

It could be. .. but that would probably be the worst reason. It was actually still in the first half, and the Steelers had scored 3 TDs.

To me, if you're going to do it, you need to do it virtually every time, whether it's a blow-out or not. Exceptions for Game Theory situations where the game is close, near the end of the game etc.

Originally Posted by Buff

From a game theory standpoint, this isn't always the case. The best example I can think of is tournament poker vs. cash game poker. If you have an edge in cash, even if it's thin, you're always going to press it. But in tournament poker, the value of preserving your tournament life sometimes outweighs the potential gain of pressing your thin edge.

If I'm trying to apply that here - perhaps some of the other variables outweighed - like the desire to not expose their 2 point plays, a potential injury to a key player, etc.

I would be interested to hear Tomlin's logic.

I would like to hear Tomlin's logic as well.

I'm not sure the tourney/vs./cash poker game exactly applies, but I understand your point.

I wondered if it was due to wind. 1st Quarter they had wind they didn't like so went for 2. 2nd Quarter, at the other end of the stadium they liked their chances better with 1. . ..but then missed it!

If you go for it every time you're going to be labeled a Belichickian.

http://www.businessinsider.com/nfl-e...nt-rule-2015-9

Originally Posted by NightTrainLayne

I would like to hear Tomlin's logic as well.

I'm not sure the tourney/vs./cash poker game exactly applies, but I understand your point.

I wondered if it was due to wind. 1st Quarter they had wind they didn't like so went for 2. 2nd Quarter, at the other end of the stadium they liked their chances better with 1. . ..but then missed it!

Yeah, it was a stretch trying to apply the poker analogy to this situation - I was just thinking of an example where you wouldn't pursue a situation that is +EV. But your original point is a good one that you would expect them to go for it every time if they believe it will net them more success in the long run. Which makes me want to understand their logic... Too bad Tomlin doesn't post on some coaching theory message board.

Roethlisberger: Steelers will keep going for two.

http://espn.go.com/nfl/story/_/id/13...?ex_cid=ESPNFB

Some interesting stuff here. Doesn't go into why they started kicking extra points last week, but I guess you can't blame them for not divulging their strategy.

Thought this was a good analysis of what we were talking about:

After the first two scores, Pittsburgh backed off a little and kicked an extra point on the third. It missed. But even that isn’t necessarily a capitulation to lower risk plays. For one, they were stomping the 49ers so badly that it could be as simple as not wanting their best two-point plays on film this early in the season. For another, there’s a time in the game when you’re best off not maximizing your Expected Points Added (that is, going for two): once the endgame scenarios present themselves in the fourth quarter—you’d much rather kick a field goal to go up 12 with three minutes left than max out your EPA on fourth-and-1 from the 10, for instance. When you’re up 29-3 in the first half, it might be a little premature to jump to your endgame protocols, but you’re not exactly wrong to begin minimizing risk in a few small ways. Plus, Josh Scobee sucks, and could use the practice.

http://deadspin.com/the-steelers-wen...s-g-1731946052

Just realized I never gave the final update form week 2. Here are the totals as they stand now:

2-Point Conversions: 8/15, 53.3%, 1.06 pts per attempt
Extra Points: 146/155, 94.1%, .94 points per attempt

So through 2 complete weeks, the two pointer is a better decision by .12 points per attempt.

Well this week is off to a nice start so far. 1/1 on 2-point conversions.

Originally Posted by NightTrainLayne

Ideally, the NFL should set this up so that the EV for each decision is fairly equal.

It did, almost by accident (because the NFL's used a different PAT spot than college far longer than it's used 2 PATs.) It decided that didn't create enough big 'splosions for the ratings it wants, so shifted the even EV in favor of 2 PATs. Which is asinine for precisely the reason you say: It should be a wash, so the risk/reward ratio's exactly the same for both, favoring neither. The only good news is that if NFL coaches got their jobs on actuarial skills they'd never kick on 4th and <6 ANYWHERE and never kick a FG inside the 10.

Originally Posted by NightTrainLayne

As it sits right now, this issue really highlights how irrationally risk averse NFL coaches are.

Ayup.

For what it's worth, I "expect" the 2 PATs EV to end up right where it's been since the NFL legalized: ~1 pt. That "~" is theoretically significant (but probably not actually IMPORTANT) because moving the PAT back couldn't help reducing its EV from ~0.995, but still didn't reduce it MUCH, as the stats so far have reflected. So the precise amount each falls shart of EXACTLY 1 determines, in principle, which try's more desirable. But few coaches are engineers, and even those who are remain human, so they'll NEVER stop seeing the kick as practically a sure thing and the scrimmage play as a job-threateningly huge gamble.

I wish the NFL had just said, "OK, from now on TDs are automatically 7 pts UNLESS you go double or nothing with a scrimmage play for 8." The problem wasn't that PATs were nearly automatic, but that the difference between "nearly" and "literally" turned the so-called "extra" (i.e. bonus) point for doing something really good into a PENALTY for doing something really good if a team has INCREDIBLY rare back luck after achieved something through great talent, skill, effort and concentration.

In other words, the 99½% of times a team made an "extra" point they weren't really getting anything more for that TD than every team got for practically every TD: The ½% of misses just DENIED them a point for being in the wrong place at the wrong time. Those are almost never blocked (how many players go all out trying to block a PAT unless they're within a point of winning/losing at the end of a game?) Nearly all are down to bad footing, bad hold, loose sod, wet grass, wet ball, fluke wind gust or something else beyond the control of any of the 53 people punished for their TD.

That's bad enough, but sometimes it DECIDES GAMES. That means draft picks, can (and has) meant playoff berths, but what if it were IN the playoffs: What if the ½% of missed PATs decided a SUPER BOWL?

Make it LITERALLY automatic, and if a team wants (or needs) to roll the dice for a genuine EXTRA point, they can. They'll make it 50% of the time, so statistically lose nothing, but the variance from the sure thing is enough teams would only try it when they had more to gain than turning 31-10 into 32-10 or 31-11.

Originally Posted by Dapper Dan

I'm waiting on someone to do a fake PAT. People always talk about how it's sometimes harder to run plays from the 3 yard line because you have no room. If you back up a bit then maybe you can fade a TE or something. It should catch them off guard too.

This is actually one of the reasons I hate moving back the PAT: Teams SHOULD'VE gone for fake 2 PATs a LOT more than they did (they broke even at 50% against opponents EXPECTING it; how much would that go up against if they DIDN'T expect it?) Now it's a moot point, because they can't send the kicking unit out at the 2 anyway (well, they can; they just won't get anything unless they pass or run into the end zone.) I doubt they're much likelier to fake it from the 15 (or wherever; for the first time in forever, I don't KNOW where the ball's spotted after a score.) Much more suspenseful now, right?