Defensive Responsibility Spectrum (DRS)

Breaking up the fielders from the pitchers

By Tangotiger

The Spectrum (from 100% pitching to 100% fielding)

If we look at all those things that the defense is responsible for, let's look to see how much should be at the hands of the pitcher and how much at the hands of the fielders.

HBP, BK, PO, K, BB, HR, WP, SB, CS, 2B, 3B, 1B, batting outs, PB, other running outs

That is, a hit batter is 100% the responsbility of the pitcher, and 0% of the fielders, while the other running outs are 100% the responsbility of the fielders and 0% of the pitchers. Everything in between is in question.

Balks, Pickoffs

When it comes to balks, you can say 99%+, as there MIGHT be some small difference if Keith Hernandez or Frank Thomas at 1B. I highly doubt it, but I suppose there is that sliver of possibility.

PO again might have some relationship to the 1B. The swipe of the 1B to tag the runner out might come into play. And I believe that a pitcher is credited with a PO, even if the runner makes it to 2B, so that we don't have to consider the relay throw to 2B. So, we are probably at the 99% level for pitching responsibility.

Walks, Strikeouts, Homeruns

For BB and K, again, the catcher MIGHT account for some difference (the way he sets up, etc), but again, we can safely say that 95%+ of the responsibility should go to the pitcher.

The HR does have some influence if you have Gary Pettis, or Devon White, etc, and the park you play in. If we ignore the park discussion, and only concentrate on the relative responsbilities between the fielders and pitchers, we are probably at least 90% the pitcher, if not at least 95%.

Wild Pitches, Passed Balls

Wild pitches might have a stronger relationship with the Catcher, and so we can say that maybe the fielders have a 20% responsbility here.

The passed balls might have a strong relationship with the pitcher, and so we can say that maybe the pitches has a 20% responsibility here.

Drawing the lines

SB, CS, 2b, 3b, 1b, Batting outs are the "grey area".

So if we draw lines, we are left with

WP, SB, CS, 2B, 3B, 1B, batting outs, PB,
other running outs

The first line is almost all pitching, the last line is all fielding, and the line in between is a mix of the two.

So, when we talk about fielding-independent stats, we are talking about the 1st line. Since HBP, BK, and PO are either hard to come by historically and have such small quanitities anyway, we are left with

Fielding-Indpendent Pitching Runs

So "basic" fielding-independent pitching can be evaluated as
FIP = (13*HR+3*BB-2*K)/IP
for reasons elaborated on in past threads.

If you want to make this number more "meaningful", you can say
runs = (league FIP - player FIP) x IP / 9

For example, Randy Johnson's FIP is -1.14, while the league average is +1.22. That makes Randy +65 runs.

The top 10 in FIP are virtually the same as the dipsERA I published earlier

Randy Johnson	 2.32 	 (1.14)	 65 
Roy Oswalt	 2.92 	 (0.33)	 24 
Mike Mussina	 3.03 	 (0.18)	 36 
Matt Morris	 3.16 	 (0.18)	 34 
Andy Pettitte	 3.13 	 (0.11)	 30 
Greg Maddux	 3.19 	 (0.02)	 32 
Terry Adams	 3.23 	 (0.02)	 23 
Curt Schilling	 3.17 	 0.05 	 33 
Javier Vazquez	 3.29 	 0.13 	 27 
Kelvim Escobar	 3.34 	 0.14 	 15 

The only difference is that dipsERA used HBP, BK, and WP. If I were to use the same components, the lists would be identical in order.


Anyway, I have not introduced anything here that we don't already know (by now) with DIPS. DIPS tries to separate the fielding from the pitching, but in so doing, Voros "assumes" league average success rates for the "grey area" components. I think this obscures some important elements of DIPS, namely $H. And introduces people's interpretation of DIPS, namely as a projection, or catch-all, etc, etc.

When discussing DIPS, I think we should be clear on this matter. We don't get hung up with OBA treating all bases as "1", so we should do the same with DIPS, which is very apparent with FIP.

Note:FIP is fielding-independent pitching, and therefore tries to isolate those skills a pitcher has that is fielding-independent. If I start using the fact that he gives up alot of HR as a stand-in for all the extra doubles he gives, I am removing this.

What you want to end up is ERA = FIP + FDP + Fielding + luck

That is why I broke up the spectrum into 3 lines, as each line corresponds to the 3 components (FIP, FDP, Fielding). Luck of course always plays a part in everything.

Therefore, I can't start mixing and matching by changing the value of the HR. I've only given you one part of the equation.