Sagarin currently ranks Cal 43th in the nation with a Predictor score of 75.18.

Looking at games already played and current Predictor scores, Sagarin had Cal going 2-3 over the first 5 games with wins over Sac State and Colorado. Cal has been 5.1 pts better than the average Predictor score YTD.

Here is how Sagarin predicts the remaining games:

Washington: Win +4.06
UCLA: Loss -7.48
Oregon: Loss -7.28
Oregon State: Loss -0.96
USC: Loss -14.18
Stanford: Loss -8.61
BYU: Loss -0.89

This would mean Cal would finish 5-7. However, if Cal maintains the trend of being 5.1 pts better than Predictor, Cal would finish with a 7-5 record.

I know this means squat, but interesting to see how one computer model views Cal.

LOL, these things get revised each game.

Here is another "predictor"



http://www.sbnation.com/college-football/2014/10/8/6945501/pac-12-football-standings-predictions-projection-2014

That near 23.9% for Colorado to go 0-9 gotta hurt

I think Sagarin might still be using data from last season, it might take another month for it to make a better prediction. The second predictor seems more reasonable at this point, it actually implies that Cal is the 6th or 7th best P12 team based on its predicted finishes of other conference foes.

0.8%!!!!!!!! So you're saying we have a chance

i can not see this team losing game after game like that .. just cant

Davidson;842378712 said:

---

0.8%!!!!!!!! So you're saying we have a chance

---

That's all I need -- a chance!

Go Bears! Beat the Huskies!

Cal88;842378709 said:

---

I think Sagarin might still be using data from last season, it might take another month for it to make a better prediction. The second predictor seems more reasonable at this point, it actually implies that Cal is the 6th or 7th best P12 team based on its predicted finishes of other conference foes.

---

Agree, Sagarin has not yet labeled his rankings "unbiased" so he is still relying on last year's data to some extent. For his model to work, he needs every team to be linked by common opponents. This takes awhile because too many teams play patsy pre-season schedules (and play different patsies), so it is impossible to know how good they really are. 5-0 Arizona could be the best team in the country for all we know. Our last minute loss to them on the road could mean we are Top 10 (not saying we are, but it makes as much sense as saying 3-2 Stanford and 3-2 USC are in the Top 20, which is where his current rankings have them).

we wont be favored that is fine ... but byu ? no hill come on

UCBerkGrad;842378680 said:

---

Sagarin currently ranks Cal 43th in the nation with a Predictor score of 75.18.

Looking at games already played and current Predictor scores, Sagarin had Cal going 2-3 over the first 5 games with wins over Sac State and Colorado. Cal has been 5.1 pts better than the average Predictor score YTD.

Here is how Sagarin predicts the remaining games:

Washington: Win +4.06
UCLA: Loss -7.48
Oregon: Loss -7.28
Oregon State: Loss -0.96
USC: Loss -14.18
Stanford: Loss -8.61
BYU: Loss -0.89

This would mean Cal would finish 5-7. However, if Cal maintains the trend of being 5.1 pts better than Predictor, Cal would finish with a 7-5 record.

I know this means squat, but interesting to see how one computer model views Cal.

---

You can't just count games that cal is favored as wins and games cal isn't losses. The right way to do it would be to estimate win probability for each matchup then compute the expected number of wins (in this case, add up win probabilities). If you do that there is a very good chance we at least get 6 wins

Tell this 30 year programmer how a computer program factors in a football team that loses a game on a hail mary but then has the balls to come back and win a home game on the last play of a shootout and then win another shootout game, this time away, on a last play missed field goal after a goal line stand. Impossible. I'd say there is no way for a programmer to factor in the grit of a team. In addition to a great offense this team has shown great grit. To go through what they did last year, to play game after game after game at an unparalleled break-neck pace and be 4-1 is pretty damn remarkable. You don't capture that kind of stuff looking at the box scores the next day and plugging data into a computer program.

Fellows, this is based a lot off last year's season. I tweaked a few things, including our hail mary loss, winning in double OT right after the painful loss, and winning yet another close one right after. I added all that and a few more data to the logarithm and calculations and below is a more accurate computation with random disputable numbers.

Washington: Win +10.06
UCLA: Win +5.23
Oregon: Win +3.28
Oregon State: Win +10.13
USC: Win +20.38
Stanford: Win +17.58
BYU: Win +6.8

:acclaim:
:gobears:

82gradDLSdad;842378740 said:

---

Tell this 30 year programmer how a computer program factors in a football team that loses a game on a hail mary but then has the balls to come back and win a home game on the last play of a shootout and then win another shootout game, this time away, on a last play missed field goal after a goal line stand. Impossible. I'd say there is no way for a programmer to factor in the grit of a team. In addition to a great offense this team has shown great grit. To go through what they did last year, to play game after game after game at an unparalleled break-neck pace and be 4-1 is pretty damn remarkable. You don't capture that kind of stuff looking at the box scores the next day and plugging data into a computer program.

---

Totally agree. I remember John Madden saying one morning on the radio: "On paper they are the better team; but the games aren't played on paper."

Whoa.....hold your horses! Are some of you saying computer models are not perfect?? Crazy talk.

ferCALgm2;842378746 said:

---

Fellows, this is based a lot off last year's season. I tweaked a few things, including our hail mary loss, winning in double OT right after the painful loss, and winning yet another close one right after. I added all that and a few more data to the logarithm and calculations and below is a more accurate computation with random disputable numbers.

Washington: Win +10.06
UCLA: Win +5.23
Oregon: Win +3.28
Oregon State: Win +10.13
USC: Win +20.38
Stanford: Win +17.58
BYU: Win +6.8

:acclaim:
:gobears:

---

ferCal,
Good work!

Again, just look at the Sagarin rankings. Arizona is undefeated and beat Oregon in Oregon, yet Oregon is ranked higher than Arizona? The rankings are still using last year to a large extent.

It took a bunch of fluke plays for Arizona to beat us in Arizona. Arizona is undefeated and beat Oregon in Oregon. We beat Washington State who beat Utah. Utah beat UCLA and Michigan. UCLA beat Virginia, Texas and ASU. ASU beat USC. USC beat Stanford and Oregon State. Stanford beat UW. Hawaii almost beat UW.

There is nothing in the above to say we are not the best team in the Pac-12 North, maybe in the Pac-12.

Let's beat the puppies this weekend!

Love this comment.

Davidson;842378683 said:

---

LOL, these things get revised each game.

Here is another "predictor"



http://www.sbnation.com/college-football/2014/10/8/6945501/pac-12-football-standings-predictions-projection-2014

That near 23.9% for Colorado to go 0-9 gotta hurt

---

Wow. That model isn't high on UW at all. I'd expect the curve for them to be much flatter than it is, considering they are largely unknown.

OTOH, Cal winning 5 conference games would be gigantic. That would give us an 8 win season if we also beat BYU. Unthinkable a month and a half ago. We'll see...

The only thing that counts is that we get to beat Arizona in the Pac 12 championship to avenge their clipping our perfect record.

Next we get to spend 60 years hating the playoff committee for selecting 4 SEC teams leaving Cal (and others) out.

BerkeleyChris;842378733 said:

---

You can't just count games that cal is favored as wins and games cal isn't losses. The right way to do it would be to estimate win probability for each matchup then compute the expected number of wins (in this case, add up win probabilities). If you do that there is a very good chance we at least get 6 wins

---

Yeah, there should be some kind of translation from points to % chance of winning such as 0 points is 50/50, 1 point is 45/55, 7 points is 20/80, or somerhing. Vegas essentially does that with the money line vs the spread, but sagarin might come up with something different.

CaliforniaGoldenBear;842378811 said:

---

The only thing that counts is that we get to beat Arizona in the Pac 12 championship to avenge their clipping our perfect record.

Next we get to spend 60 years hating the playoff committee for selecting 4 SEC teams leaving Cal (and others) out.

---

Pac-12 Championship game is at Levi's, right? Rose Bowl is #1 vs. #4....

BerlinerBaer;842378796 said:

---

Wow. That model isn't high on UW at all. I'd expect the curve for them to be much flatter than it is, considering they are largely unknown.

OTOH, Cal winning 5 conference games would be gigantic. That would give us an 8 win season if we also beat BYU. Unthinkable a month and a half ago. We'll see...

---

While i love your spirit, let's take these games one at a time. First, let's beat UW.

going4roses;842378732 said:

---

we wont be favored that is fine ... but byu ? no hill come on

---

The percentages change as the season progresses... We were way behind BYU and now we are almost favorite there. With Hill out BYU will do worse and that will flip to a should be win for us assuming we don't collapse.

going4roses;842378732 said:

---

we wont be favored that is fine ... but byu ? no hill come on

---

Sagarin does not take injuries into account for his predictions.

It is a pure mathematical model.

The good news from that model is that we have an 88% chance of finishing with 4+ conference wins, which would give us 6+ wins and make us bowl eligible. Add the BYU game at the end which is looking better and better for us, and our chances for a winning season are not too bad.

I hate computers in CFB..

FCBear;842379036 said:

---

I hate computers in CFB..

---

So you'd rather have a bunch of coaches like Mack Brown voting? Or perhaps Condi Rice deciding who makes the playoffs or not? At least the statistical models are impartial, the algorithms are fixed in advance, and they are not susceptible to appeals to emotion or allegiance.

ferCALgm2;842378746 said:

---

Fellows, this is based a lot off last year's season. I tweaked a few things, including our hail mary loss, winning in double OT right after the painful loss, and winning yet another close one right after. I added all that and a few more data to the logarithm and calculations and below is a more accurate computation with random disputable numbers.

Washington: Win +10.06
UCLA: Win +5.23
Oregon: Win +3.28
Oregon State: Win +10.13
USC: Win +20.38
Stanford: Win +17.58
BYU: Win +6.8

:acclaim:
:gobears:

---

plays game.
doesn't like outcome.
changes game.
likes outcome.
success!

i like your style.

Must not have read BYU's injury report.

Know Cal's luck it will be the Arizona game which keeps us out of the Rose Bowl. Again.

ferCALgm2;842378746 said:

---

a more accurate computation with random disputable numbers.

---

Well done, I see a bright future in comedy or politics for you.

dimitrig;842379151 said:

---

Know Cal's luck it will be the Arizona game which keeps us out of the Rose Bowl. Again.

---

One of the benefits of the expansion is that this will not happen any more, it will instead all come down to the CCG. If we can fend off the mutts, furd and the ducks we should be in. If that were in place we would have played USC for the Rose Bowl in the championship game both in 04 and in 06.

Cal88;842379266 said:

---

One of the benefits of the expansion is that this will not happen any more, it will instead all come down to the CCG. If we can fend off the mutts, furd and the ducks we should be in. If that were in place we would have played USC for the Rose Bowl in the championship game both in 04 and in 06.

---

Problem is the Rose Bowl this year is between whomever the committee ranks #1 and #4. We either would be left out for an extra SEC team or by our luck they'd rank us #3.

Cal88;842378709 said:

---

I think Sagarin might still be using data from last season, it might take another month for it to make a better prediction. The second predictor seems more reasonable at this point, it actually implies that Cal is the 6th or 7th best P12 team based on its predicted finishes of other conference foes.

---

But 3rd best in the north

BerkeleyChris;842379128 said:

---

So you'd rather have a bunch of coaches like Mack Brown voting? Or perhaps Condi Rice deciding who makes the playoffs or not? At least the statistical models are impartial, the algorithms are fixed in advance, and they are not susceptible to appeals to emotion or allegiance.

---

I'd rather just decide everything myself

Kind of amazed at the lack of math skills on this forum. First a programmer doesn't seem to understand the concept of an algorithm and then this post, seemingly clear on the concept, until calling it a logarithm. True, math instruction has changed in the 40 years since I graduated from Cal; but even then Algorithms were a step-by-step procedure for calculations, data processing, and automated reasoning. Unfortunately logarithms were somewhat forgotten when the log scales on slide rules were replaced by calculators. This algorithmic confusion I could forgive for a Chico State Grad...

Go Bears.

ferCALgm2;842378746 said:

---

Fellows, this is based a lot off last year's season. I tweaked a few things, including our hail mary loss, winning in double OT right after the painful loss, and winning yet another close one right after. I added all that and a few more data to the logarithm and calculations and below is a more accurate computation with random disputable numbers.

Washington: Win +10.06
UCLA: Win +5.23
Oregon: Win +3.28
Oregon State: Win +10.13
USC: Win +20.38
Stanford: Win +17.58
BYU: Win +6.8

:acclaim:
:gobears:

---