First, the Pac-10 remains at the top of the national heap, leading the Big Ten by about 0.9 point, and the SEC by 1.8 points. For the uninitiated, these numbers can be read to mean that today's average Pac-10 FB team - which doesn't exist - would be favored over an average Big-Ten team by 9-tenths a point if a (rose?) bowl game was played on a neutral field.

These numbers are small because they address "average" teams. As the abilities of large numbers of players are folded together, the differences between groups fall.

Second, as expected, USC remains the Pac-10 leader, though surprisingly their "rank" fell by almost two points - which means the computer is telling us that USC's win over Arizona was not quite as large as expected. In fact, Arizona gained almost 4 points in their ranking despite their 20-3 loss to USC. It seems that Arizona did better against USC than prior data predicted.

For those who wonder about the "accuracy" of these numbers, perhaps a few brief comments are in order.

First, the numbers are "exactly right", in the technical sense of "precise". They have been computed to find that set of team rankings which "best" explain the results of games so far - where "best" does not mean accurate, but instead means only that any other set of numbers would be worse. So however bad these numbers are, all others are worse. Isn't that an affirmative statement? Well it is, actually, to statisticians and mathematitions

But are they "accurate" in the sense that they can be used to predict game outcomes? Of course not. Even though the computation is done exactly, the data fit is not perfect, it is only the best that can be done given the mix of wins and upsets that occurred over the weekend.

The other serious limitation is that many games between important contenders for conference championships have not yet been played - so the computer does not have that data to fold into its mix.

Perhaps the most positive comment that can be said about these numbers is that they are completely unbiased. Every team is treated the same - valued simply by points scored in wins and losses. Coach or writer biases are excluded, as are inaccuracies that occur because most writers and coaches don't have the time to watch many games - and so rely on whatever the current hype is about teams they don't follow.

With that all in mind - here are a few more interesting bits of data gleaned from this week's results:

1. California gained 2.5 points - and Arizona State lost 7.2 points - as a result of their match up at Memorial Stadium. That should be read to say that Arizona State was more over-rated than Cal was under-rated, so the adjustments fell that way. The 7.2 point loss by ASU was the second biggest drop experienced by a Pac-10 team (but see UCAL). Does that make it a comedown?

2. The Leland Stanford Jr. University took a shellacking, and not only on the field. They dropped 3.2 points to a 56.8 rating - while Washington State (who beat Stanford 36-10) gained 5.5 points to 72.03. The Cardinal may have set a record for lowest score ever by a Pac-10 team - we will enquire with Sagarin to try to confirm this.

3. Were Cal to play Stanford today at Memorial Stadium, Cal would be favored by 35.5 points.

4. Using Sagarin's "best predictor" as the measure (as it the case in all the numbers above), Cal is ranked 6th in the nation, with USC #1, Oregon State 26th, and Oregon 10th - Cal's next two opponents. With home field advantages computed in, Cal is favored over OSU by 5.5 points, and over Oregon by 9. (Oregon is ranked higher than OSU, but Cal adds a home field advantage against Oregon, loses it against Oregon State.)

5. UCLA suffered a dreadful drop in it's ranking, from 82.5 to 70.1, a 12.4 point drop, as a result of their unexpected loss to Washington ion Seattle, 29-19. Washington gained only a half-point from the win, so again the computer speaks: "Data says Washington was ranked correctly, UCLA WAY over-ranked."

6. OSU, Cal's next opponent, is ranked closely to Arizona State.

7. The Pac-10 now looks like USC, Cal, and Oregon, the only teams with 80 points or more. ASU and OSU are in the high 70s, WSU trailing them by a few points, the rest (including UCLA!) struggling for cellar rights.

(Complete rankings for Cal's season are available at The Bear Insider's "BearPlay" game where Cal fans compete by picking winning margins for Cal football games. The game can be seen here.)

8. Get past OSU and Oregon, and Cal is favored by double-digits the rest of the way - except, of course, against the Trojans.

9. Tennessee is now ranked 83.4, 5.4 points lower than Cal. Were they to have a game this week at Memorial Stadium, Cal would be (per the computer!) favored by 9 points. Huh? The only bit of evidence available to us that lends any credence to that are the comments made by several Cal team members who today say that if they could play Tennessee 10 times again, they'd win all ten games. We have taken that with a very large grain of salt, of course, attributing it to youthful enthusiasm and budding confidence. The explanation they give is that they've seen the films, and have seen the extent to which minor breakdowns in execution swung the tide of the game. Maybe - just maybe - they know something we don't.

©Copyright 2006, BearInsider.com and Scout.com. All rights reserved.

If you haven't done so already, subscribe to The Bear Insider so you can participate in this active online Cal community and get access to the members-only content from the nation-wide Scout.com network.

Bear Insider staff writers visit the Insider discussion board regularly, and are available to discuss questions you may have about this article and Cal Athletics.