Tampa Bay Rays Show Why Baseball’s Analytics Rules Need Exceptions

| 5 min read
I’ll preserve the supplied image and credit markup, then expand the article around its existing claims and examples without adding unsupported facts.

Image credit: IMAGN IMAGES via Reuters Connect

The lesson behind a low-scoring World Series

The Tampa Bay Rays’ most useful lesson may be that good analytics leaves room for judgment. Baseball trends can identify what tends to work, but a strategy that fits one player or situation can hold another back. A useful model describes a pattern; it doesn’t automatically tell a manager what to do with the next batter or pitcher.

That distinction matters after a World Series in which the Rays and Milwaukee Brewers combined for four home runs, the fewest in a Fall Classic since 1985. Some commentators treated the result as a repudiation of modern analysis and a return to small ball. That reads too much into one series. As Rob Mains has argued, the low home run total looks more like an outlier than evidence of a broad shift.

A short series can make a striking statistic feel like a verdict. But a small sample is a poor basis for declaring that an entire approach has failed, especially when the number of home runs doesn’t tell us by itself how teams built their offenses or made decisions. The result is interesting. It isn’t proof that the game has turned away from power.

The better explanation is that both teams found ways to get the most from the players they had. As Baseball Prospectus writer Jarrett Seidler put it about Milwaukee, “I don’t think it’s so much that they’re zagging against pulled fly balls—this is the team that spent a first-round draft pick last year on Andrew Fischer—but that they are perfectly willing to maximize skill sets which just aren’t going to include that.”

That’s a more useful reading than casting the series as analytics versus instinct. A team can value power and still accept that not every hitter will produce it in the same way. This is more significant than it looks: an organization’s strategy is tested less by whether it follows a trend than by whether it knows when that trend fits the person in front of it.

When the numbers need context

Modern analysis has helped establish useful principles: pulled fly balls can produce power, pitchers tend to fare worse the third time through a lineup, and changeups often work best against hitters from the opposite side. Those are tendencies, though, not commands. They describe what often happens across many situations, while a manager has to make a choice in one specific situation.

That gap matters. A general pattern can point a team toward a question worth asking, but it can’t answer every detail about an individual player’s strengths, weaknesses, or role. Treating a tendency as an instruction may create the appearance of rigor while obscuring whether the decision makes sense for the player.

Consider two Rays hitters. Johnny DeLuca’s .429 slugging percentage came alongside Statcast measures that suggest pulled fly balls aren’t his strength. That doesn’t make the approach wrong for the team; it makes it a poor fit for every player. Chandler Simpson, for example, may need a different path to productive offense.

The distinction isn’t a license to ignore evidence. It’s a reason to read the evidence at the right level: team trends can help shape a plan, while player-level information can show where that plan needs adjustment. A single approach may be easy to explain, but simplicity doesn’t make it equally effective for every hitter.

Decisions change with the game

Pitching decisions also depend on the moment. The third-time-through penalty is real, but Drew Rasmussen faced 71 batters across three starts against the Yankees and Guardians. If an ace is pitching well, the matchup trend alone may not settle the question of when to take him out.

A manager weighing that choice has to consider what the trend says alongside how the pitcher is performing and what the game demands. Pulling a starter because a general pattern says the next trip through the lineup is riskier can be defensible; so can giving a pitcher who’s executing the chance to continue. The statistic can frame the tradeoff. It can’t make the call.

The same flexibility applies to pitch selection. Changeups are commonly used against opposite-side hitters, but a pitch that gets same-side opponents out has a job to do. A conventional matchup guideline can help establish a starting point, yet the pitcher’s actual ability to use that pitch matters too.

And this is the part most people overlook: the point of analysis isn’t to make every decision look identical. It’s to help distinguish a meaningful pattern from a rule that has been stretched beyond its evidence. Sometimes the best use of a trend is to question a habit; sometimes it’s to recognize why an exception makes sense.

What this means for teams

For a front office or coaching staff, the challenge is to keep evidence and judgment in the same conversation. Metrics can show where outcomes tend to improve, and player information can show whether a particular change is likely to suit the person asked to make it. Neither perspective is complete on its own.

If you’re working in this space, the practical question is not simply whether a trend is supported. It’s whether the trend describes this player, this matchup, and this point in the game well enough to guide a decision. That takes more care than copying a successful pattern from one context into another.

The Rays’ example points toward a practical standard: start with the evidence, then account for the player and the game in front of you. Analytics is most useful when it sharpens decisions rather than turning tendencies into rules. A reminder teams may need more often.

Source: Daniel R. Epstein · www.baseballprospectus.com