Friday, January 15, 2010

Bert Blyleven, Jim Palmer and Defensive Efficiency Rating

Bert Blyleven had 4,970 IP and had 344 RSAA. That comes from the Lee Sinins Complete Baseball Encyclopedia. It is "Runs saved against average. It's the amount of runs that a pitcher saved vs. what an average pitcher would have allowed." It is park adjusted. So he saved about .62 runs per game. Jim Palmer pitched 3,948 innings with 314 RSAA. So he saved about .72 runs per 9 IP.

But, fielders help pitchers prevent runs and many fans know that the Orioles had great fielders over the years like Brooks Robinson, Paul Blair, Mark Belanger and Bobby Grich. This is where a Bill James stat called Defensive Efficiency Rating (DER) comes in. It tells us what % of balls in play (BIP) were turned into outs by a team's fielders. The higher the rating, the better the defense.

So I looked at how good the DERs were for the Orioles during the years Palmer pitched and also for the teams that Blyleven pitched for. I found that the Orioles had better than average fielding and the teams Blyleven pitched for had below average fielding. So Palmer's fielders added to his RSAA and Blyleven's fielders reduced his RSAA. I tried to estimate how many runs this added up to and then recalculated each pitcher's RSAA per 9 IP.

I calculated DER as 1 - BABIP (batting average on balls in play). I calculated BABIP as

(H - HR)/(BFP - SO - BB - HBP - HR)

Palmer's team DER over his career (a weighted average as a % of BFP) was .738. The league average was .726 for a difference of about -.012. That is, the DER was .012 higher for the Orioles than for the league. Palmer had 12248 BIP. Times about .012 is 145. So the Oriole fielders made about 145 plays that average teams would not. If those hits had an average linear weights run value of .55, that is about 80 runs his fielders saved him.* So his career RSAA falls to 234. Per 9 IP that becomes .53.

Blyleven's teams had a DER of .718 while the league average was .722 for a difference of about .004. He had 14883. That times .004 is 57. That means that his fielders allowed an extra 57 hits. That times .55 means about 32 runs he gave up he should not have. Then his career RSAA should rise from 344 to 376. Per 9 IP that would be .68, well ahead of Palmer.

Now this all assumes that everything that happens on BIP is up to the fielders. Even if I cut the run change in half for each guy, Blyleven is still ahead .65 to .63 in RSAA per 9 IP. I also assumed that .55 is the run value of each event prevented (or not prevented). It could have been a little lower for Palmer if more of nonHRs were singles. It could have been higher for Blyleven if they were less and 2Bs and 3Bs were a higher %. It is probably not a big deal either way.

*The linear weights run value comes from Pete Palmer. Here are the run values for particular events:

1B = .47
2B = .78
3B = 1.09

That is, for example, every additional single adds .47 runs over the course of a season. I once found a weighted average of these three events of .55, weighted by each event's own frequency.

9 comments:

  1. How about Jack Morris? Many people during the HOF debates of December and early January claimed that Morris was helped by an above average defense. Is that true, by your method and how significant was it, if true?

    ReplyDelete
  2. Thanks for dropping by and commenting. I agree that Morris is worth looking at and I will try to do that soon. My guess is that he had better than average fielding behind him with guys like Trammell, Whitaker and Chet Lemon in CF.

    ReplyDelete
  3. I found your post very interesting.

    I was inspired to look at average league DER by decade, weighted by yearly BFP, to get a sense of how large in magnitude your team-to-league career adjustments were.

    1920s .709
    1930s .709
    1940s .727
    1950s .729
    1960s .734
    1970s .725
    1980s .718
    1990s .707
    2000s .704

    It seems to me that the effects of team defense you've highlighted are quite similar in magnitude to the differences across these different eras in MLB history.

    ReplyDelete
  4. Thanks for doing that and sharing your numbers. I guess that when hitting is higher and runs are higher, DER is lower (although not a perfect correlation). Probably also when fielding pct improves, DER is going up (although that is probably not a perfect correlation). Your chart puts everything in context. I think this site will be of interest to you:

    http://www.mlbnotebook.com/2010-articles/january/jim-palmer-got-a-little-help-from-his-friends.html

    ReplyDelete
  5. It is called:

    Jim Palmer Got A Little Help From His Friends

    and it lists runs saved by fielders for Morris, too.

    ReplyDelete
  6. Thanks for the link.

    Here's my cut at adjusting Jack Morris' RSAA:

    RSAA: 78
    IP: 3,824

    TeamDER: .726
    LgDER: .717
    deltaDER: -.009

    BIP: 11,805

    fielder extra hits: -106
    fielder extra runs: -58

    Jack Morris RSAA/9IP: .18
    Jack Morris adj RSAA/9IP: .05
    Jack Morris 50% adj RSAA/9IP: .11

    ReplyDelete
  7. Bob

    Thanks for doing that. I am surprised how low Morris's RSAA is to begin with. Per 9 IP he is not even close to Blyleven, no matter how we slice it.

    Cy

    ReplyDelete
  8. I was surprised too. But I looked at a few data points, and RSAA/IP seemed correlated with ERA+ at first glance, as you'd expect.

    ReplyDelete