[stupac2]

In this case, that it wasn't likely to make a difference was obvious at the time the decision was made.
In fact, presumably Yikes thought that the only chance of some kind of miracle come-back was a shake up in the offense.
And, to the extent it did make any difference, it at least coincided with c.a.l.'s offense perking up a tad.

If you actually follow that thinking to its logical conclusion (short of nihilism, anyway) then they should've just been throwing deep on every single play since extended drives clearly weren't working, may as well go for home runs. May as well also go for every 4th, which they didn't do. If you're actually doing the "throw everything at the wall and see what sticks" thing then okay, do it. But they didn't, it was just haphazard and pointless.

p.s. I also disagree with the standard board parsing of whether a decision is good or not by multiplying together a bunch of percentages that not only are unknown, but are unknowable. Nobody bothers propagating the uncertainty in these numbers, which in most cases of interest are going to make all decisions equal.

First, there's no way that's standard. I think besides myself, WiM, and BC no one here really thinks that way. Second, it's like that famous saying about democracy, it's the worst system, except for all the others. Thinking statistically about decision-making grounds you in something tangible instead of relying on vague concepts that may or may not relate to anything at all (things like "giving the offense a spark", for instance, what does that even mean?). There's a reason that coaches who are actually aware of the stats on things like punting vs going for it win more games than the ones who don't. Ron Rivera last year was a great example of that in a single person.

Anyway, in the big picture you're right, taking Goff out probably moved Cal's win percentage from like 2% to 0.5% or something, but even though that's a small absolute difference it's a big multiplier. It doesn't really matter much what the numbers are, or that Cal was pretty solidly screwed no matter what they did, but that if Dykes wanted to give his team the best chance of winning he shouldn't have given the other guy as many snaps as he did. There's obviously a place for taking crazy risks when you're down, but there's just no reason to think that's what he was doing.

[stupac2]

In this case, that it wasn't likely to make a difference was obvious at the time the decision was made.
In fact, presumably Yikes thought that the only chance of some kind of miracle come-back was a shake up in the offense.
And, to the extent it did make any difference, it at least coincided with c.a.l.'s offense perking up a tad.

If you actually follow that thinking to its logical conclusion (short of nihilism, anyway) then they should've just been throwing deep on every single play since extended drives clearly weren't working, may as well go for home runs. May as well also go for every 4th, which they didn't do. If you're actually doing the "throw everything at the wall and see what sticks" thing then okay, do it. But they didn't, it was just haphazard and pointless.

p.s. I also disagree with the standard board parsing of whether a decision is good or not by multiplying together a bunch of percentages that not only are unknown, but are unknowable. Nobody bothers propagating the uncertainty in these numbers, which in most cases of interest are going to make all decisions equal.

First, there's no way that's standard. I think besides myself, WiM, and BC no one here really thinks that way. Second, it's like that famous saying about democracy, it's the worst system, except for all the others. Thinking statistically about decision-making grounds you in something tangible instead of relying on vague concepts that may or may not relate to anything at all (things like "giving the offense a spark", for instance, what does that even mean?). There's a reason that coaches who are actually aware of the stats on things like punting vs going for it win more games than the ones who don't. Ron Rivera last year was a great example of that in a single person.

Anyway, in the big picture you're right, taking Goff out probably moved Cal's win percentage from like 2% to 0.5% or something, but even though that's a small absolute difference it's a big multiplier. It doesn't really matter much what the numbers are, or that Cal was pretty solidly screwed no matter what they did, but that if Dykes wanted to give his team the best chance of winning he shouldn't have given the other guy as many snaps as he did. There's obviously a place for taking crazy risks when you're down, but there's just no reason to think that's what he was doing.