Here, we are concerned with the dotted black line; this represents the expected win probability of the defense given they intentionally allow a touchdown while up one point. If the current situation falls below the black dotted line, teams should intentionally allow the touchdown; if it is above, they should attempt to stop or block the field goal. Once the 49ers had the 1st-and-Goal, with the impending snap coming under the 2:40 mark, the Seahawks should have immediately attempted to allow the 49ers to score.
A couple notes:
1. Since the 1st down snap came at 2:39, the two-minute warning was effectively rendered useless - it basically gave the 49ers an extra timeout to call as the play clock wound down so they did not have to take a delay of game.
2. Since the 49ers were already inside the 10-yard line, they could not achieve another first down to continue to run the clock barring a penalty.
3. As is always the case, the 49ers would not have to score a touchdown. They could kneel down before the goal line (like Brent Celek did today) to maximize their win probability. But, trying to allow the 49ers to make the mistake by scoring the touchdown does not have a downside (well very tiny downside, the difference in probability between an 18 or 19-yard field goal and 24 or 25-yard field goal is roughly 1.5%).
Now, let's look specifically at the two situations:
1. Seahawks intentionally allow a TD: According to Brian's win probability estimates, that leaves the Seahawks with a 20-25% chance to win the game if the touchdown is allowed right away. Seattle would have over two minutes and the two-minute warning to drive down and score a touchdown. This would also be dependent on a San Francisco two-point conversion - if they fail, a Seahawks TD wins the game, if they convert, a Seahawks TD ties the game. Still, getting the ball with 2:30 left down 7 is roughly a 19% proposition to win the game.
2. Seahawks try to stop San Francisco: If San Francisco runs the ball three times before kicking the field goal, the time calculator estimates Seattle gets the ball back down two with 0:24 seconds remaining (they actually ended up with 0:26 seconds). Yes, Seattle could force a fumble (~1% probability). Yes, the 49ers could miss or have their kick blocked (< 3% chance, which decreases marginally as the 49ers approach the goal line). Taking everything into consideration, the Seahawks win probability would max out around 10-11% if they attempt to stop San Francisco:
P(FG Miss or Block) * 100% + P(Fumble) * 100% + P(FG Make) * P(Win down 2 with ~0:24 remaining)
= 3% * 100% + 1% * 100% + 96% * 7% 10.7%
In all likelihood, this is actually much closer to 4-7% as that 7% win probability given the situation is based on an average number of timeouts remaining. Either way, by intentionally allowing a touchdown, the Seahawks would at least double their chances of winning the game with very minimal downside.
Keith Goldner is the Chief Analyst at numberFire.com - The leading fantasy sports analytics platform - and creator of Drive-By Football. Follow him on twitter @drivebyfootball or check out numberFire on Facebook.
Yes--but, Vikings-Ravens.
as noted before, the defense cannot force the offense to score a TD.
one assumes both teams read this website, and therefore SF would take a knee at the 1 yard line.
But that's notable in and of itself, wouldn't you agree? I have a hunch SF knew they were inside the choke hold zone based on how Gore tried to stay in bounds after his long run to put them in FG (attempt) range.
There's also something to be said for briefing defenders not to make a tackle once a ball carrier makes it inside the 20 (or insert the appropriate yd line).
It actually appeared that SF was trying not to score a TD (...or, at least, not trying to score a TD). I thought about this from the moment I saw Gore go down without really being tackled on the 51 yard run. While he did hobble a little and there was conversation by Joe and Troy about his injury, it seemed to me like perhaps Jim Harbaugh was onto the whole idea from the start. Going down at about the 15 like he did, in bounds, keep clock running at about 4 min to play and allows for another first down (which they did convert with CK running) and the "FG choke hold." All that said, it did NOT look like the Seahawks were trying to allow the TD. I knew right away that Brian or Keith would jump on this situation and am glad that you guys stepped up to the plate just as you always do with spot-on analysis. Well done.
I swear Brian's comment was not there when I began to craft my post....
also, the SF 49ers are the 3rd best scoring defense in the nfl, giving up only 16.5 points per game. They have only allowed 18 TDs in 13 games.
are you really suggesting that the best idea is to assume you can score a TD on them with 2 minutes left?
why would you base any decision on a league wide 10 year average, when you know that you are playing the 2013 san francisco 49ers? this seem to be the perfect example of 'local effects trump league averages'.
"based on how Gore tried to stay in bounds after his long run to put them in FG"
it's called running out the clock.
PS. i critique out of respect, and out of great interest in this subject. my comments are valid criticisms, that hopefully will improve the analysis when these points are addressed.
I think these "advanced" analysis are very helpful in analyzing past games. However, I have not been convinced that there is any predictive value in it, in terms of play by play decisions like the subject of this article.
note: as everyone knows, this "let the offense score" idea actually happened, because belichick did it, in the superbowl. it's not like this subject is a new idea. and the giants knew it too and tried not to, but the rb "accidently" fell into the endzone.
Anon-
Yes. It's either try to score a TD in 2 minutes, or kick a field goal in 24 seconds with no timeouts (against the same defense...). Time is more valuable here.
As for basing a decision of 10 years of data, better to have information to make an informed decision than no data at all. While you of course can't just go by the numbers (which no one here suggests you should) as circumstances do affect the probabilities, good information is always better for smarter decision-making.
Anon, you're right that local effects matter, but you're ignoring 1) Seattle's second best offense, and 2) the best offenses are better than the best defenses.
For instance, although it seems like Seattle's 3rd best offensive net yards per attempt should mostly cancel out with SF's 4th best defensive NYPA, this isn't the case. Sea's O is 1.3 NYPA above average while SF's D is only 0.7 NYPA below average. Combine those two and Sea's O vs SF's D is equal to the 9th best passing offense against an average defense.
Since the WP estimates assumes an average offense against an average defence, once you factor in team strengths it favors Seattle allowing the TD even MORE!
http://thebiglead.com/2013/12/09/should-pete-carroll-have-let-san-francisco-score-before-the-two-minute-warning/
That's really shitty that Jason Lisk would write that without citing the research done here by Brian. Really goes out of his way to torture the methodology to avoid citation.
He should be ashamed.
Are you going to analyze the Mike Tomlin decision to go on fourth down in the Steelers/Dolphins game? Would love to see that analysis.
PIT 4th-and-10 from their 20 was a toss up, ~11% chance to win whether you punt or go for it.
As a Seahawks fan I was screaming for us to let them score after Kaep got the 1st down. There's a large debate in the Seahawks community right now on this topic, and I appreciate the data to support what I felt and still feel was the obvious best decision.
What about an intentional foul to make it 1st and Goal from the 9 instead of 1st and 10 from the 18 (both with 2 timeouts and 3:37 left)? The ANS calculator certainly likes it.
The chance that the field goal is missed is quite low, even from the 18, and a shorter field makes it (marginally) easier for the defense.
Another interesting point would be the odds of a very good team like the Seahawks beating another very good team like the 49rs 3 times in one season. Looking at past history, doesn't this one road loss to their biggest rival improve the chances of the Seahawks beating them in the post season if they meet?
Chuck, as of 2010, teams that were 2-0 against an opponent in the regular season were 13-7 against that team in the playoffs. Keep in mind chances are the 2-0 team was likely the better team overall and likely had home field advantage as well.
Chuck: I'm not sure what your logic is there, that it is somehow better to be 1-1 than 2-0 against an opponent if you meet up with them a third time. Do you mean, like you shouldn't bet on red if the roulette wheel came up red 6 times in a row, because the odds of 7 reds in a row are very small? Or does beating an opponent twice somehow make a team cockier, and more likely to take the opponent lightly the third time? Anyway, James already pointed out that beating a team twice means you are quite likely to beat them a third time.
Note that, by definition, teams that are 1-1 against an opponent have a 50% chance of winning when they meet them a third time. Because both teams are 1-1 when they meet the third time.
Two points:
1) slightly off topic, but I haven't heard any discussion about Kaep's first down run. While the Seahawks did not have any timeouts to challenge, it seems pretty clear that Kaep's knee was down well short of the "yellow" first down line. On the one TV replay it's a bit difficult to see where the ball was, but it did look short. Not that it mattered, but just wondering why no discussion of this potentially bad call.
2) While the win probability analysis and discussion about the strength of each teams offense and defense is interesting, I think it confusing the decision. Let's just assume that it's highly likely that SF will not fumble and the short FG will be successful. Assuming any other outcome here is just 'hoping' for a lucky break and the decision should not be based on a lucky break. What's more, expecting to make a field goal with 25 seconds or so remaining is also going to rely on a lucky break such as a penalty, a pass or two being completed into double coverage AND the player going out of bounds; probably needs to happen twice, etc., etc. I think we can all agree that the only way for Seattle to win if they don't allow SF to score would be relying on complete luck and taking any skill factor they have out of the equation. Conversely, if Seattle does allow SF to score, and they take the gift, Seattle now has 2+ minutes to drive the length of the field. This is what they prepare for, they have great skill players, and RW and company have been very successful scoring late in games or halves when they have to -- can use all 4 downs and aren't trying to run time off the clock with their running game. We have seen RW do it time and time again -- the one exception I can think of is the Colts game, and even there, they were greatly hampered by a questionable no call on what looked like a clear PI.
So there you have it? Do you want to rely on pure luck, taking the game out of your hands, or do want a chance to put the outcome of the game back in your hands.