11 January 2013

The Year After Having Good Luck: Orioles 2013

The Orioles won 93 games, 11 more than the value of their Pythagorean Win Expectation, which was originally introduced by Bill James in his 1980 Baseball Abstract.  This formula calculates winning percentage based on the number of runs a team scores and the number of runs a team allows.
Winning Percentage = (Runs Scored^2) / (Runs Scored^2 + Runs Against^2)

In the past ten years, six teams have won ten or more games than expected based on runs scored and runs against:

Year Wins Δ PWE
Baltimore 2012 93 11
Seattle 2009 85 10
LAAA 2008 100 12
Arizona 2007 90 11
Arizona 2005 77 11
NYY 2004 101 12
To put this in context, this event happened in 2% of seasons in the past ten years.  Additionally, only one team ever duplicated this event.  The Arizona Diamondbacks managed to do this twice with a year in between.  Roughly speaking, it appears that this level of production is typically not a repeatable skill.  To explore this more fully, I looked at teams that finished seven games above their Pythagorean Win Expectation in any given year between 2002 and 2011.  The teams in this grouping include:

Year Wins Δ PWE
Houston 2010 76 8
Seattle 2009 85 10
San Diego 2009 75 8
NYY 2009 103 8
LAAA 2008 100 12
Houston 2008 86 9
Arizona 2007 90 11
Seattle 2007 88 9
St. Louis 2007 78 7
Oakland 2006 93 8
Arizona 2005 77 11
Chicago-AL 2005 99 8
NYY 2004 101 12
Cincinatti 2004 76 9
SF 2003 100 7
Minnesota 2002 94 8
Oakland 2002 103 7
I then compared that performance to what the team accomplished the season before and after the season in question. 

N-1 N N+1
Wins 80.3 89.6 81.7
StDev 14.3 10.4 13.3
The N-1 and N+1 years were measured as being significantly less than the year where the team out performed their expectation.  Something similar can be seen when looking how teams perform against their expectations.


N-1 0 4 *
N 8.9 1.7
N+1 0.4 3.8 *
What the above table shows is that the population of teams on average outperformed their Pythagorean Winning Expectation by 8.9 +/- 1.7 wins.  That performance was significantly more than how the team performed according to this metric the year prior or afterward.  In other words, outperforming the Pythagorean Win Expectation is quite unlikely to be an actual skill.  If it was a skill then one would expect a team would continually outperform the expectation.

What does this all mean?  That whatever the reasons are behind the Orioles winning 11 more games than expected that it is highly unlikely that they can repeat that effort in 2013.  In a relatively generic way, it underscores the team's need to improve the roster this off season.


Andrew said...

One thing I notice in your analysis is that from the year-before outperformance to the year-after outperformance, teams improved by a not-insignificant 1.7 wins.

It seems there's some relationship between a team outperforming its Pythag W-L total and a team that's generally in the process of improving. Or, that there's coincidentally a number of recent teams who outperformed their Pythag W-L and also happened to be in the process of improving.

I think most smart O's fans know we're much closer to a true 80 win team than a 93 win team. But if you can get to the level of a true 85 win team, you're within the range where you can accidentally wind up with the second wild card, depending on what happens around you.

Jon Shepherd said...

I'm not sure where you got that from that article. That was not in my findings.

It was a while ago that I ran this, but I think I found an insignificant difference between the N-1 and N+1 years.

Can you provide that information?

Renaissance Man said...

How much of the PWE for 2012 was skewed negatively towards the part of the season where Arrieta, Matusz, and Hunter were making regular starts, and where, pre-Manny, their defense was atrocious? Even without making any upgrades from outside the organization, I think the O's will get some improvement just by having a better selection of starting pitchers to choose from, and by hopefully not ever having to see the likes of Wilson Betemit put on a glove and trot out to third base again...

Jon Shepherd said...

How much of their PWE is skewed due to McLouth performing well, Chris Davis having a breakout season, Chen's exceptional ERA in comparison to his peripherals, Strop's season etc.?

I think we need to be careful when looking to the future. Being aware of both the good and bad luck. Do you think the team really had an exorbitant amount of bad luck or good luck?