Why I Don't Care About Platoonpalooza 2014

Replacing Dexter Fowler with Drew Stubbs means Colorado will have less of a platoon advantage in 2014. - Jeff Hanisch-US PRESSWIRE

There's been a lot of talk this winter about the Rockies' plans to rely heavily on platoons next year. The point has been made that, while the plan sounds nice in the abstract, it's unlikely to actually materialize, because a) injuries happen, and when they do, they'll break up the platoons, and b) we can't trust that the team will consistently place platooning front and center, even if there aren't any injuries (i.e. what happens if we face lefties on back-to-back days, and Stubbs goes 6-8 with a HR or something? Is Weiss going to do the smart thing and sit him against a righty the next day, or is he going to trust in the "hot streak" and play Stubbs the next day anyway, forsaking the platoon advantage?). Those are absolutely important things to keep in mind. But there's one point that hasn't yet been made clearly enough: even if executed perfectly, our platoons aren't likely to accomplish more for us than what other teams are getting from their own platoon arrangements. Let me explain.

First things first: sample platoon splits are extremely unreliable. According to The Book, we have to regress lefties’ platoon splits against 1000 PA against LHP of league-average splits for left-handed hitters, and righties against 2200 PA of league-average splits for RHB against LHP. So for a guy like Drew Stubbs, who has 451 PA against lefties over the past three years (any data older than that is too remote to be of any real use in projections), his estimated true platoon split will be regressed 2200/(2200+451)= 83% towards the league average platoon split for a RHB. Going through this procedure for all the hitters on our projected roster, we get the following wOBA splits:

Name Bats Sample Split True Split Estimate
Justin Morneau L .124 .062
Carlos Gonzalez L .068 .045
Corey Dickerson L .087 .034
Charlie Blackmon L -.039 .024
Josh Rutledge R .041 -.017
DJ LeMahieu R .014 -.019
Troy Tulowitzki R -.006 -.020
Charlie Culberson R .008 -.022
Jordan Pacheco R -.017 -.022
Michael Cuddyer R -.036 -.025
Nolan Arenado R -.082 -.026
Drew Stubbs R -.077 -.032
Wilin Rosario R -.117 -.033

Those numbers are just straight wOBA differentials, wOBA vs. RHP minus wOBA vs. LHP. The "sample split" is the player's actual performance over the past three years. So my estimate for Morneau is that he's 62 points better vs. RHP than LHP; on the opposite end of the spectrum, Rosario is estimated at 33 points better vs. LHP than RHP. For reference, the average splits are 32 points for LHB and -23 points for RHB... so Morneau's estimated split is nearly double that of the average left-handed hitter (but still only half his observed split - I cannot emphasize enough how much randomness goes into sample platoon splits, and how important it is to regress them).

So now we have estimated true platoon splits for all of our hitters. The next step is to put those numbers in context and convert them into runs and wins. Now, there are two logically distinct arguments being made in favor of our platoon plan:
1) Platooning is beneficial in general.
2) Platooning is especially beneficial with the collection of hitters that we've assembled.

Let's start with the second argument. Armed with the estimates provided above, combined with the knowledge of how players are used on average (league-wide, LHB get 20% of their PA vs. lefties and 80% vs. righties; RHB get 36% vs. lefties and 64% vs. righties), we can determine how much our hitters would benefit from "perfect" usage, relative to the amount that average hitters would benefit from same. Here's that table, along with some rough playing time estimates:

Name MaxBenefit Proj. PA
Wilin Rosario .007 474
Justin Morneau .006 509
Drew Stubbs .006 256
Carlos Gonzalez .002 529
Nolan Arenado .002 585
Michael Cuddyer .001 477
Corey Dickerson .000 567
Jordan Pacheco .000 256
Charlie Culberson -.001 256
Charlie Blackmon -.002 256
Troy Tulowitzki -.002 531
DJ LeMahieu -.002 577
Josh Rutledge -.003 256

This means that Rosario's wOBA, if he were used exclusively against opposite-handed pitching, would increase by 7 points more than the average RHB's wOBA would increase in those same conditions (and that Morneau's wOBA would increase by 6 points more than the average LHB's wOBA would, etc). The negative numbers for the guys at the bottom of the list indicate that they have smaller than average platoon splits, and therefore that they'd derive less benefit from being platooned than an average hitter would.

Now, if we weight each player's maximum platoon benefit by his projected PA total, we get the overall team average: the Rockies would benefit from perfect platooning by a margin of 1.4 points of wOBA more than the average team would. 1.4 points of wOBA is equal to .0011 runs per plate appearance. Over a full season, that works out to about 6.5 runs. So that's the value of our hitters' extreme splits: 6.5 runs, if absolutely maximized. But of course that "absolute maximization" is an absurd hypothetical; no team can possibly ensure that its hitters face opposite-handed pitchers 100% of the time, or anywhere close to it. In the 2013 NL, the team that had the platoon advantage the least often was the Brewers, 41.7% of PA; the team that had the platoon advantage the most often was the Giants, 64.9% of PA. League average was 52.0%. So going from 52% to 100% would mean a 6.5 run advantage for us, but in practice, it's basically impossible to get up above 65%. Which means that, more realistically, our hitters' extreme splits give us an advantage of no more than 1.5-2 runs over the course of a season.

So all this talk about Morneau and Stubbs' extreme platoon splits is much ado about nothing. But what about the other question, the value of platooning in general? Well, let's go back to the numbers I gave in the previous paragraph: the Giants this year led the league in percentage of PA with the platoon advantage, at 65%, compared to a league average of 52%. That extra 13% amounts to around 750 platoon-advantage PA above average over the course of a season. And what's the value of a platoon-advantage PA, you ask? This is easy: as I said earlier, the average platoon advantage for a LHB is 32 points, for a RHB 23 points. 63% of platoon-advantage PA are LHB vs. RHP, 37% are RHB vs. LHP. So therefore, on average, the platoon advantage is 28.6 points of wOBA, or .0224 runs per PA. Over 750 PA, that's 16.8 runs.

So leading the league in platoon-advantage PA is worth close to 2 extra wins per year. That sounds pretty good. But the crucial question here is how likely we are to actually pull it off. And unfortunately, our chances are slim to none. Why? It's simple. The #1 factor in having more platoon-advantage PA is giving more PA to left-handed hitters, because the overwhelming majority of pitchers are right-handed. Looking at all 30 teams this year, the correlation between percentage of PA given to LHB and percentage of PA with the platoon advantage was .83, which is absolutely huge.

And we do not have a lefty-heavy offense; quite the contrary, in fact. In this year's NL, 40.9% of PA went to left-handed hitters; for the Rockies, that number was only 36.4%. And the 2014 Rockies aren't going to be any more lefty-heavy than the 2013 team was; all we've done is replace Helton with Morneau and Fowler with Stubbs, which obviously means a shift in the other (unwanted) direction - next year we'll have even more PA going to right-handed hitters. Based on the rough PA projections I gave above, we'll be giving just 33.7% of PA to LHB.

This year, the team that best maximized the platoon advantage relative to its overall team handedness was the Astros, who had the platoon advantage 57.5% of the time despite giving only 37.9% of PA to left-handed hitters. That's a 19.6% difference... if we add that to the 2014 Rockies LHB PA % forecast from the previous paragraph, we can conclude that if deployed optimally (well, not "optimally", but "as well as or better than any other team in baseball", which is plenty optimistic in its own right), we could have the platoon advantage as much as 53.1% of the time. But again, the league average is 52.0%. Remember, having the platoon advantage in an extra 13% of PA, as the Giants did this year, is worth 16.8 runs... so having it in an extra 1.1% of PA is worth only 1.4 runs. And that 1.4 runs is our best-case scenario.

So there we have it: even if we deploy our players as well as is realistically possible, the expected value of our platooning (relative to the platooning done by an average team) is 1-2 runs over the course of a season. That's not to say that platooning is pointless. It's not. Just as leading the league in platoon advantage PA can be worth almost two extra wins per year, being at the bottom of the league in that category can cost you the same amount. So there's value in avoiding that scenario. But the more important point is this: it's almost inconceivable that we will derive significantly more value from platooning than the average team will. In other words, we're not getting a leg up on the competition here, all we're doing is keeping pace. Which means that there's still no real reason to think that this team will be more than the sum of its parts. And the sum of its parts isn't anything to get excited about. All of this platoon talk is just a way of distracting people from the fact that Morneau and Stubbs are bad baseball players.

Eat. Drink. Be Merry. But the above FanPost does not necessarily reflect the attitudes, opinions, or views of Purple Row's staff (unless, of course, it's written by the staff [and even then, it still might not]).

Log In Sign Up

Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

Join Purple Row

You must be a member of Purple Row to participate.

We have our own Community Guidelines at Purple Row. You should read them.

Join Purple Row

You must be a member of Purple Row to participate.

We have our own Community Guidelines at Purple Row. You should read them.




Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.