Hey, Bill

Win Shares questions...  In the original book, you had Wally Berger drop from 33 to 21 WS between 1934 and 1935 despite remarkably similar numbers (150 games each season, three points higher on-base and two points higher slugging in 1935).  The obvious difference here is that the team was far worse in 1935, one of the worst teams of all-time.

So, two questions.  1) Does Win Shares "punish" players on historically bad teams?  Berger wasn't doing anything differently in 1935 than he had before.  2) Will Berger's Win Shares/Loss Shares under the new system still show the seeming plummet in value despite the steady numbers?

--Mac Thomason

1)  It was not my intention to “punish” (as you say) players on historically bad teams.   However, when teams are really, really bad, it puts a lot of pressure on the system, and sometimes we get results which are a little atypical simply because the team is so bad that the system has difficulty locking onto the normal performance parameters. 

2)  Your question says that “the obvious difference here is that the team was far worse in 1935”, but that’s not the only difference; there are actually several other things going on here.  

3)  Berger, whose fielding percentage was normally OK, led National League outfielders in errors in 1935, with 17.  

4)  Berger created an estimated 116 runs in 1934, 112 in 1935, a 3% drop.

5)  There were 108 more runs scored in the National League in 1935 than in 1934, an increase of about 2%, although most of that actually is explained by more games played. . .it’s really just a ½ of one percent increase in runs per game. 

These are small changes, but we’re talking about a 36% drop in Win Shares.   If you put together several 2, 3% changes, it adds up. 

6)  The bigger thing is park effects.  In 1934 the Braves scored and allowed 560 runs in their home park, as opposed to 837 on the road, a very low park effect (.669).  There were 63 home runs in their home park, 98 on the road.

            In 1935, adding Babe Ruth, I am assuming that they pulled in their right field fence; I don’t know for sure.   I am assuming this because 1) it is something one would naturally do when adding an aging Babe Ruth, 2) teams in that era moved their fences in and out at the drop of a hat, and 3) the park effect changed dramatically in 1935.   In 1935 the Braves scored and allowed 690 runs at home, 737 on the road (home runs 75 at home, 81 on the road), for a raw park effect of .936. 

            Berger, a right-handed hitter, was not apparently much affected by the change, but, of course, that’s not what we’re adjusting for.   What is relevant is not the effect on this particular hitter, but the effect on the value of his runs.   The higher park effect makes the context of his runs substantially higher, and thus reduces his value, although his numbers are nearly the same.   Or, stated another way, the introduction of more runs into the games he was playing made the runs that he created somewhat less significant.

7)  These are Wally Berger’s career numbers:

And these are the Win Shares and Loss Shares I assign to him:

The 1934 record breaks down as 22-2 on offense, 4-6 in the field, whereas the 1935 breaks down as 18-6 on offense, 2-8 in the field. 

8)  I continue to believe that Berger would have been as good a Hall of Fame selection as Hack Wilson or Earl Averill, two contemporary center fielders who have very similar career numbers.   This chart compares the Win Shares and Loss Shares for the three players, matched at the same age: