Comments on: WS Prediction: AL over NL in 4 maybe 5 games. This and that about baseball stats. Tue, 16 Jul 2013 17:01:55 +0000 hourly 1 By: David in Toledo Sat, 27 Sep 2008 15:22:24 +0000 It's terrific to see the detailed number crunching -- thanks!

However, it also comes down to who can actually play in October. Will the Red Sox have Lowell, Drew, Ortiz, and a full complement of pitchers?

By: whiz Fri, 26 Sep 2008 21:19:19 +0000 Good stuff Sean. I have been posting similar rankings over at for the last month or two (the older articles that describe the system in more detail are available in the archives there).

My algorithm is a little different (it actually uses individual game scores and not run totals), although it also skews the rankings heavily towards AL teams (the Cubs are the best NL team and do no better than 6th or 7th).

Actually I have four different ranking systems: one that uses runs scored and runs allowed (RS/RA), only wins and losses (W/L), Runs Created (RC), and a hybrid system that interpolates between W/L and RS/RA. They all incorporate a home field advantage and account for schedule strength.

The RS/RA and RC rankings calculate both an offensive and defensive rating and then use the Pythagenpat formula to find a hypothetical win percentage against an average team at a neutral site. That's better than a differential since a one-run difference is more important (and leads to a better win pct.) for low-scoring games than for high-scoring games.

The final regular season rankings should be posted Monday, if there are no tie-breaker games. I will also be using the rankings to make predictions for the division series and beyond; that will hopefully be posted on Tuesday.

As a teaser, Anaheim will probably host Boston if things stay as they are now; using my current rankings, Boston has a 62% chance of winning the series, 42% of winning in 4 games or less. For more details, see my article on next Tuesday.