Skip to content

Player

Shelby Miller

MILPBats RThrows RActive

Season 2025 pitches thrown. Sequences are 2-, 3-, and 4-pitch windows. Count filters apply to 2-pitch type rows only. Location sequences use a 4-region plate map. Date windows stay on Arsenal and Hitting. Tunnel Score on Arsenal.

2-pitch windows

533

CSW%

26.6%

Whiff%

28.8%

Count+

49.5%

Sequence Entropy

50

Sequence Lift

5.6%

Sequence Effectiveness

+5.1%

Location sequence
Type plus a 4-region plate map (up-in, up-away, down-in, down-away). Pairs under 20 sequences are hidden. Count filters do not apply here.
Transition flow
Two-pitch sequences. Link width is how often the next type follows.
Four-seam FastballSplitterSweeperUnknownSliderFour-seam FastballSplitterSweeperUnknownSliderCutter
Transition lift
P(next | first) minus the unconditional next-pitch mix. Positive means that follow-up is used more than usual after the first type.
Four-seam FastballSplitterSweeperUnknownSliderFour-seam FastballSplitterSweeperUnknownSliderCutter
Sequence tree from Splitter
Branches are 2-pitch, then 3- and 4-pitch n-grams when the sample clears the display floor. Count filters hide 3/4-grams.
SplitterFour-seam FastballFour-seam FastballSplitterFour-seam FastballSplitterFour-seam FastballSplitterSweeperFour-seam FastballSplitterFour-seam FastballFour-seam FastballSplitterSplitterSweeper
Whiff% after Splitter
Last-pitch rates on the follow-up. Sample floors hide tiny pairs.
Pooled model after Splitter
Observed mix versus a league-pooled next-pitch model given last pitch and count. Smoothing for rare contexts, not a sit or pick.