Strength of Schedule and Announcers

I promise to get to those burning questions of the CWS like who to start on Friday and who to root for on Wednesday, but first let's have one last note on Cal State Fullerton now that they have been eliminated by the plucky Cavaliers.  In one of the comments thread, it was pointed out that the ISR's might overrate west coast teams.

Now, I'm loathe to criticize the amazing work of Boyd World.  His site really is one of the very best resources for a college baseball fan.  And let's be honest, there's no way I could construct an even halfway decent computer rating system.  This, of course, is a long-winded preamble to me criticizing the amazing work of Boyd World. 

When the field was announced, Boyd crunched the numbers and found that CSF had a 46.4% chance of making the final.  That's, for all intents and purposes, even money odds.  That is also, patently absurd.  Nothing against CSF, but there's no team that should have been even money to make the championship series, particularly a team in the same bracket as such heavyweights as UC-Irvine, Florida St, and LSU.  Favorites?  Sure.  But overwhelming favorites?  By contrast, LSU had a 46.2% chance of just making it Omaha.  While LSU had a 9.2% chance of making the final series and a 3.7% chance of winning it, CSF had a 32.6% of winning the title. 

Did anyone honestly believe that CSF had a ten times better chance of winning the title than LSU when the field was announced?  Hell, I don't even think Boyd believed that.  Check out the pre-CWS odds:

Final  Champ  Team

30.8   14.0   Texas

2.4    0.3   Southern Mississippi

27.6    9.8   North Carolina

39.1   21.4   Arizona State

57.9   40.1   Cal State Fullerton

14.6    4.6   Arkansas

17.2    6.5   Louisiana State

10.3    3.4   Virginia

For those of you scoring at home, that means he gave nearly the same odds for CSF to win the title as Texas, Arizona St, and LSU COMBINED.  There has to be a point you look at the numbers your system pumps out and ask "does this make sense?" 

Believe it or not, this isn't to tear apart Boyd.  I meant it when I said that I deeply respect his work.  There is a point here.  Because I believe I know the cause of the problem, even if I don't have the solution (going back to the idea I can't build a rating system).  The problem is a faulty strength of schedule rating. 

Strength of schedule is usually calculated by some variation of a factor of opponents' winning percentage added or multiplied to a factor of opponents' opponents' winning percentage.  While this is the industry standard, I believe that it is completely incorrect.  It does give a good baseline, but there are two fundamental flaws. 

ONE. Assuming the same opponents' opponents' winning percentage, this method of strength of schedule overrates teams that play lots of mediocre teams.  It is better to play two 5-5 teams and a 6-4 team (to use football), than to play a 10-0 team, a 5-5 team, and a 0-10 team.  But it is harder to go undefeated against the second schedule despite it being rated as the easier schedule.  This is a basic flaw of strength of schedule that it does not have a component to rate top tier games.  But really, it is a function of the more basic flaw...

TWO.  Strength of schedule ratings rate a game against a lesser opponent as easier than an open date.  Think about that, playing anyone, no matter how poor of a squad, is ALWAYS tougher than not playing at all.  Any system that punishes a team for playing a team in comparison to not playing at all is fundamentally flawed.  A truly good rating system, like Boyd's, can be so well constructed as to minimize this flaw, but this flaw cannot be completely ignored. 

Playing is always harder than not playing.  I don't have the solution to strength of schedule, but I simply do not believe strength of schedule ratings.  Which makes me horribly skeptical of computer rankings.  Hopefully, someone better at math than I am can figure out the workaround to this basic problem.  ,  

I don't know much about computer rankings, but I do know that there is no earthly way CSF had a ten times better chance in Omaha than LSU. 

....

I like to rip on the MSM, particularly ESPN, but the College World Series is one of those things ESPN does right. 

Orel Hersheiser and Mike Patrick mangled every name on LSU's roster, but other than that, they had great chemistry and made an entertaining broadcast despite having to kill time during a blowout.  They were funny, engaging, and I actually learned a few things.  Having a great pitcher like Orel break down pitcher mechanics is a real treat. 

Their discussion on hit batsmen was funny and also pretty interesting.  And Orel got off one of the best lines of the night when he admitted that there was no way Patrick was ever going to convince him about anything that helped a hitter. 

I'm pretty critical, so I think it's only fair that I praise when ESPN does something well.  They have done an outstanding job in Omaha so far, especially in how they let the games speak for themselves.  Good job, guys.