Are Platoon Splits Almost a Certainty for the #Nats – 1 of 2?

Don: So that is the question of the day. Given the budget, who the available FAs are; and lack of tradeable assets, platoons seem like the best/only option for the Nats. TalkNats commenter DChamps asked us to do:

Here is what I would like to see if you can do it is take 2019/2020 combined and do a platoon split to see what a Brantley [Substitute Schwarber] platoon would look like with Pillar or Duvall. Do the same for Bell and Zim and Luis Garcia and Harrison and at catcher Gomes for 100 games and Jason Castro for 62 games and in CF Robles for 110 games and Stevo for 52 games.

Steve: Wow, I would love to see platoon splits or what Rizzo likes to call matchups for 2 players at the same position. Riz has never liked the pure lefty/righty thing. I think we have to determine how many at-bats those players would get because Duval and Pillar would play more than the 30 starts a year against lefties. This is also much different to look at since Josh Bell is decent from both sides of the plate as a switch-hitter.

Andrew: I think that is smart. This could be the Nats reality. My fear is that if they go with Brantley Schwarber that they play him too much.

BREAKING: Deferred by a day because the Nats have signed Kyle Schwarber. So we will replace Brantley with Schwarber in DChamps list.

Don: So let’s do this. We can start with Ryan Zimmerman and Josh Bell. That is the most likely since Bell is already on the roster and signing Zim as the backup is almost a certainty. So here is the logic for how I created these tables.

• As DChamps suggested, I used all the GameDay AtBat data for the two players in the platoon for 2019 and 2020. I have labeled that grouping as All. I show the players both individually and combined – and the combined results are shown as Combined.
• Since the Combined line won’t necessarily include the number of PAs for a full season, I created two other combinations both by combining the the players data and pro-rating the numbers to 684 PAs. That is roughly the number Baseball Reference uses for the their 162 game line item. I also looked at all the data I have downloaded and the average number of PAs/game for a player was 4.22. Do the arithmetic and you get 162*4.22=664 (rounded). I did two combinations:
  • 25/75 Share
  • 50/50 Share
• I then created a grouping called Platoon. It has the same five detail lines. But the data included is filtered to reflect a platoon. One of the two players is designated to start against righties and one against lefties. Using Bell and Zim as the example, if the opposing starter in the Game Day data was a righty, I included all of Zim’s PAs from that game; likewise if the starter was a lefty I included all of Bell’s PAs from that game. Note that this is different from the Baseball Reference platoon splits since those numbers are based on who the opposing pitcher is for each PA. That is not how a lineup for a given game is constructed. A platoon is going to be based on who the starter is.

So let’s look at the results for our 1B split: Bell against righties; Zim against lefties.

Method=All

Batter	PA	AB	BA	OBP	SLG	OPS	Hit	Sngl	Dbl	Trpl	HR	BB	SO
Bell	838	724	0.264	0.351	0.514	0.865	191	103	40	3	45	96	177
Zimmerman	250	226	0.257	0.320	0.416	0.736	58	38	12	0	8	22	56
Combined	1088	950	0.262	0.344	0.491	0.834	249	141	52	3	53	118	233
25/75	684	598	0.262	0.343	0.489	0.832	157	89	33	2	33	74	147
50/50	684	605	0.260	0.335	0.464	0.799	157	94	33	1	29	69	149

Method=Platoon

Batter	PA	AB	BA	OBP	SLG	OPS	Hit	Sngl	Dbl	Trpl	HR	BB	SO
Bell	647	560	0.271	0.355	0.527	0.882	152	80	35	3	34	72	127
Zimmerman	58	53	0.283	0.345	0.566	0.911	15	8	3	0	4	5	13
Combined	705	613	0.272	0.355	0.530	0.885	167	88	38	3	38	77	140
25/75	684	600	0.274	0.353	0.537	0.890	165	87	37	2	39	72	139
50/50	684	609	0.277	0.350	0.547	0.897	169	89	36	2	42	68	144

Steve: One of my favorite points to be made is that a good platoon creates a “whole that is better than the sum of the parts.” Somewhere you find the matchups where there would be a righty who Zim might hit better than Bell such as Michael Wacha for just one example. So this does not have to be strict platoon for Zim to only face lefty pitchers if there is a matchup advantage. Zim for instance hits righty Aaron Nola at a .250/.273/.531/.804 slash while Bell has never hit Nola (0-8). There are other examples like Johnny Cueto, Zack Wheeler and Adam Wainwright who Zim hits better than Bell.

It is up to the Nats analytics to create matchups instead of pure platoons. It gets more complicated if the Nats were facing Vince Velasquez as both Bell and Zim hit him hard. That is a good problem to have.

The point is, these should never be pure lefty/righty platoons unless a batter just can’t hit same-handed pitchers. I can tell you that Luis Garcia for instance hit righties very well (.323/.356/.444/.800) but he was bad against lefties with only 5 hits in 35 at-bats and no walks. In fact, if the lefty pitcher had crossfire action, fuggetaboutit. The rookie had a 37.1 K% against lefties. Against righties, he was looking like a budding All-Star and that is the point here. Matchups work. That is why today I would start Garcia at 2nd base with Harrison partnering with him and slide Starlin Castro to 3rd base.

Andrew: Those Garcia numbers are eye-opening. I agree the word ‘platoon’ gets people agitated and it needs to be explained that the key is playing matchup baseball. It does not always work-out as you could still go 0-4 when the matchup was favorable.

Clearly you came up with some examples of when Zim should face some RHPs. The same could be done for every player, and should be. This is also great for rest schedules and to strengthen the bench.

As statistically strong as Houston was supposed to be, they didn’t pick good matchups for Brantley, and we all saw Dusty botch the matchups with Yuli Gurriel in the postseason in obvious spots where they should have pinch-hit for him. Those are the ‘little things’ as Dave Martinez calls them.

With all of this said, Rizzo has never really embraced matchup baseball, and I think he needs to with players who have shown clear split issues then as Steve says take advantage of matchups where you can no matter if it is righty-on-righty.

Don: I agree on the match-up discussion. What these tables are intended to do is to provide some insights. To truly do match-ups, more information is needed – and I am willing to to bet the Nats analytics team is doing just that. Doing similarity analysis on pitchers so they can be grouped in order to have more PAs to analyze.

I wonder how the Schwarber signing impacts whether the Nats bring Zim back. I could see it could go either way: bring him back because of how much better he is against lefties and why would you want to sit Bell against righties for Schwarber; or the roster spot has to be used for someone else.

But time to shift gears now and look at platoons with Kyle Schwarber. So lets match him up with Kevin Pillar, Adam Duvall and maybe even Josh Harrison.

Schwarber – Duvall

Method=All

Batter	PA	AB	BA	OBP	SLG	OPS	Hit	Sngl	Dbl	Trpl	HR	BB	SO
Duvall	372	341	0.240	0.298	0.531	0.829	82	40	13	1	28	24	109
Schwarber	841	724	0.232	0.332	0.492	0.823	168	81	35	3	49	103	223
Combined	1213	1065	0.235	0.322	0.504	0.826	250	121	48	4	77	127	332
25/75	684	598	0.234	0.323	0.502	0.825	140	68	27	2	43	74	186
50/50	684	608	0.236	0.315	0.512	0.827	144	70	26	2	46	64	191

Method=Platoon

Batter	PA	AB	BA	OBP	SLG	OPS	Hit	Sngl	Dbl	Trpl	HR	BB	SO
Duvall	103	93	0.280	0.340	0.559	0.899	26	13	6	1	6	8	26
Schwarber	717	614	0.243	0.344	0.520	0.864	149	71	31	2	45	92	183
Combined	820	707	0.248	0.344	0.525	0.869	175	84	37	3	51	100	209
25/75	684	594	0.252	0.343	0.530	0.873	150	72	32	3	42	79	174
50/50	684	602	0.262	0.342	0.540	0.882	157	77	35	4	41	70	174

Schwarber – Pillar

Method=All

Batter	PA	AB	BA	OBP	SLG	OPS	Hit	Sngl	Dbl	Trpl	HR	BB	SO
Pillar	868	819	0.266	0.300	0.440	0.739	218	136	49	6	27	31	130
Schwarber	841	724	0.232	0.332	0.492	0.823	168	81	35	3	49	103	223
Combined	1709	1543	0.250	0.315	0.464	0.779	386	217	84	9	76	134	353
25/75	684	603	0.241	0.324	0.478	0.801	145	76	31	3	35	69	162
50/50	684	617	0.250	0.316	0.464	0.780	154	87	34	4	31	54	142

Method=Platoon

Batter	PA	AB	BA	OBP	SLG	OPS	Hit	Sngl	Dbl	Trpl	HR	BB	SO
Pillar	286	271	0.306	0.336	0.531	0.867	83	48	21	2	12	11	33
Schwarber	717	614	0.243	0.344	0.520	0.864	149	71	31	2	45	92	183
Combined	1003	885	0.262	0.342	0.523	0.865	232	119	52	4	57	103	216
25/75	684	601	0.260	0.342	0.523	0.865	156	79	35	3	39	72	151
50/50	684	617	0.276	0.340	0.526	0.866	170	91	40	3	36	57	127

Schwarber – Harrison

Method=All

Batter	PA	AB	BA	OBP	SLG	OPS	Hit	Sngl	Dbl	Trpl	HR	BB	SO
Harrison	237	215	0.214	0.270	0.321	0.591	46	32	9	1	4	12	38
Schwarber	841	724	0.232	0.332	0.492	0.823	168	81	35	3	49	103	223
Combined	1078	939	0.228	0.318	0.453	0.771	214	113	44	4	53	115	261
25/75	684	597	0.227	0.316	0.447	0.764	136	72	28	3	33	71	163
50/50	684	605	0.223	0.301	0.404	0.705	135	79	27	3	26	59	146

Method=Platoon

Batter	PA	AB	BA	OBP	SLG	OPS	Hit	Sngl	Dbl	Trpl	HR	BB	SO
Harrison	77	69	0.275	0.312	0.449	0.761	19	13	3	0	3	3	10
Schwarber	717	614	0.243	0.344	0.520	0.864	149	71	31	2	45	92	183
Combined	794	683	0.246	0.341	0.512	0.854	168	84	34	2	48	95	193
25/75	684	593	0.251	0.336	0.501	0.838	149	80	29	1	39	72	153
50/50	684	599	0.259	0.328	0.484	0.812	155	92	28	1	35	57	132

Frankly, none of these jumps out at me as something I would definitely want to do. At first glance, Duvall seems marginally better – slightly better On Base (i.e., keeping the line moving) and a little more pop – but more Ks. The question is whether he or Pillar are worth signing give that splits with Harrison aren’t that much worse.

Steve: Defense and strengthening the bench is the further consideration. This is a good dive in the shallow end showing how platoon baseball in simplicity improves the situation while strengthening the bench. To repeat, this is how the Dodgers have reached 3 World Series in 4 years. Moving Joc Pederson and Kiké Hernandez around in platoons and playing to their strengths. They did it with Cody Bellinger until they felt comfortable playing him every day.

Most teams don’t do this because it is expensive spending money on a quality bench. You need good players to pull this off.

Andrew: Yes and some good points and this isn’t inventing the mousetrap, it is just copying what others have done and customizing it to your own needs. If you believe in analytics and have positions of weakness, you have a chance to improve a couple of positions with matchups. With a 4-man bench plus a backup catcher, you can’t do too much matchup baseball. The Dodgers did it regularly at 2nd base and in one outfield spot, and as Steve said it they used to do it with Bellinger also.

I could see the Nats doing this with Schwarber and with Garcia for sure. I think if you have the ability to do it with Bell some, you want to do it. What you don’t want to do is never let your players ever face same-handed pitchers which is why I like matchups. Garcia clearly has to keep improving, but yes, a .323 batting second baseman who is making league minimum is attractive.

Steve: Solid points. You can’t do this at every position. Doing it with Schwarber would hopefully also replace him with a better defensive player. Rest schedules are enhanced, and you then have Schwarber available off the bench as a potential switch-hitter. That is why this is a positive in total.

I’m also a believer in young Luis Garcia. People are assuming that Kieboom is the starting 3rd baseman, and I wonder why? He’s the one that has to prove himself. The 20-year-old Garcia came in and proved himself. Yes, he struggled mightily against lefties. Again, expose him less against lefties and get Harrison in there where he mashed lefty starters last year.

This isn’t traditional baseball and purists hate this type of stuff, but the Dodgers have proved that it works.

Don: This approach has not only been used and validated by other teams; it is also a matter of necessity for the Nats. It will also be interesting to see what they do if we end up with the DH in 2021.

In part 2, we can look at other splits. This post is already long enough; plus I need a little time to update the logic in my programs to add more flexibility to the subset of the data that the 25% gets applied to.