“It’s ‘yes’ or ‘no’ from pitch to pitch,” Pat Murphy said. He waited almost a full fortnight into Cactus League play before putting an ABS challenge system strategy into action for his team, but the activation was not the slow turn of a dimmer dial. It was a hard flick of the switch. Beginning Thursday afternoon at Salt River Fields, Brewers batters no longer had free rein to feel out the system and test-drive it at their leisure. Two Brewers coaches were assigned to use colored cards to signal whether the situation justified a challenge on a borderline call or not, and every player in the lineup was expected to heed the signals.
One day later, the league had issued a hard flick of its own, thwarting that particular version of dugout-to-player communication on the premise that it falls outside the spirit of the new structure. Players are supposed to make challenge decisions without the help of the dugout; that’s been one of the key principles of the system since the league began testing it in the minor leagues. It’s a flimsy and misbegotten prohibition, though, and it will have little effect. The Brewers’ system wasn’t in place to tell batters (or catchers) to challenge or not; it was just a way to communicate when a challenge is contextually appropriate.Â
There are really two layers to every challenge decision, which is the major source of complexity in this new system. A player has to quickly gauge their own confidence that a call that went against them was wrong, but then, they also have to figure out whether the count, base-out situation and game state dictate that flipping that call is worth the risk of a failed (and thereby lost) challenge. The card system was one of several possible ways to quickly tell players whether they had the go-ahead to issue a challenge based on all those factors beyond the question of their confidence about having been wronged. In that sense, it doesn’t really violate the spirit of the new rule.
Communicating that information to players also isn’t something the league will have any luck preventing teams from doing. Banning cards from dugout railings will stop players from being able to do a quick check after a call and before a challenge, but the team can still relay signals to the batter via the base coaches that contain that information, before each pitch. The league won’t be able to prove that’s what any team is doing, and even if they could, they shouldn’t care.
So, despite an early setback, the wheels of the Milwaukee ABS Challenge Machine are in motion. Murphy described a “huge” session in his office Thursday to devise the plan, which goes much, much deeper than signal cards, anyway.
“We probably had one, two, three, four, five R&D guys in here with our baseball staff, saying, ‘How are we gonna present this?'” Murphy said. “‘Does this make sense?’ ‘What are you worried about from a strategy standpoint?’ ‘Here’s what we’re worried about from an implementation standpoint.'”
Murphy said the hope and expectation is for hitters (especially) to learn when challenges make sense or not themselves, anyway, rather than relying on any guidance from the bench. He also emphasized that, even if the league had allowed cards in the dugouts, the system was bound to become “way more complicated than [yes or no].”
“I don’t want to reveal our strategy,” Murphy said, but “The key is not to lose challenges early, but it’s also not to have challenges left at the end of the game.”Â
That’s the delicate balance all teams will have to learn to strike, and by forcing players to make their decisions about whether to challenge in the heat of battle on the field, the league is compelling them to become a high-functioning collaborative organism. Each player has to attune themselves to the variables that determine whether a given call is a good challenge or not, including things far beyond how strongly they feel that the umpire was wrong. Maxfield Lane and Owen Riley, who work together under the name Oyster Analytics, developed an app that models all those variables and the resulting changes within a game, with startling results.Â
Say Sal Frelick takes the first pitch of the game, a good two inches off the outside corner. Should he challenge? The answer is clear: No.
The odds that Frelick saw the ball well enough and will win his challenge are roughly 5:3, which isn’t bad. Given the count (0-0; a pitch either way matters, but isn’t decisive), the score (by definition, 0-0), the lack of anyone on base, and the long time left in the game, the Oyster model estimates that he’d need to be more like 9:1 certain to make the challenge worthwhile. Now, however, imagine that it’s the sixth inning, with the tying run on second base and two outs, in a 2-2 count. It’s a no-brainer again, but going the other way.
The yellow borders in each of these visuals show you the rough boundaries outside which a batter should challenge, given all the variables plugged in. That’s the break-even boundary. Here, even if Frelick were only about 20% sure he was right, he should issue the challenge.
Though these are two relatively extreme examples, they’re far from the most extreme. There are cases in which the same two pitches can demand even more obvious ‘yes’ or ‘no’ challenge decisions—and there are also plenty in which it’s something very close to a toss-up. Look what happens if there are runners on the corners with nobody out in a tie game, in the third inning, with a 2-1 count on the batter and just one challenge remaining, rather than two.
This is why the Brewers tried to start with a simplistic binary—not to effectively simulate their plan for the season ahead, but to begin the long learning process that will prove inevitable. Murphy has talked multiple times about not wanting to put too much into hitters’ heads as they go to the plate, and this system threatens to do that almost constantly. Each player will need to learn to intuit some of these variables’ impacts, so they don’t have to compute them as they make decisions in the box. Otherwise, they won’t be sufficiently focused on their actual task: hitting.
So much, too, will depend on the keystone skill of every hitter: vision.
“It still comes down to the players, which is what I love about the rule,” Murphy said. “It comes down to the players dictating it. They have to become really efficient, at knowing situations and—more importantly—at knowing what is a strike. First, what is a strike in the heart of the plate, and what is a strike when the ball is actually one [ball’s width] on, or what’s not a strike that’s actually one ball off, or way off?”
On Feb. 28, long before Murphy was ready to implement a system team-wide, the club lost both of their challenges in the first inning. In his typically wry form, Murphy said he wished he “could have tased some of them, when they do that,” but the prime directive of the early Cactus League games was for players to get a feel for the system by practicing. That plate vision comes from sheer acuity, but also from a good approach and a proper plan. Calibrating it, however, requires focused practice.
The Trajekt machine is a vital tool for that development and calibration. Though multiple Brewers hitters expressed some skepticism about it in their own experience, Murphy is insistent that Trajekt is the only way (other than live reps) to steadily gain a better sense of the edges of the zone, especially vertically. In addition to being programmed to deliver pitches that mimic those of any pitcher a batter might face, that machine allows the operator to set the specified strike zone of each batter, so players can practice discerning when the ball is (and isn’t) in their new, unique zone.
Murphy and the Brewers will give greater leeway to hitters who show facility with the zone both in games and in practice settings than to those who consistently lose challenges or seem to make poor swing decisions when they let the system factor into their at-bats too much. Trajekt can inform that, but so can data the Brewers have on players’ responses to the ball, which aren’t publicly available—or, in some cases, even well-understood by the players themselves. Using early hand movements as proxies, the team can estimate when hitters tend to see the ball and how early they make their swing decisions against certain pitcher types—based on handedness and arm slot, especially. That data can be used to coach players, but it can also be used as a scouting/grading tool—and to help them decide how free the player’s hand should be to issue challenges.
The front office, the coaching staff and the players themselves will be in constant communication throughout the season. The ABS system will be a subject in most pre-series hitter meetings and a constant source of feedback from the analytics team to Murphy and company. It will change the game too much not to work strenuously to do well with it, for a team that relies on winning on the margins. Doing that will mean a patient (but urgent), collaborative and comprehensive embrace of complexity.