The discipline of teams
“Team potential exists anywhere hierarchy or organizational boundaries inhibit good performance.”
Let’s say the Mariners can add only one batter and it has to be either Trevor Story or Brett Phillips. Ignoring cost, who should they take for 2022?
Story is one of the best hitters in MLB and plays a premium position; Phillips is a platoon outfielder. Story was worth 3.5 fWAR last year; Phillips was worth 1.8.
Story is the better player.
But Story and Phillips would fill unique roles on the Mariners roster, and either’s presence would shift the entire depth chart. Some players would gain at bats and others would lose.
The goal for the Mariners isn’t simply to collect good players — it’s to win games. And while those often require the same conclusions, understanding how they diverge is crucial to building a competitive team.
If the Mariners are to truly prioritize winning, they must view their roster as a discrete unit of performance and not just a set of positive values.
Naked WAR
A player’s expected value changes every game.
For instance, a hitter is likely to produce fewer runs in a game started by Jacob Degrom than in a game started by, say, Vinny Nittoli. Similarly, a second baseman is likely to prevent fewer runs when asked to play out of position at catcher.
Most of these shifts in game value can be generalized with two variables: player position and opposing pitcher handedness. I measured these changes using a stripped down (naked, if you will) version of WAR.
The inputs are quite simple:
- wRAA per 650 plate appearances against each handedness
- DRS per 1,200 innings at each position played
This allow us to compare compare a player’s run value in different game situations. Here’s Abraham Toro’s run value splits from 2021 as an example:
There are two ways to conceptualize these numbers:
- Think of them as answering, “How many runs would Toro be worth if he played a full season exclusively at 3B and against LHP?” Basically, these are season-scaled WAR splits. This requires a bit of imagination but is fairly intuitive if you can let go of reality.
- Or, think less about the numbers themselves and more in terms of relative value, almost like the basis for tOPS+ but for WAR.
We see that Toro was anything from solidly above average to borderline unplayable depending on the day and the circumstance. This tracks with his actual season scaled wRAA + DRS of -4.23, given that he mostly played second base and faced lefties more often than league average.
This framework can be applied to all sorts of relevant questions. For instance, what is the value of playing Toro exclusively at 2B versus 3B? To find out, I added LHP runs to RHP runs and then weighted for pitcher handedness (about 70% of pitches are thrown by RHP). With that method, a full season of Toro at second base would be worth -8.26 runs and compared to 3.15 runs at third base — a difference of about 11.5 runs over a full season.
Usage can accentuate or stifle production.
Accounting game
This method can be expanded further to compare lineups or entire rosters. For instance, do the Mariners produce more runs with Story or Phillips?
I established a pair of lineups for each scenario. I didn’t follow any sort of method — I just went with what felt right. I filled in the run value splits for the lineups using the method above, added them up for each lineup, and weighted by the handedness frequency.
(Story plays second base in this hypothetical. Since he’s never actually played there, I gave him a DRS of 15, which would make him the best defensive second baseman in MLB.)
Here are the results with Story lineups in blue, Phillips lineups in green, and status quo lineups in orange:
The Story and Phillips rosters produce identical value despite Story being worth 2–3 times as many runs. A platoon and some minor lineup shuffling closed the gap entirely.
Now, let’s say the Mariners replace Jarred Kelenic with someone whose values are less extreme — I’ll call him Player JK. On defense, Player JK is exactly average in left field and quite bad but not necessarily unplayable in center field. And on offense, Player JK is overall productive but with pronounced splits. (Perhaps this is similar to our expectations for Kelenic in 2022.)
Here’s that scenario:
The Phillips Mariners produce nearly 10 runs more than the Story Mariners, and even the status quo roster is slightly more productive.
Why?
Story’s presence forces Adam Frazier to play left field where he isn’t nearly as strong; Frazier alone is about 8 runs less valuable with Story commanding second base. And Frazier in left field means Player JK is forced to play center field where he’s also relatively worse.
Story is good but not good enough to make up for that forfeited value. So while the Story Mariners technically have the better collection of individual players, the Phillips Mariners are better positioned to maximize the value of each of their players.
The discipline of teams
So would I take Phillips over Story?
Probably not in a cost-neutral environment.
There are many things this method doesn’t address, and my fingers are all over the scales. To be entirely transparent, I tried this hypothetical with several players before settling on the pair I found the most convenient.
But despite its mechanical flaws, I think this framework has some merit.
The game environment changes every day, and players are primed for more or less success. We assume this is unavoidable and evaluate them based on their ability to adapt to all situations. Good players are generalists and good teams are looking for the most good generalists.
But the framework presented here turns that around. It puts the onus on teams to read the environment and adapt in a way that positions its players for success. Rather than finding a collection of individual players who each offer a package of strengths and weakness, teams build a cohesive unit that highlights collective strengths and hides collective weaknesses. Generalists here are still quite valuable, but “platoon” is no longer a pejorative. Teams are able to see players as the sum of their strengths.
This suggests that, from the team perspective, most batters produce similar value. Any player between 1.5 and 4 fWAR (or thereabouts) could potentially be replicated with clever shuffling. And while I don’t think I’ve accurately identified the margin, it’s not unreasonable for a team to look at a star shortstop and a platoon outfielder and come to a similar conclusion.
This has massive implications for team building, of course. It places roster flexibility as a top priority — a true competitive advantage — and means few players are worth more than the roster spot they occupy. Marginal value at the team level becomes the driving force behind all decision making.
This accounting framework, with some adjustments, could be a useful tool for understanding player value and team building.
This is not to say good players don’t matter. Certainly no amount of shuffling could have made Pirates good last year. Teams should still focus on creating, acquiring and holding onto good players. But teams should truly prioritize collective performance and redefine who and what is “good.”
Katzenbach and Smith wrote about this concept for the Harvard Business Review in 1993.
“Team potential exists anywhere hierarchy or organizational boundaries inhibit good performance,” they wrote.
Teams perform at their highest when they can leverage their members unique abilities to get a level of production greater than the sum of all their individual bests — this is what distinguishes a team from a working group or a committee.
“When a small group of people challenge themselves to get over a wall or to reduce cycle time by 50%, their respective titles, perks, and other stripes fade into the background,” Katzenbach and Smith wrote. “The teams that succeed evaluate what and how each individual can best contribute to the team’s goal and, more important, do so in terms of the performance objective itself rather than a person’s status or personality.”
####
