[stupac2]

http://insider.espn.go.com/college-footb...e-football

For those that don't have insider, it's basically looking at ESPN's EPA and QBR, and throwing on some windows ("since Oct 7") to show that ASU is peaking while we're sliding. What's a bit odd is that there's no attention paid whatsoever to the fact that the two teams have already played, it's not even mentioned in the article, which is sort of baffling to me. These teams haven't changed all that much since that game, you should address how Stanford managed to demolish ASU if you're saying that ASU is clearly better.

That said, it's clearly a scary game, especially if Arizona can't magic us up another upset.

[stupac2]

http://insider.espn.go.com/college-footb...e-football

For those that don't have insider, it's basically looking at ESPN's EPA and QBR, and throwing on some windows ("since Oct 7") to show that ASU is peaking while we're sliding. What's a bit odd is that there's no attention paid whatsoever to the fact that the two teams have already played, it's not even mentioned in the article, which is sort of baffling to me. These teams haven't changed all that much since that game, you should address how Stanford managed to demolish ASU if you're saying that ASU is clearly better.

That said, it's clearly a scary game, especially if Arizona can't magic us up another upset.

[oman]

ASU just looked meaner to me this last weekend than they did against us.

Maybe it is those all whites that look kind of intimidating.

[oman]

ASU just looked meaner to me this last weekend than they did against us.

Maybe it is those all whites that look kind of intimidating.

[Griffins78]

The 4th quarter of our game against ASU was significant - in that I think it gave ASU hope and kept them from going into depression. They came into our game with high hopes that we obliterated in the first half. But they didn't give up and came back in the second half fighting and hung in during the 3rd quarter. Then they caught fire in the 4th quarter and it restored a belief in themselves that I think has made the difference in their season.

They are a different team now than the one we faced in the first half. They are the team that we played in the second half - that scored more points (28) in a single half than any team has scored against us in an entire game this season (except UW which scored 28).

This is going to be a tough game. Especially if we play in Tempe.

[Griffins78]

The 4th quarter of our game against ASU was significant - in that I think it gave ASU hope and kept them from going into depression. They came into our game with high hopes that we obliterated in the first half. But they didn't give up and came back in the second half fighting and hung in during the 3rd quarter. Then they caught fire in the 4th quarter and it restored a belief in themselves that I think has made the difference in their season.

They are a different team now than the one we faced in the first half. They are the team that we played in the second half - that scored more points (28) in a single half than any team has scored against us in an entire game this season (except UW which scored 28).

This is going to be a tough game. Especially if we play in Tempe.

[washingtonismoney]

While I don't disagree with the ultimate conclusion, the article demonstrates nicely why stats aren't what's important, it's the ability to perform meaningful analysis with them. Joyner, the guy who wrote the article, can be remembered from such legendary pieces of analysis as "Stanford's wide receivers -- going to be good!" (about the 2011 Stanford offense) and "Matt B'arkley -- better than Luck!" The hallmarks of Joyner's analyses are an inability to distinguish why small sample sizes are bad, a love of selective stats, and wild extrapolations.

This article -- which unfortunately I can't access at the moment -- shows all of these things. For example, Joyner writes about the total QBR of the team -- which ASU exceeds Stanford by a healthy margin at. But that's not a great take on the stat: most of the time, with quarterback-related statistics, you care about the stats of the best QB rather than all of the QBs. Chopping it that way reveals that Hogan's adjusted QBR (78.8 ) is virtually indistinguishable from Kelly's adjusted QBR (79.1). Viewed naively, there's not much reason to prefer one to the other. (Of course, we know that Hogan's stats vary greatly based on whether he's at home on the road. Kelly is a bit better in that respect.)

The other silly thing is naively using win-loss record. When evaluating performance, you need to ask a few questions -- what was the strength of schedule? The margin of victory? The other key question to ask, when putting statistics in context, is whether there are any anomalous factors affecting the statistics. A computer sees ASU's 20 point drubbing of SC and thinks, "Wow, that's one of the best victories of the year." A human knows that that victory precipitated a firing that completely changed SC's performance. So I would've liked to see Joyner, at the very least, chop up the EPA statistics to see whether the idea that ASU is peaking (or not) is actually verified by more performance-level statistics. But given Joyner's other silly analyses, I'm not exactly shocked he didn't. (My favorite example was the Stanford-wideouts-actually-good piece: he based it on, like, a sample of 10 throws during the spring game. )

[washingtonismoney]

While I don't disagree with the ultimate conclusion, the article demonstrates nicely why stats aren't what's important, it's the ability to perform meaningful analysis with them. Joyner, the guy who wrote the article, can be remembered from such legendary pieces of analysis as "Stanford's wide receivers -- going to be good!" (about the 2011 Stanford offense) and "Matt B'arkley -- better than Luck!" The hallmarks of Joyner's analyses are an inability to distinguish why small sample sizes are bad, a love of selective stats, and wild extrapolations.

This article -- which unfortunately I can't access at the moment -- shows all of these things. For example, Joyner writes about the total QBR of the team -- which ASU exceeds Stanford by a healthy margin at. But that's not a great take on the stat: most of the time, with quarterback-related statistics, you care about the stats of the best QB rather than all of the QBs. Chopping it that way reveals that Hogan's adjusted QBR (78.8 ) is virtually indistinguishable from Kelly's adjusted QBR (79.1). Viewed naively, there's not much reason to prefer one to the other. (Of course, we know that Hogan's stats vary greatly based on whether he's at home on the road. Kelly is a bit better in that respect.)

The other silly thing is naively using win-loss record. When evaluating performance, you need to ask a few questions -- what was the strength of schedule? The margin of victory? The other key question to ask, when putting statistics in context, is whether there are any anomalous factors affecting the statistics. A computer sees ASU's 20 point drubbing of SC and thinks, "Wow, that's one of the best victories of the year." A human knows that that victory precipitated a firing that completely changed SC's performance. So I would've liked to see Joyner, at the very least, chop up the EPA statistics to see whether the idea that ASU is peaking (or not) is actually verified by more performance-level statistics. But given Joyner's other silly analyses, I'm not exactly shocked he didn't. (My favorite example was the Stanford-wideouts-actually-good piece: he based it on, like, a sample of 10 throws during the spring game. )

[76lsjumb]

Since we're talking about statistics, I'm surprised that no one has mentioned the one thing we can do that -- based on statistics, at least -- would seem to guarantee us a win against ASU.  That thing would be to...lose to N.D.

I believe that Shaw has NEVER lost 2 games in a row as head coach.  Thus, while his winning percentage is pretty good after wins, it is PERFECT after losses. 

Hey, don't look at me that way...  It's statistics.  ;D

[76lsjumb]

Since we're talking about statistics, I'm surprised that no one has mentioned the one thing we can do that -- based on statistics, at least -- would seem to guarantee us a win against ASU.  That thing would be to...lose to N.D.

I believe that Shaw has NEVER lost 2 games in a row as head coach.  Thus, while his winning percentage is pretty good after wins, it is PERFECT after losses. 

Hey, don't look at me that way...  It's statistics.  ;D