About this report
The first P/PA report was some follow up research on a problem that High Heat Stats released in 2012. It asked a simple historical question: did MLB batting pitches per plate appearance keep rising after 2012?
This report asks the follow-up question that matters more to baseball people:
Does seeing more pitches actually help teams hit better or win more?
The answer is not at all clean. It's really messy, in fact it's so messy that I'm not really finding a lot of correlations. From 1988 through 2025, higher team P/PA is strongly associated with strikeout rate, somewhat associated with walk rate, and only weakly associated with OPS, runs per game, win percentage, and run differential.
That does not mean patience is useless. It means P/PA is a crude way to measure patience. In future reports, we will dig deeper into counts faced and results to try to shed some more light into what kind of patience is useful in baseball. And we will also look at this problem through the lens of modern day pitching.
On correlations (please don't ban ice cream)
A correlation is a measure that describes the size and direction of the relationship between things. If two things are correlated, they move together. If they are positively correlated, when one increases the other will too. And if they are negatively correlated, when one increases the other will decrease.
On average, tall people weigh more than shorter people. So you could say that height and weight are positively correlated. When the temperature rises, winter coat sales drop. Temperature and winter coat sales are negatively correlated. Shoe size and income are totally unrelated. They have no correlation.
Pearson correlations are expressed between -1 and 1. -1 would be strongly negatively correlated; there is a strong relationship between these things and when one rises the other drops. +1 would be strongly positively correlated. And everything around 0 is really weak; there's no real relationship there.
It's important to remember that just because there is a relationship between two things it doesn't necessarily mean that one causes the other. There is an old adage 'correlation does not equal causation'. The classic example of this is the correlation between ice cream sales and drowning. Ice cream does not cause drowning. Instead more people do things in the water and/or eat ice cream when it's warm outside.
The problem with P/PA
Pitches per plate appearance sounds like it should measure patience. Sometimes it does. A team that sees more pitches may be controlling the strike zone, taking bad pitches, walking, and waiting for good pitches their players can drive.
But a team can also see more pitches because their players fall behind, foul off two strike pitches, swing through good fastballs, and strike out too much. That's not very good hitting. It could be good strategy. In the postseason, it's a totally legitimate strategy to try to force the other manager to go to the bullpen earlier and squeaking out a few more pitches per plate appearance is a good way to do that.
The same P/PA number can contain:
- selectivity
- passivity
- strike-zone control
- swing-and-miss
- foul-ball survival
- pitcher quality
- reliever usage
- era effects
- umpiring changes
- team construction
Pitches per plate appearance is interesting, but you can't rely upon it too much.
What the data says
The strongest relationship in the full dataset is between P/PA and strikeout rate. Intuitively that makes sense. Hitting a baseball is incredibly hard and pitchers always have a major advantage. The more pitches per plate appearance over a 162 game season and the more two strike counts and so pitchers have more opportunities to get strikeouts. The hitter always has to solve the pitcher to get on base, and modern baseball analytics has made that problem even nastier.
The relationship between P/PA and walk rate is positive too, but weaker than the relationship with strikeout rate. This wasn't as intuitive; when I started working on this I started with the thesis that strikeouts would be more strongly correlated than walks, but the walk rate wouldn't be too far behind. It is a positive correlation and it's the second strongest correlation I found with actual data. But it's significantly weaker than the relationship between P/PA and strikeout rate. I guess that's why we run numbers and don't just blog about things we think are true.
That is the central tension of this report:
More pitches are associated with more walks, but they are even more strongly associated with more strikeouts.
The correlations with OPS, runs per game, win percentage, and run differential are much weaker. That does not prove that taking pitches is bad. It does suggest that seeing more pitches is not the same thing as being a better offense. I still don't totally understand this and don't have enough data to figure out the why but that will come (I hope).
Why this matters
Baseball has a lot of old sayings that are partly true. "Be patient at the plate" is one of them. When I was a young hitter, my coaches would always tell me to be patient, though I had a habit of trying to drive the ball hard instead of making good contact so that was good advice for me. But at the highest levels of baseball, we see signs that patience is a virtue but passivity is a curse.
Patience can be valuable. Taking a walk is valuable. Making a pitcher work can be valuable. Refusing to chase bad pitches is very valuable.
But if patience crosses the line into passivity, the value disappears quickly. A hitter who takes hittable pitches and then strikes out lost the battle. He failed to solve the pitcher. A team that drives up pitch counts but does not hit, score, or win has not discovered anything new. It has just discovered a new way to make sure that fans get the best value out of their ticket. Which I mean, that's great and all but it sounds like something Bill Veeck would have been into.
That is why this report goes into the data.
Moneyball, but carefully
It would be easy to turn this into an anti-analytics rant. That would be fun because analytics have really changed the game over the coure of my life, but it would also be lazy. Besides Moneyball is one of my favourite baseball movies ever so I can't bite the hand that entertains me. Or something like that.
The real Sabre-based insight was not simply "take more pitches." The better version was that on-base percentage was undervalued, and teams could exploit market inefficiencies by valuing skills that other teams misunderstood. That is different from saying P/PA itself is magic because it's clearly not, it's just a data point. The real magic seems to come from elsewhere and I'm not sure where yet.
This report is not arguing against analytics. It is arguing for better analytics. A single number can be useful, but only if we understand what it can and cannot measure. P/PA does not cleanly separate good patience from bad passivity. Hopefully we'll find more data on what separates good patience from passivity but I suspect that will require deep dives into players.
Why 2020 has a toggle
The 2020 season was really really weird. And not just because of Covid. It had a 60-game schedule, different strategic incentives, and an expanded playoff format. Some team-season totals also behave strangely because raw wins and losses come from a shortened season.
This report includes 2020 by default because it happened and belongs in the historical record. The toggle lets you remove it when you want a more normal season-to-season comparison. In this dataset, removing 2020 does not change the main conclusion very much. That is useful. The central finding does not depend on one strange season. And it makes me feel a lot better about my report than my payroll versus wins simulation which was dramatically different with 2020 included.
How to read the correlation table
The correlation table uses Pearson correlation.
A value near 1.000 means two values tend to rise together. A value near -1.000 means one tends to rise as the other falls. A value near 0.000 means the relationship is weak or not very linear.
This report compares team P/PA with:
- strikeout rate
- walk rate
- batting average
- on-base percentage
- slugging percentage
- OPS
- runs per game
- wins
- win percentage
- run differential
- playoff result
The table is not a proof of cause because correlation is never causation. It does not say P/PA causes strikeouts or wins. It shows how team P/PA moves with other team-level outcomes in this dataset. So it shows relationships between the numbers but does not imply that one causes the other. It's like banning ice cream to prevent drownings.
How to read the team table
The full team-season table contains every team-season from 1988 through 2025.
You can search by team, season, league, division, or playoff result. You can also filter by era, league, playoff outcome, and P/PA group.
The P/PA group filter is especially useful:
- Top 10 by season shows the ten highest P/PA teams in each season.
- Bottom 10 by season shows the ten lowest P/PA teams in each season.
- Above league average shows teams above their season's league-average P/PA.
- Below league average shows teams below their season's league-average P/PA.
This keeps comparisons fairer across eras. A 3.80 P/PA season meant something different in 1991 than it did in 2025.
Methodology
This report uses Baseball-Reference team batting pitch-count pages and standard batting pages from 1988 through 2025. Those rows were normalized into team-season records. Then the data was joined to Lahman team records to add wins, losses, league, division, run differential, and playoff results.
P/PA uses Baseball-Reference pitch-data plate appearances. Strikeout rate, walk rate, home run rate, and offensive rate statistics use standard batting plate appearances.
That distinction matters because Baseball-Reference pitch-data plate appearances and standard batting plate appearances can differ, especially in older seasons. The report preserves both instead of pretending baseball data is tidier than it is. This is a really neat part of baseball; the data is structured but it's not tidy. It's as tidy as it is because of amazing volunteers who spend their days copying box scores.
What this report does not prove
This report does not prove that taking pitches causes strikeouts. It does not prove that teams should swing early. It does not prove that patience is bad.
It also does not know whether a team was taking good pitches, taking bad pitches, fouling off tough pitches, or getting carved up by someone throwing 100 mph with a breaking ball that should require a permit. It doesn't have enough data incorporated to make any claims like that.
The report only says this:
At the team-season level, P/PA is strongly tied to strikeout rate, somewhat tied to walk rate, and weakly tied to broad offensive success and winning.
That is enough to make the old patience slogan a lot more complicated.
What did we learn?
Seeing more pitches is not the same thing as hitting better. That is the main lesson.
Good patience still matters. Zone control matters. Walks matter. Forcing a pitcher into bad counts matters. But P/PA alone cannot tell us whether a team is being selective or passive. A good plate appearance is not just long. A good plate appearance produces something useful.
That could be a walk. It could be a hard-hit ball. It could be a sacrifice fly. It could be forcing a pitcher into a worse situation for the next hitter. In other cases, it might be about forcing a manager to dig into their bullpen sooner than they want.
But length by itself is not virtue. Baseball, once again, refuses to be reduced to a rule of thumb. I love this game so much. You might even say I'm down with P/PA (yeah you know me).