The Game-Theory Gods Were Nauseated Yesterday

Submitted by oakapple on

Three of yesterday’s games featured atrocious game-theory decisions by the losing coaches. None should have been even close calls, and yet the coaches got them wrong. I don’t just mean “wrong,” in that the strategy turned out badly, but in that no coach knowledgeable in the odds should have made those decisions.

In the Outback Bowl, Georgia was in the driver’s seat, as its defense intercepted Michigan State QB Kirk Cousins in the first overtime. All the Bulldogs needed was a field goal to win. On 1st & 10, the Bulldogs gained 2 yards. On 2nd & 10, coach Mark Richt called a run, to position the ball in the middle of the field, which lost 2 yards. He then attempted a field goal on 3rd down, which Kicker Blair Walsh missed wide right. Michigan State went on to win in triple overtime.

Walsh was 19-31 (61%) on the year, so his 42-yard attempt was no gimme. (Walsh also missed a 47-yarder in the 3rd OT that would have extended the game.) Had Richt elected to run a real play, even a conservative one, on 3rd down, there is almost certainly a far better than 61% chance that it would have gained positive yards, either making a first down or making the FG attempt on 4th down an easier one.

Except in situations where there is very little time left (which is not a factor in OT), it hardly ever makes sense to attempt a field goal before 4th down.

In the Rose Bowl, Wisconsin took over at its own 13 yard line with 23 seconds to play, trailing by 7. Incredibly, Oregon played practically no defense at all, allowing Badger QB Russell Wilson to complete a 29-yard pass to a receiver who then went out of bounds, and a 33-yard pass to a receiver who remained in-bounds. There were 2 seconds left, with the ball on the Oregon 25 yard line. Wisconsin could run one more play, but they had to line up in a hurry. Incredibly, coach Bret Bielema called for Wilson to spike the ball, stopping the clock at 0:00, ending the game.

For Bielema’s strategy to work, the play to spike the ball needed to take just 1 second. Since any play must take some time, it can only work if the center snaps the ball the very instant the referee starts the clock. That is extremely difficult to do, as Michigan and Washington State fans will recall from the 1998 Rose Bowl, when Cougar QB Ryan Leaf attempted to do the identical thing, with the identical result. The “spike-the-ball” strategy almost always takes two seconds, something every coach should know.

Of course, it was no certainty that the Badgers would score from the 25 yard line, but Bielema’s decision deprived them of the chance to try.

In the Fiesta Bowl, the Stanford Cardinal had a 1st & 10 from the Oklahoma State Cowboys’ 25 in a tie game, with 52 seconds to play. With the Heisman Trophy runner-up Andrew Luck at QB, coach David Shaw didn’t even try to make another first down, electing instead to drain the clock and attempt a 35-yard field goal on 3rd down as time expired.

The problem with playing for a FG, is that the kick is not automatic, and you’re foregoing the chance to get more. Just like the Georgia game, the FG was far from certain: kicker Jordan Williamson had already missed a 41-yarder earlier in the game, and he would miss a 43-yarder in overtime. The odds were much more in the Cardinal’s favor if coach Shaw had tried to get another first down or two, which with 52 seconds remaining there was plenty of time to do, and Andrew Luck had been nearly perfect all day.

In OT, Shaw once again took the ball out of Luck’s hands, losing 2 yards on two running plays (with a false start sandwiched in between), before calling a safe pass on 3rd down that set up kicker Williamson’s third miss of the evening.

rbgoblue

January 3rd, 2012 at 10:33 AM ^

I actually can understand Stanford's decision to let the clock run at the end of the game.  IIRC, they were lining up for about a 35 yd FG, which should be very makeable.  Oklahoma St. had a time out left as well, so running additional plays could have put them in a position where they would be forced to kick a 4th down FG with about 30 seconds left, giving a very high powered Ok St. offense (with Justin Blackmon!) a chance to drive for the win or tie.  By running out the clock, they guaranteed overtime with what should have been a very good chance to win.

That said, Mark Richt's playcalling in OT was moronic.

rbgoblue

January 3rd, 2012 at 10:35 AM ^

Also, I didn't understand Okie State's decision to take the ball off the 6 inch line on 1st and Goal to set up the chip shot for the win in OT.  Why put the game in the hands of a kicker (anyone remember Iowa St) without taking at least a couple shots to sneak it in?

rbgoblue

January 3rd, 2012 at 10:59 AM ^

two totally different situations, boss.  stanford's decision was that a 35 yd fg with no chance for okie state to answer gave them a better chance to win than slightly improving their fg odds but leaving justin blackmon with a chance to answer.

okie state was on the 6 inch line.  i would think that trying the qb sneak would make more sense that risking a bad snap/blocked kick/shank and settling for the fg automatically.

Inertia Policeman

January 3rd, 2012 at 2:21 PM ^

All I'm saying is that continuing to run your offense does not mean that OK State automatically has a chance to answer. Luck hadn't thrown an incompletion on that drive. Maybe they get 10 yards closer and then run out the clock, or better yet, maybe they just don't go all puntasaur and end up scoring a touchdown, which is what I think would have happened after watching Standford effortlessly march down the field. Even though OK State's offense is great, I'd rather make them go 80 yards in 30 seconds to tie than trust the game to a freshman kicker on a far from automatic field goal.
Besides, we all know only one team in the NCAA can go 80 yards in 30 seconds and score a touchdown.

Ali G Bomaye

January 3rd, 2012 at 2:56 PM ^

Kicking from the center of the field on the 3 yard line makes the kick equivalent to an extra point.  OK State's kicker had made 79/80 extra points on the year (98.8%).  It's reasonable for Gundy to think that they had more than a 1.2% chance of turning it over if they were to run a play.

Inertia Policeman

January 3rd, 2012 at 10:36 AM ^

When Stanford decided to play for the field goal with a first down at the OK State 25, 3 timeouts, 52 seconds left, and ANDREW LUCK AT QB, I knew they were going to lose. Luck was 5 of 5 on that drive, and you decide to not let him throw again. Simply absurd. I don't know about you, David Shaw, but I would trust my future number one overall draft pick quarterback over my freshman kicker in that situation. 

ND Sux

January 3rd, 2012 at 10:54 AM ^

Sheesh, my TV almost didn't survive that one.  I understand not throwing for a 1st down when all you need is a FG to win, but to "position" the ball on 2nd down was utter pussification.  Hell, even two QB sneaks would have netted at least two yards and accomplished the same thing.

That said, the coaches blew the game even before OT.  With a 16 point lead, they basically did NOTHING on offense in Q4.  On the final MSWho drive, they executed the dreaded three-man-rush, allowing Cousins (who hadn't done much all day w/ normal pressure) to drive for the tie. 

Georgia Tech coaching staff managed a similar debacle, IMO. 

Ziff72

January 3rd, 2012 at 11:05 AM ^

I like your post basically because I posted it yesterday.   Yet I don't see any of the usual suspects coming in to say your post is stupid.   So I have to surmise  the attacks on my post have become personal.

Gene

January 3rd, 2012 at 11:11 AM ^

I don't think kicking on 3rd down is in itself that crazy: you eliminate the risk of a penalty making the kick harder, and allow for the risk of a botched snap (as you would still have another chance to kick.) That said, it only makes sense when the kick is an easy one where gaining 5 yards isn't going to make any real difference. At 42 yards, it's just stupid.

StoneCold

January 3rd, 2012 at 1:00 PM ^

I've never seen the 'let's kick on 3rd down just in case something bad happens so we can still kick on 4th down' strategy play out.  Sure, plenty of these 3rd down kicks have been made but I can't recall a botched snap and then a kick on fourth down.

In this instance, let's say it was a botched snap and Georgia recovered at the 32 yard line.  Now, they're looking at a 49 yard field goal on fourth down.  That's far.  (Although, Gibbons would have drilled it.)  Georgia should've just ran a play on third down and been content with gaining a few yards for their kicker.   

Ali G Bomaye

January 3rd, 2012 at 2:59 PM ^

The botched-snap argument is pretty weak.  When was the last time you saw a botched snap on a FG where the holder had the wherewithal to fall on the ball cleanly and set up another attempt?  And with the best QB in college football, I'd take the risk of a penalty reducing the odds of the kick by a few percent to gain the upside of gaining 10-15 yards and making the kick significantly easier.  As you said, at 42 yards, not going for more is just stupid.

Gulo Gulo Luscus

January 3rd, 2012 at 12:09 PM ^

has anyone mentioned d'antonio's 2 point conversion decisions?  that game is over in regulation if he just kicks the XP all 3 times instead of going 1/3 on the 2 pointers.  i'm with the mathlete (game theory manifesto), who argues:

Rule #1 of 2 point conversions is don’t even think about them until the fourth quarter. No exceptions.

in this case, i actually think the one in the 4th quarter was the worst of all.  trying to go up 3 instead of 2 makes sense if the opponent has 1 possession left or you don't think you'll get the ball back, but at that point MSU was dominating and if you're up 2, another TD and XP pretty much seals the deal.  alas, sparty pulled it out and richt is the one taking heat.  i tend to agree, although there was a tweet from the mathlete that confused me a bit:

Odds of making a 42 yard FG, 56%, odds of making even a 32 yard FG, 79%. Easy call to kick it there and not try for a single yard.

honestly i couldn't tell if this was sarcasm or not, but the mathlete doesn't appear to be a particularly sarcastic guy.  it's possible he disagrees with all of us regarding richt's OT decision making right there.

 

 

El Jeffe

January 3rd, 2012 at 2:00 PM ^

No, I think your sarcasm interpretation was right, though it is hard to tell with the twitters. I think the penultimate word "single" gives it away, as does increasing your chances of winning by (79/56 -1) * 100 = 41.1%!!!!

Not to say that UGA would necessarily have picked up those 10 yards, but still.

raleighwood

January 3rd, 2012 at 12:12 PM ^

Dantonio went for two with eight minutes to go in the 3rd quarter.  He might get a bit of a pass because his team was not able generate much offense up to that point so he probably planned on two scoring drives instead of three but it was a clear violation of the 4th Quarter rule.  Sparty could have won in regulation if Dantonio followed the rules and kicked after all three TD's.

 

Braylon 5 Hour…

January 3rd, 2012 at 12:15 PM ^

What seems to also be going unmentioned about Stanford (although not precisely a game theory issue I guess) is that in Overtime they get the ball immediately after this redshirt freshman kid botches a field goal attempt.  They run for 5 yards on first and 10, but are back at 2nd and 10 after a fall start and they run the ball again for a 3 yard loss.  3rd and 13, Andrew Luck throws a check down pass that gains 3 yards, and you put the same kid who just botched a much more makeable field goal in to attempt a 42 yarder probably no more than 5 minutes of real time after he already botched it in regulation.  You have Andrew Luck, why are you bailing out the defense??? Overtime you need to be playing for TDs, especially when you have the ball first, not setting up an awkwardly distanced field goal for a mediocre kicker.

BlueHills

January 3rd, 2012 at 1:09 PM ^

Agree. I could hardly believe that with so much time on the clock, Stanford, with perhaps the best "pro style" quarterback in the country, simply left it up to a kicker.

Has to be nuts.

joeh200

January 3rd, 2012 at 1:18 PM ^

None of the bad decisions shocked me like Richt's yesterday. Kicking on 3rd down isthe right decision in a lot of cases. The throw away 2nd down play, however, was a great example of why Richt is always mentioned as one of the coaches on the hot seat. Richt wasn't coaching to win there, he was coaching not to lose, and that bites you every time.

Tater

January 3rd, 2012 at 1:19 PM ^

1.  I hate "prevent defense."  Insert tired "adage" here.

2.  Coaches don't get retrospect in realtime.  The feces is usually in contact with the rotary blade in realtime.

3.  It all comes down to success.  If it works, it was a great decision.  If it doesn't, it's time for "firecoachblahblah dot com."

Number 7

January 3rd, 2012 at 2:08 PM ^

But I do believe Wsionsin's first punt of the day came from on 4th and 3 from the Oregon 43 (or so).  Perhaps they were worried about not making it and giving up a quick scoring drive.  Well, they downed the punt at the 5 . . .  and two plays later DeAnthony (not that -- well, actually yes, THAT DeAnthony) was breaking Tyrone Wheatley's Rose Bowl record for longest TD run.

 

[Edit -- Hey!  The Mathlete said the same thing!  But oops -- he posted it 26 minutes earlier.  My bad.]

jmblue

January 3rd, 2012 at 3:56 PM ^

I'd be curious to know how Richt's decision to run the ball on 3rd and 8 with 2 minutes left in regulation affected his team's odds of winning (per the Mathlete).  That seemed like a really dubious decision.  MSU had a timeout, so it burned no clock and (given UGA's consistent failure to run the ball) was practically guaranteed to result in a punt.  All it functionally accomplished was to get MSU to burn a timeout, but with 2 minutes left, they still had plenty of time.