Platooning: The value for a team (Part 2)

Here’s a scary thought if you are an AL West team with only one geographical entity in your name: the Angels will have Josh Hamilton and Albert Pujols in their lineup this year, and there’s a good chance that neither will be their best hitter.

True, those guys did cost $365 million to get—which, as our astute readers immediately noticed, would enable you to buy your own island in the Caribbean every day of the year—but such is the harsh reality of owning and generally managing a baseball operation nowadays. If you want premier production, you have to pay for it. A lot.

Here is a funny fact, though. Compare these 2012 productions:


The first one is the Angels’ new first baseman/left fielder duo. Second one is 2012 Oakland’s first baseman/left fielder quartet, a construction of cheap, one-armed misfit toys, if you are rich like me and consider anything below $5 million to be but change money.

The two are eerily similar in everything but financial commitment and number of heads, and it was the production of the latter that was pinpointed by many as the main reason for the A’s having won the division. But which was it? Was it another brilliant stroke made by the market inefficiency explorer extraordinaire or rather a standard case of unicorn power turning 2012 versions of Jonny Gomes, Seth Smith, Brandon Moss and Chris Carter not only into more enlightened and better human beings, but also into more productive baseball versions of their old selves?

The four players named above offer different stories of success, some more expectable, some less. To make a long story short, Carter showed some platoon split last year, in line with his minor league performance, although the big news with him is that he hit anything, let alone both left-handers and right-handers.

Moss, he of neutral major league career platoon splits, just posted a wRC+ of 172 against right-handers, a mark 70 percent better than his career norm and six points higher than what both Miguel Cabrera and Mike Trout managed to scrape together.

Smith, probably platooniest of the four platoon players, performed just as he always did, having his wRC+ splits of 47/120 (versus career norms of 46/122) and facing only 18 percent of left-handed pitchers (versus a career norm of 17 percent).

If the A’s deserve big credit somewhere, it’s with Gomes, not as much for his improved performance against both left-handed and right-handed pitching, as for his improved usage, since they increased the percentage of his plate appearances against left-handed pitching from his 33 percent career norm to 59 percent in 2012.

So, the Athletics got mostly lucky on two platoon players, received just what they bargained for on the third one and were lucky and smart on the fourth one. But even if the output was on the extreme positive end of the expected spectrum, did the A’s try harder than other teams? In other words, did they invest more in seeking the platoon advantage than other 29 teams did?

To answer that, we have to make the number of platoon match-ups somehow quantifiable. There are probably different ways to do that, and what I chose might not be the best, but the odds are you haven’t done it at all, so my approach is still better than what you have, much like fixing the broken transmission on your four-wheel drive in the middle of the Gobi Desert by hammering sewing needles around the axle. No, I’m not kidding.

I decided to look only at starter-versus-starter confrontations. Why? Because I liked the purity of it. It’s like Effron’s dice—the pitching team is forced to make the first choice (it’s not really a choice, as starting pitchers are used in order, not to create match-up advantages), which enables you to choose your hitting team in a way that will give you the best odds to win.

Everything that happens later in the game, once the pinch-hitters and relief pitchers enter the stage, is a game of action and reaction where both teams have a chance to turn the handedness of the confrontation in their favor. As I said, it’s not the only way to go about it, but it is the one I chose.

First we need a baseline. We can not simply compare the number of same-handed or opposite-handed plate appearances for any two players without controlling for the pitching environment their teams faced.

As you see, getting your right-handed hitting specialist enough at-bats was significantly more difficult in the National League Central than in the NL West last year. Now, knowing this, we can compare the splits in any hitter’s plate appearances versus opposing starters to what the baseline for his team was.

Let’s stick with Gomes for a second. Against starting pitching, he had 210 plate appearances, 165 of which were against left-handers. That’s a staggering 79 percent, or more than double of what Oakland as a team had (36 percent).

A Hardball Times Update
Goodbye for now.

Had Gomes been used as often versus left-handers and right-handers as the team baseline suggests, we would expect only 75 (36 percent of 210) of his plate appearances to come against left-handers. So, by platooning him, the Athletics were able to “convert” 90 of those plate appearances in his favor. That number is what I call “PlatoonPA.” The bigger the number, the more platooned the player was in 2012.

There were 25 players who received at least 50 PlatoonPA in 2012.

I actually have this Pavlov reflex where you will say Raul Ibanez, and I will say lawn dart. However, this is almost equally impressive: in 2012, Ibañez had 251 plate appearances versus right-handed starters and only two against a left-hander. You might remember the game. It was when Robinson Cano, Nick Swisher and Andruw Jones went back to back-to-back on Johan Santana. Luckily for Santana, the next batter was Ibañez.

However outlandish that 251:2 ratio might look, it created only three more PlatoonPA than Gomes’ 165:45. That’s because all teams face more right-handed starters, and even had his 253 plate appearances been spread out, Ibañez would have had 158 of them against right-handers. Life is good when you are left-handed.

How about the added value of a single PlatoonPA? Surely there is more value in a Gomes plate appearance vs. a left-hander or an Ibañez plate appearance versus a right-hander than against the pitcher of the same handedness. So by switching their regular mix of opportunities into a platoon mix and increasing the run expectancy, the value is being created. But how much value?

Let’s look at our 25 most platooned players of 2012 again. For each, I have listed his observed career wOBA split and his regressed one. For the regressions I used the method described in the first part of this mini-series.

The weighted average of observed platoon splits is about 0.050 points of wOBA, that of the regressed one is 0.035. Now, this is obviously a very small sample size, but the fact that the values are higher than generally used to estimate the platoon splits could also be because we are not looking at the whole population of major league hitters. Instead, we are looking at the players who, for some reason or the other, earn their bread—there might be even a little something left to buy the butter—as platoon hitters.

It is a perfect example of a biased selection, so there is a possibility that by regressing them toward the mean of the general population, we neglect their true skill. Or, it might be that the managers are biased in selecting such players for the platoon roles based on their unattainable past performance.

So, there is a good chance that the following numbers are too conservative, as they are calculated with league average numbers (0.035 for LHB and 0.023 for RHB). Using these numbers to calculate the added value due to platooning for each player and then adding those added values on a team level, we can both compare which teams used the tactic the most and roughly estimate the expected increase in run creation due to it.

And no, it wasn’t the A’s who platooned the most but the New York Mets. Oakland came in second, followed by Chicago Cubs, San Diego Padres and Boston Red Sox. The teams that used platooning the most gained just over a win from it.

Apart from that number being an imprecise science, one has to consider further factors. Additional platoon plate appearances through pinch hitters and/or against relievers should bring that number up. The cost of a roster spot also must be considered, as it is of limited supply, and it should bring that number somewhat down.

For now, though, a somewhat educated estimate is that by aggressively matching up a lineup—provided you have the players to do so—a team stands to gain about one to one-and-a-half additional wins per season.


Bojan Koprivica played soccer in the USA and baseball everywhere else. While some use this fact to question his competitiveness, he talks of it using expressions like "international flair".
4 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
studes
11 years ago

Great stuff again, Bojan.  I’m a little unclear how you calculated the final table.  You used average splits instead of batter-specific splits?  Also, I don’t understand your comment about bringing the number up through pinch-hitting.  Wasn’t that factored in already?

Bojan Koprivica
11 years ago

Thanks, Dave.

Yes, I used average splits instead of batter specific ones. Using batter specific ones would be better, but I didn’t have everybody’s career splits in my DB, meaning that I had to regress manually, so I cut a corner there.

I could either parse the retrosheet data for further years into my DB, or just use last 7 years for everybody. I might rerun the data like that later to see if and how much the numbers increase.

As for pinch-hitting, it is not accounted for in the PlatoonPA, as the number looks only into starter versus starter confrontations for the reasons I mentioned.

Take Jonny Gomes again. He had 165 PA versus LH starters, when he was in the starting lineup, 45 against RH starters. That is 59% of PA versus left.

All of his other plate appearances (the ones when he faced a reliever, or was a pinch-hitter himself) he faced LHP 31 times and RHP 92 times. That’s only 25% of PA versus left, much more “normal” number. It basically means that the whole platoon advantage he accrued came from facing starters as a starter.

David Gassko
11 years ago

Great stuff, Bojan. As Tango would put it, excellent rolling-up-your-sleeves work.

Bojan Koprivica
11 years ago

And as on cue, the A’s go and acquire Jed Lowrie