[BostonCard]

So I ran the numbers sampling from the actual data from the game against Cal (with the caveat that the sample size especially for the running plays, was small) To be fair, I considered sacks and Mills scrambles to be pass plays as I assume they were not designed runs.  Pretty much regardless of what we called, the probability of "success", defined as getting a first down is between 75 and 80% (because it is a simulation, it fluctuates a bit).  The probability of getting a touchdown (a play of >26 yards) on that set of downs was low regardless of what we did, but zero if we ran twice (we did not have two running plays that together added up to 26 yards), 15% if we passed twice, and about 6% if we passed once.  Of course, a fresh set of downs would have offered a chance to score a touchdown from somewhere inside of the Cal 22, but I didn't want to program and run that.  The flip side of that is that around 7% of the time Mills will throw an interception if he passes it twice.  You could make an argument that this is an unacceptable risk at that stage of the game.  







That being said, to the extent that previous plays in the game are representative of what might have happened, and taking sample size into consideration, it is not clear to me that the numbers prove that one or the other call is clearly a better strategy.





BC

I am taking teejers side on this one, at least mostly.



I see a couple of problems with much of the discussion above.

First of all, the numbers you sample from this game are fairly meaningless; too small a sample size, all runs are not the same play - even more so all passes are not the same play, defenses were both tiring, etc.



Secondly, I don't think the issue is so much run vs pass, but whether the immediate tactical goal should have been to make a first down or take the chance for a quick TD. 



I'd argue that the primary goal at that moment in time should have been a first down. We were already in FG range, making the key element to run clock. cal isn't a quick strike team, but it was clear Stanford's chances to hold them for very long weren't particularly high. Taking time off the clock was critical.  A first down gives at least another three downs to run clock, perhaps in addition to the second down play itself.



A first down would also seem rather achievable goal at that point as well, since it was second and four (a short four).



That doesn't necessarily mean a run play however, given how bad Stanford's run game has been. It could have meant a short safe pass with a much higher completion probability than the end zone shot we took.



Obviously a possibility is that either two runs don't get a first down, or a short pass still ends up incomplete and we're back where we ended up anyway, but I still think the fundamental tactical imperative at that moment in time was a first down rather than the chance at an immediate TD.



So, I'd say it was a tactical mistake.

It could have worked out anyway, and the opposite could still have ended in a loss. It's probably not the most critical issue surrounding the game, but still a mistake I think.

I don't disagree inasmuch as your argument introduces some subtlety that teejers' doesn't.  His point is that we should have run the ball twice and that is obvious to everyone who knows anything about football.  My argument is not so much that passing is better than running, but rather that both options are defensible.

In looking at the situation, the goals of the drive at that point are twofold: milk as much time off the clock (certainly once we got first and 10 with about 3 minutes left; Murky does have a point that it is hard to run a 7 1/2 minute drive in the best of circumstances, and the worst of circumstances is the one that hands Cal the ball with just enough time to mount the winning drive) and score points (ideally a TD, but a FG is better than nothing, and a FG as time expires is better than a TD any time).  The best way to do that is to get a fresh set of downs, to force cal to use up its timeouts and to drive the clock down to less than a minute left.  I would say that scoring (especially a TD) would be preferable to a FG attempt with less time on the clock (that is, you would not instruct your player to take a knee at the 1 inch line in order to be able to milk more time off the clock).

I haven't had a chance to look at the play in detail, but the play may not have been designed specifically to try to score a TD.  There were likely multiple receiving options, one of which was short and one of which was in the end zone.  If I were the one giving the QB instructions, I would say try to hit the short receiver if you can, but don't force it, and if you see a favorable, low-risk, match-up in the end zone take a shot.  But I'm not privy of any discussions.

BC

[BostonCard]

So I ran the numbers sampling from the actual data from the game against Cal (with the caveat that the sample size especially for the running plays, was small) To be fair, I considered sacks and Mills scrambles to be pass plays as I assume they were not designed runs.  Pretty much regardless of what we called, the probability of "success", defined as getting a first down is between 75 and 80% (because it is a simulation, it fluctuates a bit).  The probability of getting a touchdown (a play of >26 yards) on that set of downs was low regardless of what we did, but zero if we ran twice (we did not have two running plays that together added up to 26 yards), 15% if we passed twice, and about 6% if we passed once.  Of course, a fresh set of downs would have offered a chance to score a touchdown from somewhere inside of the Cal 22, but I didn't want to program and run that.  The flip side of that is that around 7% of the time Mills will throw an interception if he passes it twice.  You could make an argument that this is an unacceptable risk at that stage of the game.  







That being said, to the extent that previous plays in the game are representative of what might have happened, and taking sample size into consideration, it is not clear to me that the numbers prove that one or the other call is clearly a better strategy.





BC

I am taking teejers side on this one, at least mostly.



I see a couple of problems with much of the discussion above.

First of all, the numbers you sample from this game are fairly meaningless; too small a sample size, all runs are not the same play - even more so all passes are not the same play, defenses were both tiring, etc.



Secondly, I don't think the issue is so much run vs pass, but whether the immediate tactical goal should have been to make a first down or take the chance for a quick TD. 



I'd argue that the primary goal at that moment in time should have been a first down. We were already in FG range, making the key element to run clock. cal isn't a quick strike team, but it was clear Stanford's chances to hold them for very long weren't particularly high. Taking time off the clock was critical.  A first down gives at least another three downs to run clock, perhaps in addition to the second down play itself.



A first down would also seem rather achievable goal at that point as well, since it was second and four (a short four).



That doesn't necessarily mean a run play however, given how bad Stanford's run game has been. It could have meant a short safe pass with a much higher completion probability than the end zone shot we took.



Obviously a possibility is that either two runs don't get a first down, or a short pass still ends up incomplete and we're back where we ended up anyway, but I still think the fundamental tactical imperative at that moment in time was a first down rather than the chance at an immediate TD.



So, I'd say it was a tactical mistake.

It could have worked out anyway, and the opposite could still have ended in a loss. It's probably not the most critical issue surrounding the game, but still a mistake I think.

I don't disagree inasmuch as your argument introduces some subtlety that teejers' doesn't.  His point is that we should have run the ball twice and that is obvious to everyone who knows anything about football.  My argument is not so much that passing is better than running, but rather that both options are defensible.

In looking at the situation, the goals of the drive at that point are twofold: milk as much time off the clock (certainly once we got first and 10 with about 3 minutes left; Murky does have a point that it is hard to run a 7 1/2 minute drive in the best of circumstances, and the worst of circumstances is the one that hands Cal the ball with just enough time to mount the winning drive) and score points (ideally a TD, but a FG is better than nothing, and a FG as time expires is better than a TD any time).  The best way to do that is to get a fresh set of downs, to force cal to use up its timeouts and to drive the clock down to less than a minute left.  I would say that scoring (especially a TD) would be preferable to a FG attempt with less time on the clock (that is, you would not instruct your player to take a knee at the 1 inch line in order to be able to milk more time off the clock).

I haven't had a chance to look at the play in detail, but the play may not have been designed specifically to try to score a TD.  There were likely multiple receiving options, one of which was short and one of which was in the end zone.  If I were the one giving the QB instructions, I would say try to hit the short receiver if you can, but don't force it, and if you see a favorable, low-risk, match-up in the end zone take a shot.  But I'm not privy of any discussions.

BC