2016 MLB Rankings – PLA Consensus Model Final Pre-Season Projections

By: Joe Botana

Proverbs 11:14  New American Standard Bible (NASB)

14 Where there is no guidance the people fall,
But in abundance of counselors there is [
a]victory.

Proverbs 15:22  New American Standard Bible (NASB)

22 Without consultation, plans are frustrated,
But with many counselors they [
a]succeed.

Last week, we introduced a new statistical model, based on consensus ratings from four popular existing models. Inspired by the Biblical wisdom quoted above, as well as by “poll of polls” and “analyst consensus” estimates used in political science and finance, our “PLA Consensus Model” seeks to predict the outcome of the upcoming baseball season based on the collective wisdom of others. Over the last week, we’ve used the output generated by the model to project the final results of all the NL and AL Divisions. To say we’ve gotten a lot of questions, disagreement, and incredulous comments would be an understatement. One even began with the phrase “If your wife was tied to the railroad tracks and the train was approaching, and you had to correctly predict the winner of “x” division, would you really pick “y” team to win it?” Our response has uniformly been the same – “It’s not our pick. It’s the model’s pick.” That in turn is the consensus pick based on the “collective counsel” of our four advisers, the four external models whose input we use.

In today’s final installment of this series, we’ll make several additional predictions. First, we’ll develop the first weekly “Power Rankings” for ATHT, and use the model’s rankings as our individual input to the ballot. Second, we will project the post-season picture based on the model’s consensus predictions. Finally, we’ll reaffirm our projections, and rank which ones we most expect to be right, and which we expect to be wrong, so we can check back in October.

PLA Consensus Model Pre-Season Power Rankings for ATHT Poll

Teams are ranked by a composite score based on consensus projected wins and consensus finishing position in division. The initial power rankings are as follow:

(1) Chicago Cubs       30         (2) Houston       29               (3) NY Mets          28

(4) San Francisco      27         (5) Pittsburgh    26               (6) Washington    25

(7) LA Dodgers          24         (8) Boston           23              (9)  Arizona          22

(10) Cleveland           21          (11) St. Louis      20              (12) Texas            19

(13) Toronto               18         (14)  NY Yankees   17           (15) Tampa Bay   16

(16)  Kansas City        15         (17) White Sox   14               (18)  Detroit         13

(19) Seattle                 12         (20)  Twins          11               (21)  Miami          10

(22)  LA Angels           9           (23) Brewers        8                (24) San Diego       7

(25) Oakland              6           (26)  Baltimore    5                (27)  Philadelphia  4

(28) Colorado             3          (29)  Atlanta         2                (30) Cincinnati       1

Post Season Projections

American League – Wild Card

NY Yankees (#5) at Toronto Blue Jays (#4) (Winners – Toronto)

ALDS

Toronto Blue Jays (#4) at Houston Astros (#1) (Winners – Houston)

Cleveland Indians (#3) at Boston Red Sox (#2) (Winners – Boston)

ALCS

Boston Red Sox (#2) at Houston Astros (#1)  (Winners – Astros)

National League – Wild Card

LA Dodgers (#5) at Pittsburgh Pirates (#4) (Winners – Pirates)

NLDS

Pittsburgh Pirates (#4) at Chicago Cubs (#1) (Winners – Cubs)

San Francisco Giants (#3) at NY Mets (#2) (Winners – NY Mets)

NLCS

NY Mets (#2) at Chicago Cubs (#1)  (Winners – Cubs)

WORLD SERIES

Houston Astros vs. Chicago Cubs (Winners – Cubs)

Other Predictions:

Based on the model’s consensus projections for wins and division placement, and also taking into account the degree of variance for the various teams and divisions, the division results are ranked in order of confidence:

  • Chicago Cubs win NLCD (Lock – 30 points)
  • Houston Astros win ALWD (Highly confident – 25 points)
  • New York Mets win NLED (Confident – 20 points)
  • San Francisco Giants win NLWD (Hopeful – 15 points)
  • Boston Red Sox win ALED (Tentative – 8 points)
  • Cleveland Indians win ALCD (Uncertain – 2 points)

Readers are welcome and encouraged to make a similar set of projections. The proposed scoring system allows for a possible 100 points. Save them, so that at the end of the season we can compare notes, and see whether there is “wisdom in a multitude of counselors” when it comes to picking baseball season finishing positions.

We can also rank the model’s predictions in terms of order of finish within the respective decision. In October, we will compute a score based on two factors. The first will be the number of teams that finish in other than the projected position. The second will be the “absolute value” of the differences in team finishing position within the division. (A team predicted to finish first but finishes fourth generates an absolute difference of 3. A team predicted to finish fifth but finishes third generates an absolute difference of 2.) The division with the lowest composite score will be the one with “highest order” while the one with the highest composite score will be deemed the one with “highest chaos.” Based on the model, these predictions appear below:

  • National League East — Highest Order/Lowest Composite Score
  • American League West
  • American League West
  • National League Central
  • National league West
  • American League Central — Highest Chaos/Highest Composite Score

Final Comments

This model was developed in response to the significant degree of variance observed in the predictions of several respected existing models. In theory, using their multiple inputs, based on different sets of assumptions and analytics, and developing a consensus prediction, should “smooth out” extremes and errors to produce a better overall prediction. We’ll come back in October and see whether the actual results support our theory.

Several readers have asked me why I referred to the model as the “PLA Consensus Model.” The answer is twofold. First, while I have seen and used models which apply this approach both in politics and economics/finance, the underlying wisdom of the approach comes from my favorite book, the Bible. Thus, the name “Proverbs Logic Analytics” was shortened to “PLA.” Coincidentally, they are also the initials of my grandfather, Pio Luis Anca, an innovative Cuban-Spanish-American entrepreneur who was also an avid baseball fan, and who taught me to love the game. I hope he is honored that this new model is in part named after him, and not too upset that its first consensus projections don’t have his beloved St. Louis Cardinals making the 2016 playoffs.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s