World Series Projections: Texas over St. Louis in Five
Note: the original chart reversed the distribution of game outcomes for St. Louis. The error has been rectified. I thank our vigilant readers for catching this and apologize for any confusion.
Defying all odds, the St. Louis Cardinals find themselves in the World Series, ready to take on the heavy favorites: the Texas Rangers.
Point of fact: I did not expect the Cardinals to be here, not in August, not in September, not last week. My simulator didn't think any more highly of them. Yet, here we are. Our projections have the Rangers as 2:5 favorites, with the most likely outcome being a Texas win in five games.
So says our Log5 postseason simulator based on numbers from Baseball Prospectus. Raw numbers and figures of note after the jump.
Figures of Note:
- As of right now, all four St. Louis winning outcomes are less likely than all four Texas winning outcomes.
- That means we expect a Texas sweep before we expect a Cardinals win in seven games.
- The Cardinals' victory over the Brewers boosted Texas' chances by 1.5%.
- For the first time since 2005, both World Series representatives hail from the Central Time Zone.
| Probability of Postseason Series Victory | 10/17/11 | ||||||
| Inputs | Outputs | ||||||
| Team | Seed | Talent | LCS% | WS% | WS HFA | ||
| TEX | 2 | 0.640 | 100.0% | 72.2% | -2.0% | ||
| STL | WC | 0.530 | 100.0% | 27.8% | 2.0% | ||
| AL | 72.2% | -2.0% | |||||
| NL | 27.8% | 2.0% | |||||
Methodology: Our simulator predicts the outcome of every possible single-game match-up in every potential five- and seven-game series match-up between the eight teams currently in line for a playoff berth. Baseball Prospectus' Adjusted Hit List serves as proxy for true talent. We use the Log5 method to predict single-game match-ups, adjusting each number to reflect a 0.540 home field advantage (based on Matt Swartz's findings). The simulator does not adjust for roster discrepancies, pitching rotations, or any stadium-specific home field advantages that may or may not exist.
21 comments
|
0 recs |
Do you like this story?
Comments
70-30? That seems like an unrealistically high level of certainty.
Baseball in short series just doesn’t work like that. I doubt I would give the rangers more than 60-40, just because of the randomness of baseball.
Agree 100%
It’s what my numbers tell me, but my intuition says 60:40. Xei tells me that Vegas says 58:42.
Jesse-Douglas Mathewson, Ph.D. Candidate in Government and Politics at UMD-College Park.
Columnist for Beyond the Box Score specializing in projections, PitchF/X and infographics.
Blogger and Editor, Rational Pastime Blog. Twitter: @RationalPastime.
I could see a Ranger sweep
But wouldnt be surprised if it went 6. Either way I dont think the Cards have any gas left in that pen to deal with the Ranger offense. And as shaky as their rotation looked against the Brewers, if they are equally shaky, it wont be pretty.
Could be going to game 2, so no matter how it goes that should be exciting.
There's an error in the chart
Fixing it right now
Jesse-Douglas Mathewson, Ph.D. Candidate in Government and Politics at UMD-College Park.
Columnist for Beyond the Box Score specializing in projections, PitchF/X and infographics.
Blogger and Editor, Rational Pastime Blog. Twitter: @RationalPastime.
It's fixed.
Jesse-Douglas Mathewson, Ph.D. Candidate in Government and Politics at UMD-College Park.
Columnist for Beyond the Box Score specializing in projections, PitchF/X and infographics.
Blogger and Editor, Rational Pastime Blog. Twitter: @RationalPastime.
I posted this on lonestarball
The probability of the Rangers winning Game 7 given that there is a game 7 is 15.4/(15.4+2.2) = 87.5%. That seems awfully high for a single game.
That's not the right number.
These aren’t the numbers indicating what the likelihood of a win in any individual game is, it’s simply the distribution of expected outcomes of the eight current possibilities.
The Log5 probability of a team with Texas’ numbers winning a game on the road against a team with St. Louis’ numbers is 57.3%.
Jesse-Douglas Mathewson, Ph.D. Candidate in Government and Politics at UMD-College Park.
Columnist for Beyond the Box Score specializing in projections, PitchF/X and infographics.
Blogger and Editor, Rational Pastime Blog. Twitter: @RationalPastime.
Right...
That’s why I wrote “given that there is a game 7”.
It estimates that the Rangers win in 7 games 15.4% of the time and the Cards win in 7 games in 2.2% of the time, meaning it estimates the series goes 7 games 17.6% of the time.
Of that 17.6%, it assumes the Rangers win game seven 15.4/17.6 of the time?
Okay, nevermind, I see what you're saying and you're right. I have to fix something here.
Jesse-Douglas Mathewson, Ph.D. Candidate in Government and Politics at UMD-College Park.
Columnist for Beyond the Box Score specializing in projections, PitchF/X and infographics.
Blogger and Editor, Rational Pastime Blog. Twitter: @RationalPastime.
Cool thanks :)
And thanks for doing all this work over the entire playoffs. It’s been extremely interesting :)
by cjcheng on Oct 17, 2011 2:22 PM EDT up reply actions 1 recs
Fixed.
Jesse-Douglas Mathewson, Ph.D. Candidate in Government and Politics at UMD-College Park.
Columnist for Beyond the Box Score specializing in projections, PitchF/X and infographics.
Blogger and Editor, Rational Pastime Blog. Twitter: @RationalPastime.
Hey, no problem, I enjoy it.
I only wish I had more time so that I could better fine-tune my projections.
Jesse-Douglas Mathewson, Ph.D. Candidate in Government and Politics at UMD-College Park.
Columnist for Beyond the Box Score specializing in projections, PitchF/X and infographics.
Blogger and Editor, Rational Pastime Blog. Twitter: @RationalPastime.
The first chart would be neat if the STL bars were flipped.
Then you’d have a continuous distribution, with the extreme outcomes on the outside.
It'd be nice this time
But there are times when that labeling sequence wouldn’t have the same effect, plus I feel it’d be less intuitive.
Jesse-Douglas Mathewson, Ph.D. Candidate in Government and Politics at UMD-College Park.
Columnist for Beyond the Box Score specializing in projections, PitchF/X and infographics.
Blogger and Editor, Rational Pastime Blog. Twitter: @RationalPastime.
I'll see if I can come up with something that's consistent on both axes.
Jesse-Douglas Mathewson, Ph.D. Candidate in Government and Politics at UMD-College Park.
Columnist for Beyond the Box Score specializing in projections, PitchF/X and infographics.
Blogger and Editor, Rational Pastime Blog. Twitter: @RationalPastime.
Theoretically, is there any reason why the distribution wouldn't be uni-modal?
I suppose if home field advantage is more significant (like in basketball) or the starting pitcher matchups are drastically different between games, which is certainly possible.
It can be slightly bimodal thanks to home field advantage...
…if the two teams are evenly matched or if the underdog takes an early lead.
Jesse-Douglas Mathewson, Ph.D. Candidate in Government and Politics at UMD-College Park.
Columnist for Beyond the Box Score specializing in projections, PitchF/X and infographics.
Blogger and Editor, Rational Pastime Blog. Twitter: @RationalPastime.

by 



















