Thursday, October 16, 2008

Mailbag! Your Questions Answered

  • I am confused. It shows that your site was updated yesterday…but when I look at the computer rankings (based on the links provided) I get different numbers than what you show. For instance, I assume “JS” must equal the Jeff Sagarin rankings. You have Utah as 22 and BYU as 0. I followed the link on your site to Sagarin’s rankings and see Utah as 12 and BYU as 13. This is just one example.

Jeff Sagarin actually publishes two rating - Predictor and ELO_CHESS. The ranking of the ratings is a synthesis of these two - that's where you saw Utah at No. 12 and BYU at No. 13.

Now, look at the ELO_CHESS rankings - ELO_CHESS is the unbiased formula where margin of victory isn't not taken into consideration, i.e., a 59-point win is equal to a 1-point win on a blocked last-second PAT (you get my drift). The BCS uses Sagarin's ELO_CHESS ranking, and now you see that Utah is ranked 4th and BYU 35th. In the BCS system, first place gets 25 points, second 24 and so on. Therefore Utah gets 22 points and BYU gets 0 for being outside of top 25.

Obviously, BYU's blowout wins are not helping because of this scheme. What matters is that UCLA and Washington need to win as many games as possible to help out the Cougars.

  • Virginia Tech has a lead in the cumulative scores of the Coaches and Harris Polls. According to you, Michigan State has a computer average of .63, while Virginia Tech has a computer average of .47. However, after subtracting the top and bottom scores of both teams, Tech averages a ranking of 14.25, while MSU averages a ranking of 15.75. Can Michigan State really own such a sizeable advantage in computer scores?

The computer scores are calculated as such: For each team, a value is assigned to each computer ranking in inverse of 25 - i.e. No. 1 gets 25 points, 2 gets 24, and so on. So the points you see in each column is NOT rankings, but values. And for that purpose, the bigger the number, the better. Michigan State, after discarding the highest and lowest rankings, has a point total of 63, which translates into .630 in the computer average, and VT has 47, so .470. And yes, MSU's computer scores more than offset VT's better poll numbers.

  • I'm surprised you said "chances for two non-BCS teams are virtually nil". I think it looks like two non-BCS are virtually assured!

As it is right now, six BCS conference champions, a second team from both the SEC and Big 12, and one non-BCS team are assured of berths, so that leaves one.

That pick will end up with the second-place Big Ten team, in my estimation. Any combination of PSU-OSU-MSU will yield two teams in the top 14. An outside chance also is there for Cal to finish 10-2 and get into the top 14.

I'd love to see two non-BCS teams in the top 14 and force the BCS to take both, but that's just so unlikely.

  • IF, in the final BCS poll, there are TWO non-BCS in the top 14, (the first pool of at-large teams) and the PAC10 runner up and BigTen runner up are NOT in the top 14, then the BCS HAS TO TAKE the second non-BCS team. Right?

1. The top 14 rule is sacrosanct UNLESS there are not 14 eligible teams. Therefore, if none of the second place teams in the Pac-10, Big Ten, ACC and Big East are in the top 14, then by rule the BCS MUST take a non-BCS team that's eligible. The top 18 rule can only be invoked if there are not 10 eligible teams in the top 14 (regardless of conference affiliation).

2. All that said, I just don't think it's likely that only the SEC and Big 12 will have more than one team in the top 14, here's why:

The top 14 currently is populated by five Big 12 teams - four from the South Division - and that just won't hold. You'll see at least two of those teams, maybe three, drop from the top 14 after they all play each other.

Remember how far down Illinois was last year? Yet, at the end of the season, a 9-3 record was still good enough to be ranked No. 13 and a Rose Bowl invitation (a criminal one, but I digress).

  • Do you have to win your conference championship to go to the BCS title game? There was talk of Georgia maybe deserving to go last year. What are the chances of a rematch if Texas and Oklahoma both win out and of course Penn State and Alabama both loose?

You don't have to win the conference to get into the BCS title game. But a rematch of OU and UT in the BCS title game is unlikely because OU can't even advance to the Big 12 title game and will be punished by the voters for it. (As Georgia did a year ago)

  • What if OU, Texas & Texas Tech all have 1 conf loss at the end of the year (i.e OU beats Tech, Tech beats UT). Who makes the B12 title game?

In the three-way tie scenario, it'll probably come down to the placement in the BCS standings on Nov. 30. Here's the tiebreaker procedure:

1. The records of the three teams will be compared against each other.

2. The records of the three teams will be compared within their division.

3. The records of the three teams will be compared against the next highest placed teams in their division in order of finish (fourth, fifth and sixth).

4. The records of the three teams will be compared against all common conference opponents.

5. The highest ranked team in the first Bowl Championship Series Poll following the completion of Big 12 regular season conference play shall be the representative.

6. The team with the best overall winning percentage [excluding exempted games] shall be the representative.

7. The representative will be chosen by draw.

  • Who would be your BCS bowl matchups if the were to be picked for this Sunday?

Wow, this is waaaaaay too early. But what the heck, here's my crystal ball -

BCS - Alabama vs. Penn State

Rose - USC vs. Michigan State

Fiesta - Texas vs. BYU

Sugar - Florida vs. Oklahoma

Orange - Wake Forest vs. Pittsburgh

Finally, the BCS standings tracking has been updated to give better balance to unranked teams. And be sure to come back Saturday night as the Guru will unveil projections for the first official BCS standings. It will be published shortly after the Texas-Missouri 8 p.m. ET game. A live blogging comment forum will be available for those of you want to banter while you wait.

No comments: