Recent:
1. Oregon #14
2. Stanford #22
3. Arizona #27
4. Colorado #38
5. Utah #63
6. Washington #65
7. USC #66
8. ASU #77
9. OSU #92
10. Cal #149
11. WSU #168
12. UCLA #200

Cal moving up. UW was a good win (they are all good, but this one more so) for us at this point and indicates it won't be our last.

Ucla at 200 and worst in the conference??? You have to go back to the 1940s to find UCLA so bad.

Chabbear said:

---

Ucla at 200 and worst in the conference??? You have to go back to the 1940s to find UCLA so bad.

---

And for those who believe Cal can't win without top recruits, well, UCLA can't win, even with seven, count'em, SEVEN RCSI Top 100 ranked recruits on the roster, and a formerly successful coach to boot. Want to bet Cal knocks them off this season?

#219 up to #149 in one week. I think that's a strong indication at how volatile Sagarin IS at this stage of the season. Also how fluid the rankings in the middle of the distribution are (let's say 50-300). Getting into the top 50 or out of the bottom 50 takes more consistency (think of a the ends and slope of a bell curve).

Sagarin has three different ratings, PREDICTOR, GOLDEN MEAN and RECENT. Then those three ratings are synthesized to come up with his overall RATING.

I would kind of caution against using the "RECENT" rating exclusively. Sometimes when the recent rating is significantly different than PREDICTOR or GOLDEN MEAN, it is an indication of true "trend", and might make the RECENT rating more indicative than the others. But RECENT could differ for a lot of other reasons as well. Injuries, hot or cold play by key players, confidence levels, luck, officiating, etc can cause a temporary difference in the RECENT rating, but it could regress towards the mean over time. By definition, RECENT involves heavy weighting towards a smaller sample size, and I think it is important to keep that in mind.

Here are the PREDICTOR ratings (with RECENT rating in parentheses) as of this morning:

PREDICTOR:
1. Oregon #11 (#13, was #14 before Sunday games)
2. Arizona #14 (#63, was #27)
3. Colorado #27 (#26, was #38)
4. Washington #57 (#65, was #65)
5. Stanfurd #61 (#22, was #22)
6. OSU #66 (#51, was #92)
7. ASU #69 (#76, was #77)
8. USC #75 (#66, was #66)
9. Utah #77 (#83, was #63)
10. UCLA #114 (#204, was #200)
11. WSU #158 (#168, was #168)
12. Cal #182 (#151, was #149)

On Sunday, OSU beat Arizona, and their RECENT rank climbed from 92 to 51, while Arizona dropped from 27 to 63. I think that illustrates how volatile the RECENT ranking can be.

Note that Stanfurd is much better in RECENT play than PREDICTOR, while UCLA and Arizona are much worse in RECENT play than PREDICTOR. It'll be interesting to see if those RECENT trends hold, or does Arizona and UCLA trend back up, and Stanfurd back down?

LOUMFSG2 said:

---

Sagarin has three different ratings, PREDICTOR, GOLDEN MEAN and RECENT. Then those three ratings are synthesized to come up with his overall RATING.

I would kind of caution against using the "RECENT" rating exclusively. Sometimes when the recent rating is significantly different than PREDICTOR or GOLDEN MEAN, it is an indication of true "trend", and might make the RECENT rating more indicative than the others. But RECENT could differ for a lot of other reasons as well. Injuries, hot or cold play by key players, confidence levels, luck, officiating, etc can cause a temporary difference in the RECENT rating, but it could regress towards the mean over time. By definition, RECENT involves heavy weighting towards a smaller sample size, and I think it is important to keep that in mind.

Here are the PREDICTOR ratings (with RECENT rating in parentheses) as of this morning:

PREDICTOR:
1. Oregon #11 (#13, was #14 before Sunday games)
2. Arizona #14 (#63, was #27)
3. Colorado #27 (#26, was #38)
4. Washington #57 (#65, was #65)
5. Stanfurd #61 (#22, was #22)
6. OSU #66 (#51, was #92)
7. ASU #69 (#76, was #77)
8. USC #75 (#66, was #66)
9. Utah #77 (#83, was #63)
10. UCLA #114 (#204, was #200)
11. WSU #158 (#168, was #168)
12. Cal #182 (#151, was #149)

On Sunday, OSU beat Arizona, and their RECENT rank climbed from 92 to 51, while Arizona dropped from 27 to 63. I think that illustrates how volatile the RECENT ranking can be.

Note that Stanfurd is much better in RECENT play than PREDICTOR, while UCLA and Arizona are much worse in RECENT play than PREDICTOR. It'll be interesting to see if those RECENT trends hold, or does Arizona and UCLA trend back up, and Stanfurd back down?

---

RECENT does not use a smaller sample size, it just gives more weight to the more recent data. Early in the season (entering league play) this is helpful in assessing the relative strengths of the PAC-12 teams and minimizing initial bias and flukey early results against much better or much worse teams, teams adjusting to new personnel, finding their best lineups, new coaches, etc. once teams have more games under their belt I lean more toward PREDICTOR. However, I do think PAC-12 results SHOULD be weighted more heavily in assessing relative strength among PAC-12 teams.

A key point is that PREDICTOR and GOLDEN MEAN are based on the idea that each team has an innate strength that does not change but is unknown and revealed over time with more data.

Comparing PREDICTOR with RECENT gives a good indication of trend. The fact that Cal's RECENT is improving dramatically is a strong indication that the TEAM may be improving. I would argue it generally gives a better indication of our chances of winning our next game. How good are the teams right now?

Of course, if you know a key player has been out and that has effected recent results, but that player is now back in the lineup, you would ignore RECENT.

Though as you point out, RECENT can be greatly influenced by (what we believe to be) upsets (OSU over Arizona for example).

Also, I think the volatility makes it more fun

calumnus said:

---

LOUMFSG2 said:

---

Sagarin has three different ratings, PREDICTOR, GOLDEN MEAN and RECENT. Then those three ratings are synthesized to come up with his overall RATING.

I would kind of caution against using the "RECENT" rating exclusively. Sometimes when the recent rating is significantly different than PREDICTOR or GOLDEN MEAN, it is an indication of true "trend", and might make the RECENT rating more indicative than the others. But RECENT could differ for a lot of other reasons as well. Injuries, hot or cold play by key players, confidence levels, luck, officiating, etc can cause a temporary difference in the RECENT rating, but it could regress towards the mean over time. By definition, RECENT involves heavy weighting towards a smaller sample size, and I think it is important to keep that in mind.

Here are the PREDICTOR ratings (with RECENT rating in parentheses) as of this morning:

PREDICTOR:
1. Oregon #11 (#13, was #14 before Sunday games)
2. Arizona #14 (#63, was #27)
3. Colorado #27 (#26, was #38)
4. Washington #57 (#65, was #65)
5. Stanfurd #61 (#22, was #22)
6. OSU #66 (#51, was #92)
7. ASU #69 (#76, was #77)
8. USC #75 (#66, was #66)
9. Utah #77 (#83, was #63)
10. UCLA #114 (#204, was #200)
11. WSU #158 (#168, was #168)
12. Cal #182 (#151, was #149)

On Sunday, OSU beat Arizona, and their RECENT rank climbed from 92 to 51, while Arizona dropped from 27 to 63. I think that illustrates how volatile the RECENT ranking can be.

Note that Stanfurd is much better in RECENT play than PREDICTOR, while UCLA and Arizona are much worse in RECENT play than PREDICTOR. It'll be interesting to see if those RECENT trends hold, or does Arizona and UCLA trend back up, and Stanfurd back down?

---

RECENT does not use a smaller sample size, it just gives more weight to the more recent data. Early in the season (entering league play) this is helpful in assessing the relative strengths of the PAC-12 teams and minimizing initial bias and flukey early results against much better or much worse teams, teams adjusting to new personnel, finding their best lineups, new coaches, etc. once teams have more games under their belt I lean more toward PREDICTOR. However, I do think PAC-12 results SHOULD be weighted more heavily in assessing relative strength among PAC-12 teams.

A key point is that PREDICTOR and GOLDEN MEAN are based on the idea that each team has an innate strength that does not change but is unknown and revealed over time with more data.

Comparing PREDICTOR with RECENT gives a good indication of trend. The fact that Cal's RECENT is improving dramatically is a strong indication that the TEAM may be improving. I would argue it generally gives a better indication of our chances of winning our next game. How good are the teams right now?

Of course, if you know a key player has been out and that has effected recent results, but that player is now back in the lineup, you would ignore RECENT.

Though as you point out, RECENT can be greatly influenced by (what we believe to be) upsets (OSU over Arizona for example).

Also, I think the volatility makes it more fun

---

I believe PREDICTOR is in fact the best predictor of future results. Thus, I generally focus on the PREDICTOR rankings. If you want to flavor in recent trends, then RATING combines them. RECENT by itself is too fluky and volatile to offer much insight, IMHO.

BGGB2 said:

---

calumnus said:

---

LOUMFSG2 said:

---

Sagarin has three different ratings, PREDICTOR, GOLDEN MEAN and RECENT. Then those three ratings are synthesized to come up with his overall RATING.

I would kind of caution against using the "RECENT" rating exclusively. Sometimes when the recent rating is significantly different than PREDICTOR or GOLDEN MEAN, it is an indication of true "trend", and might make the RECENT rating more indicative than the others. But RECENT could differ for a lot of other reasons as well. Injuries, hot or cold play by key players, confidence levels, luck, officiating, etc can cause a temporary difference in the RECENT rating, but it could regress towards the mean over time. By definition, RECENT involves heavy weighting towards a smaller sample size, and I think it is important to keep that in mind.

Here are the PREDICTOR ratings (with RECENT rating in parentheses) as of this morning:

PREDICTOR:
1. Oregon #11 (#13, was #14 before Sunday games)
2. Arizona #14 (#63, was #27)
3. Colorado #27 (#26, was #38)
4. Washington #57 (#65, was #65)
5. Stanfurd #61 (#22, was #22)
6. OSU #66 (#51, was #92)
7. ASU #69 (#76, was #77)
8. USC #75 (#66, was #66)
9. Utah #77 (#83, was #63)
10. UCLA #114 (#204, was #200)
11. WSU #158 (#168, was #168)
12. Cal #182 (#151, was #149)

On Sunday, OSU beat Arizona, and their RECENT rank climbed from 92 to 51, while Arizona dropped from 27 to 63. I think that illustrates how volatile the RECENT ranking can be.

Note that Stanfurd is much better in RECENT play than PREDICTOR, while UCLA and Arizona are much worse in RECENT play than PREDICTOR. It'll be interesting to see if those RECENT trends hold, or does Arizona and UCLA trend back up, and Stanfurd back down?

---

RECENT does not use a smaller sample size, it just gives more weight to the more recent data. Early in the season (entering league play) this is helpful in assessing the relative strengths of the PAC-12 teams and minimizing initial bias and flukey early results against much better or much worse teams, teams adjusting to new personnel, finding their best lineups, new coaches, etc. once teams have more games under their belt I lean more toward PREDICTOR. However, I do think PAC-12 results SHOULD be weighted more heavily in assessing relative strength among PAC-12 teams.

A key point is that PREDICTOR and GOLDEN MEAN are based on the idea that each team has an innate strength that does not change but is unknown and revealed over time with more data.

Comparing PREDICTOR with RECENT gives a good indication of trend. The fact that Cal's RECENT is improving dramatically is a strong indication that the TEAM may be improving. I would argue it generally gives a better indication of our chances of winning our next game. How good are the teams right now?

Of course, if you know a key player has been out and that has effected recent results, but that player is now back in the lineup, you would ignore RECENT.

Though as you point out, RECENT can be greatly influenced by (what we believe to be) upsets (OSU over Arizona for example).

Also, I think the volatility makes it more fun

---

I believe PREDICTOR is in fact the best predictor of future results. Thus, I generally focus on the PREDICTOR rankings. If you want to flavor in recent trends, then RATING combines them. RECENT by itself is too fluky and volatile to offer much insight, IMHO.

---

I believe Helltopay1 is in fact the best predictor of future results. The man is uncanny.

Yeah, he knows a lot, he'll be the first to tell you that, and what he doesn't know he makes up.

The funny thing is that if we continue to get better and upset the higher rated P12 teams, we're going to put a big dent into their Sagarin/RPIs...