The pitch, in your hand.

What it is

The part a chart cannot hold.

A dashboard can tell you what the ball did. It cannot always tell you where the thumb sat, which seam caught the pad, or why one pitcher swears by a hold another pitcher cannot make work.

That is the gap. Pitch Atlas treats the grip as the specimen and the source label as part of the interface.

Where it came from

It started as one hand on one ball.

The first grips filed here are photographs of the builder's own hand. Austin H. worked a four-pitch mix in games: four-seam, two-seam, a three-finger changeup, and a 12-6 curve, with a splitter that rode along for certain spots. There is no circle change in the set. His hands were too small to form one, so the file says that instead of faking a textbook page.

That account seeded the whole source model. It is filed as a pitcher's own words, not tracked data, because that is what it is. And it surfaced the gap the atlas now works: where the pads sat, which seam the fingers rode, why one pitch stayed in the pocket and another never fit. That knowledge lived in exactly one place, the hand that learned it, and nobody was writing it down.

See the grip library →

Useful imperfection

The rough edge is not a style trick.

Pitch Atlas has a wabi-sabi streak when it leaves the record imperfect on purpose: a missing grip stays missing, a community note stays a field note, and a schematic does not call itself measured if the constants are not pinned.

That is good only when it protects trust. The useful roughness is source honesty, hand texture, and visible limits. Fake patina would be worse than a plain page.

Why it exists

A grip dies with the arm.

A grip is oral tradition. A coach reshapes a kid's fingers in a bullpen. A teammate shows a seam trick between starts. A pitcher feels his way to a hold nobody taught him. Almost none of it gets written down with a name attached, and the measurement era did not fix that. Cameras and radar now record what the ball does after release down to the inch. The hand before release stayed where it always was: in dugouts, in scattered clips, in memory.

The lost-pitches wing shows the price. Holds that won real games survive as fragments because nobody filed them while the arm was still around to ask. Pitch Atlas exists to file the hold while it can still be asked about. Sourced, because memory earns trust through provenance. Uncorrected, because the variant a textbook would flag is sometimes the one that fit a real hand.

Nobody built this before because nobody had to own it. Coaching sells the next adjustment. Media sells what the ball just did. Trackers sell the measurement. An archive of sourced grips pays off years later, when someone wants to know how a pitch was actually held. Preservation is the job nobody claims until the thing is already gone.

Walk the lost-pitches wing →