Wednesday, March 11, 2009

2009 Baseball Predictions Primer

In the past (2008 and 2007) I have made predictions about the outcome of the baseball season. I based these on the previous year's record and then added in my own personal adjustment. It is a fairly simple formula.

Previous Years Wins +/- Off-season Moves.

This year I am taking a different approach. I have run some regression statistics and will be using those. I figured that I would explain the process before I actually posted the predictions. I have run a regression analysis on the premise that Win Total is a function of OPS and WHIP. To make it simpler, Winning is a combination of Hitting and Pitching. I chose OPS and WHIP as indicators of each respectively. Running a regression analysis on the past 5 seasons, I came up with this formula:

Win Total = 55.19 + (192.8 * OPS) - (86.2 * WHIP)

Now, you may ask, "How accurate is that formula?" A good question. According to the regression model, the adjusted R squared is .7835. In other words, The above formula accounts for 78.35% of the reason wins are the total that they are. For baseball, this is a fairly strong relationship. There is obviously no way to account for random occurrences fully, so a good 21.65% is still left to chance and situational factors.

So with that formula in mind. I calculated how many wins each team would have won if that formula were perfect. I then took the difference between reality and prediction and multiplied it by 78.35%. So with that, I have come up with the Adjusted Win Total for each team last year.

I will use this as my base prior to adjustment this year rather than actual results. Why do I want to do this? I want to know how good teams were, not how lucky they were. Therefore my formula this year is:

Previous Years Adjusted Wins +/- Opinion of Off-season Moves.

So what exactly are the adjusted win totals that I will be working with?

AL East
  1. 96 - 66 Boston Red Sox

  2. 92 - 70 Tampa Bay Rays

  3. 89 - 73 Toronto Blue Jays

  4. 87 - 75 New York Yankees

  5. 68 - 94 Baltimore Orioles

AL Central
  1. 91 - 71 Chicago White Sox

  2. 84 - 78 Cleveland Indians

  3. 84 - 78 Minnesota Twins

  4. 76 - 86 Detroit Tigers

  5. 75 - 87 Kansas City Royals

AL West
  1. 88 - 74 Los Angeles Angels

  2. 77 - 85 Texas Rangers

  3. 72 - 88 Oakland Athletics

  4. 61 - 101 Seattle Mariners

NL East
  1. 88 - 74 Philadelphia Phillies

  2. 85 - 77 New York Mets

  3. 82 - 80 Florida Marlins

  4. 77 - 85 Atlanta Braves

  5. 63 - 99 Washington Nationals

NL Central
  1. 98 - 64 Chicago Cubs

  2. 87 - 75 Milwaukee Brewers

  3. 87 - 75 St. Louis Cardinals

  4. 81 - 81 Houston Astros

  5. 71 - 91 Cincinnati Reds

  6. 61 - 101 Pittsburgh Pirates

NL West
  1. 86 - 76 Arizona Diamondbacks

  2. 85 - 77 Los Angeles Dodgers

  3. 74 - 88 Colorado Rockies

  4. 70 - 92 San Diego Padres

  5. 68 - 94 San Francisco Giants


Which team was adjusted the most?
Los Angeles Angels (-12 wins)

Feel free to post comments or suggestions. Is there a stat that you think would be better to use than either OPS or WHIP?

2 comments:

III said...

Neat enough idea. But some anomalies that stand out to me...

* Toronto with 89 wins? Based on what? Burnett's gone. McGowan's starting the year injured. So is Vernon Wells. What's to like?

* 60 wins for Seattle?! HaHa, awesome. I'm tempted to make a wager with you on them. I say they're over 70 wins.

* I see the Reds as better than just a 70-win team this season.

Otherwise, very nicely done.

Jordan said...

Ok, let me clarify what this is. Instead of using the standings from last year as a base, I will be using the above standings as a base of performance from last year. In a couple weeks, I will post a prediction for this years standings that have my personal opinion adjustments added in. These will take into account off-season moves and inklings.