The Thin Line Between 1B and DH

[agrudez]

Despite being a front office that prioritizes ground ball pitchers, I've never really prioritized 1B defense. That's where your put your "2nd DH", right? Well, I'm re-thinking that a touch after seeing just how brutal Norris Rutledge was last season - a -14.9 ZR. And that shouldn't be a surprise, he had a -15.6 the year prior, -14.9 before that and -12 two seasons before that. He's *always* been a brutal defensive 1B - I just never even bothered to check since I didn't think it mattered. So, why do I think it matters now? Well, because across the rest of my team (C, 2B, 3B, SS, LF, CF, RF) we were "pretty average" (noting that I am assuming a '0' is relatively "average" - which may be a very poor assumption) - netting a -2.1 ZR - yet were 12th in BABIP in the JL. So, it seems Rutledge is the low hanging fruit for why we were near the bottom of the barrel defensively as a team. So, that got me asking myself... if 1B "matters" after all, then what is the 'cut-off' for when a player needs to be a DH instead of a 1B? Below is a compilation of the 1B that logged the most innings at the position for their teams in the JL last year with a relevant plot attached.

ZR IP ZR/IP Range Error Player with most innings at 1B on team
MON -12.5 819 -0.015262515 1 4 Ettienne LaFitte
HSV -14.9 1350 -0.011037037 1 4 Norris Rutledge
WIC -5.8 615 -0.009430894 1 3 Brian Clough
JAC -4.6 1294 -0.003554869 2 3 Manuel Martinez
RCK -2.1 1313 -0.001599391 2 5 Manuel Marino
LOU -1.6 1262 -0.001267829 3.00 4.00 Pedro Debesa
AC -0.4 1084 -0.000369004 2 5 Adrian Salazar
NSH 0.6 1251 0.000479616 4 2 Miguel Suarez
LV 0.8 995 0.00080402 3 7 Gervasio Ridder
BRK 2.6 1320 0.001969697 3 5 Cisco Guerrero
SA 2.1 658 0.003191489 2 5 Manuel Nunez
PHO 4.5 1366 0.00329429 3 4 Mario Deortez
MC 4.1 1138 0.003602812 3 6 Mike Ellis
CC 2.7 492 0.005487805 6 7 Wendel Clark
NO 10.4 1367 0.007607901 4 7 Yancy Cravat

Looking at the plot attached, it seems pretty clear that 1 range is a no-go as all 3 players that had it were brutal at 1B (though, granted, we didn't get a 1 range, high error rating sample to see how that might affect the level of disfunction). In case you're having trouble interpreting ZR/Inning, a -0.01 ZR/Inning is a -14.58 pace over a full season (162 games * 9 innings = 1458 innings). On the next level of range (2), you start to see a pretty wide range of competence. The poor error rating sample is the worst of the four - as you might expect; however, the other three are pretty widely spaced despite the same error rating. If I had to guess, I'd say the one sample with a positive ZR at 2 range is an outlier (either in-season luck or a rating that is close to the next level on the 1-250 scale). It isn't until we get to 3 range that you start to see consistent positive ZR (4 samples positive vs. 1 sample negative). You also see almost no consistent trend in error rating vs. performance, though with a dataset so small that doesn't necessarily lend to any conclusions. At 4 range you have both positive with the low error rating being just barely so and the high error rating being the best of the dataset - even better than the 6 range sample with the same error rating (not that this should be a conclusive comparison).

So... what's my conclusion from this? Blankly, 1 range is unacceptable. 2 range is fine, but don't except solid, consistent defensive contributions even if the error rating is good. 3 range is when you start really getting into the realm of competence and 4+ range is when you start to see "good" defense with, potentially, some diminishing returns on range from there. Within that, it seems like <=3 error will start to tank your range's contribution to defensive performance. 4-5 is where you start to feel confident in competence - though the tangible fluctuation in performance in this particular SSS leads me to believe this isn't a 'comfort' zone. That makes >=6 the real benchmark for "good".

Error color scale for chart:
2 = red
3 = orange
4 = yellow
5 = light green
6 = dark green
7 = cyan

[bcslouck]

Nicely done. I do always try to have a decent defensive 1B. Unless the bat it truly elite. I always figured a bad 1B hurts your whole defense as you have your 2B, SS, and 3B making regular throws over there while the P and C make the occasional throw. Expanding on this exercise would be to compare overall infield defenses and sorting by 1B by ratings and results. May be tough to do and a lot of excel formulas since there are a lot of factors.

[RonCo]

Nice work. You've got me thinking about cutting a "Plays Above Average" chart for each position out of my defensive script.

My time is getting short, but I'll see what I can do. But in the interim, here are Huntsville's PPA for all 1B zones:

Code: Select all

PLAYS	3L	3	34
SHAL	0	-6.24	-1.91
MID	-4.64	-4.19	-6.39
DEEP	0	-1.46	5.17

Discounting the "34" zones (where the 2B probably makes a bigger difference, Huntsville's 1B were 16.5 plays worse than if you had fielded an average 1B.

Line: -4.6 Plays
Main 1B Zones: -11.9 Plays

[RonCo]

Nicely done. I do always try to have a decent defensive 1B. Unless the bat it truly elite. I always figured a bad 1B hurts your whole defense as you have your 2B, SS, and 3B making regular throws over there while the P and C make the occasional throw. Expanding on this exercise would be to compare overall infield defenses and sorting by 1B by ratings and results. May be tough to do and a lot of excel formulas since there are a lot of factors.

Yes, the saving of throwing errors by a good 1B is hard to quantify, but I'm sure it's there.

To back up your CCJ 1B defense's value above that, here's your CCJ Plays Above Average on ground balls in each 1B zone:

Code: Select all

PLAYS	3L	3	34
SHALLOW	0	-1.79	2.28
MID	1.62	3.45	4.99
DEEP	0	-1.46	0.59

[RonCo]

So... what's my conclusion from this? Blankly, 1 range is unacceptable. 2 range is fine, but don't except solid, consistent defensive contributions even if the error rating is good. 3 range is when you start really getting into the realm of competence and 4+ range is when you start to see "good" defense with, potentially, some diminishing returns on range from there. Within that, it seems like <=3 error will start to tank your range's contribution to defensive performance. 4-5 is where you start to feel confident in competence - though the tangible fluctuation in performance in this particular SSS leads me to believe this isn't a 'comfort' zone. That makes >=6 the real benchmark for "good".

Yes to all this, though the one caveat is to remember that (assuming OOTP is even close to properly coded) ZR is a relative stat...so if every team in the league punts defense for offense at 1b (puts a 1 range there), then a 1 will be just as competitive as anything else. This is important to think about in the context that, assuming every GM in the league reads this and begins to put better defenders in the field, this scale shifts upward.

I love the game theory part of this world.