What the Pitch
Was Doing

Statcast is not a sample. The rig sits in all thirty major-league parks, so every pitch in every game is measured the same way, by the same instrument. Here each installation is sized by its home staff's season strikeouts - the loudest thing a pitch can do. Seattle and the Bronx run big; Denver, where the thin air flattens every breaking ball, runs quietest in the league.

Every park, in a table

Ballpark	Club	K	4-seam mph	Spin rpm
American Family Field	Milwaukee, WI MIL	1,497	94.4	2,347
Angel Stadium	Anaheim, CA LAA	1,279	93.8	2,293
Busch Stadium	St. Louis, MO STL	1,316	93.5	2,276
Camden Yards	Baltimore, MD BAL	1,466	94.1	2,325
Chase Field	Phoenix, AZ ARI	1,402	93.9	2,298
Citi Field	Queens, NY NYM	1,420	94	2,309
Citizens Bank Park	Philadelphia, PA PHI	1,519	94.3	2,344
Comerica Park	Detroit, MI DET	1,488	94.2	2,329
Coors Field	Denver, CO COL	1,187	93.4	2,244
Dodger Stadium	Los Angeles, CA LAD	1,544	94.5	2,352
Fenway Park	Boston, MA BOS	1,389	93.7	2,287
Globe Life Field	Arlington, TX TEX	1,394	94	2,307
Great American Ball Park	Cincinnati, OH CIN	1,454	94.6	2,333
Guaranteed Rate Field	Chicago, IL CWS	1,298	93.5	2,270
Kauffman Stadium	Kansas City, MO KC	1,341	93.6	2,281
loanDepot Park	Miami, FL MIA	1,408	94.2	2,318
Minute Maid Park	Houston, TX HOU	1,476	94.3	2,336
Nationals Park	Washington, DC WSH	1,263	93.6	2,272
Oakland Coliseum	Oakland, CA ATH	1,224	93.3	2,258
Oracle Park	San Francisco, CA SF	1,387	93.8	2,296
Petco Park	San Diego, CA SD	1,498	94.4	2,340
PNC Park	Pittsburgh, PA PIT	1,372	94.1	2,301
Progressive Field	Cleveland, OH CLE	1,523	95.1	2,360
Rogers Centre	Toronto, ON TOR	1,359	93.7	2,288
T-Mobile Park	Seattle, WA SEA	1,608	94.9	2,371
Target Field	Minneapolis, MN MIN	1,461	94.7	2,338
Tropicana Field	St. Petersburg, FL TB	1,471	94.6	2,349
Truist Park	Atlanta, GA ATL	1,512	94.4	2,341
Wrigley Field	Chicago, IL CHC	1,431	94	2,312
Yankee Stadium	Bronx, NY NYY	1,533	94.8	2,356

Illustrative stand-in figures in the real Statcast shape. When the pitch-level export is ingested, strikeouts, average four-seam velocity and spin roll up per home staff (pitching team = home club in the top of the inning). Coordinates are curator-authored - see Methodology.

Rank the arsenals by Stuff+, the model's read on raw pitch quality from velocity, spin and movement alone - 100 is league average, every 10 points above it is roughly a tier nastier. This is pure process: what the pitch is doing, before a batter has anything to say.

Bars share one scale; the hairline marks the league-average 100. Whiff% is whiffs per swing on the pitcher's whole mix. Illustrative stand-in figures - the physical inputs come straight from the pitch feed; Stuff+ is Baseball Savant's model (see Methodology).

The tracking measures the pitch; the scoreboard measures the outcome. They do not always agree. Every pitcher below is placed by Stuff+ (how nasty, going right) against the xwOBA they allow (how much damage, better going up). The parade of triple-digit fastballs clusters to the right - but plenty of them sit low, giving up more than their stuff says they should. That distance is the whole story.

Illustrative stand-in figures. xwOBA (expected wOBA) rolls up straight from the pitch feed's batted-ball model; Stuff+ is the pitch-quality model - see Methodology.

Two trends the tracking caught in the act. The league keeps throwing harder: the average four-seamer is up 1.7 mph since 2015, a tenth or two every season, no plateau in sight. And the mix keeps bending away from the heat that defines it - the four-seam recedes year over year while the sweeper, not even a named pitch a decade ago, floods in.

Every season, in a table

Year	4-seam mph	4-seam %	Sinker %	Sweeper %
2015	92.6	36.2	21.5	-
2016	92.8	35.9	21.0	-
2017	93	35.6	20.4	-
2018	93.2	35.2	19.9	-
2019	93.4	34.9	19.3	-
2020	93.5	34.5	18.8	-
2021	93.7	34.1	18.3	-
2022	93.9	33.6	17.9	-
2023	94	33.2	17.4	4.6
2024	94.2	32.7	16.9	6.1
2025	94.3	32.5	16.6	6.8

The charted series only; the full eight-pitch mix by season sits in the table under The Whole Repertoire. The sweeper is blank before 2023 because Statcast bucketed it inside the slider until then. Illustrative stand-in figures in the real Statcast shape.

Averages flatter the change; the distribution shows it. In 2015 the league's four-seamers piled up at 91 and 92, and 15.4 percent cleared 95. In 2025 the whole curve has slid right and 38.4 percent sit at 95 or better. What used to be a scouting-report exclamation point is now the third pitch of an average inning.

Each pair of columns is one 1-mph bin; heights are the share of that season's four-seamers, so both series sum to 100. The hairline marks 95 mph. Illustrative stand-in distribution shaped to the real drift - the real one bins straight off the pitch feed's release_speed column (see Methodology).

Every bin, in a table

Bin, mph	2015 share %	2025 share %	Change
87-88	2.3	0.2	-2.1
88-89	3.8	0.6	-3.2
89-90	7.2	1.8	-5.4
90-91	11.4	4.2	-7.2
91-92	15.1	8.0	-7.1
92-93	16.8	12.6	-4.2
93-94	15.7	16.4	+0.7
94-95	12.2	17.5	+5.3
95-96	8.0	15.5	+7.5
96-97	4.3	11.2	+6.9
97-98	2.0	6.7	+4.7
98-99	0.8	3.3	+2.5
99-100	0.2	1.3	+1.1
100-101	0.1	0.4	+0.3

The book still opens on a fastball: nearly six pitches in ten are four-seamers, sinkers or cutters. But the book is being rewritten one percentage point a season. Every fastball's share is drifting down, every bender's is drifting up, and the sweeper went from unclassified to 6.8 percent of all pitches in three seasons. Eight panels, one scale - the flat lines are the story as much as the moving ones.

All eight panels share one 0-40% scale, so a flat low line really is a flat low line; the faint hairline in each panel is 20%. The end figure is the 2025 share. The sweeper panel starts in 2023, when Statcast first split it from the slider. Illustrative stand-in figures (see Methodology).

Every season's full mix, in a table

Year	FF	SI	FC	SL	ST	CU	CH	FS
2015	36.2	21.5	5.9	14.8	-	10.1	10.6	0.9
2016	35.9	21.0	6.3	15.2	-	10.0	10.6	1.0
2017	35.6	20.4	6.7	15.6	-	9.9	10.7	1.1
2018	35.2	19.9	7.0	16.0	-	10.0	10.7	1.2
2019	34.9	19.3	7.3	16.3	-	10.2	10.6	1.4
2020	34.5	18.8	7.6	16.6	-	10.3	10.6	1.6
2021	34.1	18.3	7.9	16.8	-	10.4	10.7	1.8
2022	33.6	17.9	8.3	16.9	-	10.5	10.9	1.9
2023	33.2	17.4	8.6	12.9	4.6	9.8	11.0	2.5
2024	32.7	16.9	8.9	12.2	6.1	9.4	11.1	2.7
2025	32.5	16.6	9.0	11.9	6.8	9.2	11.2	2.8

Usage share of all pitches, percent. Codes are Statcast's own pitch_type classification; the sweeper (ST) is blank before 2023 because it lived inside the slider until then.

Radar guns flatter velocity; the cameras also see shape. Every pitch here is placed by how far it moves versus a ball thrown without spin - ride up, run to the arm side, sweep away. The four-seam lives alone at the top because its backspin fights gravity; the sweeper lives alone at the far left because sixteen inches of sideways is the whole trick. A repertoire is really a set of distances between these dots.

Every pitch's movement, in a table

Pitch	Family	Horiz, in	Vert, in	Velo, mph	Spin, rpm
Four-Seam FF	fastball	+7.9	+15.8	94.2	2,320
Sinker SI	fastball	+15.1	+6.9	93.5	2,178
Slider SL	breaking	-5.9	+1.7	85.4	2,431
Changeup CH	offspeed	+13.2	+4.6	85.1	1,782
Curveball CU	breaking	-9.3	-8.7	79.6	2,564
Cutter FC	fastball	-1.8	+8.4	89.1	2,384
Sweeper ST	breaking	-14.6	+2.8	82.1	2,648
Splitter FS	offspeed	+9.6	+1.2	86.4	1,497

Illustrative stand-in figures. The real values are the per-type means of the pitch feed's pfx_x / pfx_z movement columns, converted to inches and mirrored to arm-side-positive for left-handers (see Methodology).

Put the two eras side by side and every line tilts the same way: batters swing through every single pitch type more than they did in 2015. Harder fastballs buy a couple points; the benders buy more; the splitter, thrown by almost nobody, misses nearly two bats in five. This is the arms race scored from the hitter's side of the plate.

Every slope, in a table

Pitch	From	Then %	Now %	Change
Splitter FS	2015	33.6	37.9	+4.3
Sweeper ST	2023	34.8	36.2	+1.4
Slider SL	2015	31.2	34.6	+3.4
Curveball CU	2015	28.9	31.7	+2.8
Changeup CH	2015	28.1	30.8	+2.7
Cutter FC	2015	22.4	25.7	+3.3
Four-Seam FF	2015	18.9	22.1	+3.2
Sinker SI	2015	13.5	15.3	+1.8

Illustrative stand-in figures. The real rates divide swinging strikes by all swings, per pitch type per season, straight off the pitch feed's description column (see Methodology).

The closing ledger: eight pitches, one tradeoff, every number in one place. Sorted fastest to slowest, the pattern reads almost like a law: the hard pitches (amber) are the ones batters put in play, and the slow benders - sweeper, splitter, slider - are where the swings and misses (slate) live. Velocity buys strikes; shape buys whiffs. Every good arsenal is a bet placed somewhere on this line.

Velocity bars share one scale (floored at 74 mph so the spread reads); whiff bars share another. xwOBA is expected wOBA against - lower is nastier. Illustrative stand-in figures pulled to the real per-pitch shape; every value here aggregates directly from the pitch feed (see Methodology).