https://ssp3nc3r.github.io/tags/eccentricity/ Recent content in Eccentricity on Dr. Scott Spencer Hugo en Tue, 30 Jun 2026 00:00:00 +0000 https://ssp3nc3r.github.io/post/framing-the-eccentrics/ Tue, 30 Jun 2026 00:00:00 +0000 https://ssp3nc3r.github.io/post/framing-the-eccentrics/ <!-- DRAFT (revised to the reference-width / structure-B fit, 1,912,808 pitches 2021–2026): eccentricity is SUPER-LINEAR, p≈2.14 (90% CI [1.5,2.8]), λ≈0.16; max edge shift ~3.9pp. The model also carries a stance-driven top/bottom edge asymmetry (τ_t/τ_b) that had to be modeled for p to read true; it beats the old symmetric fit by ~+60 ELPD (LOO). Narrative flipped from "saturating→remembered line" to "super-linear→reading fast peripheral motion". --> <p>A called strike zone, drawn from enough pitches, is not a crisp rectangle. Its edges are soft — there is a band, an inch or two wide, where a taken pitch is sometimes a ball and sometimes a strike — and the softness is not the same on every side. The boundary is tightest in the upper-inside corner and loosest down and away. A descriptive fit — a binned heatmap, a spline, a boosted-tree model — renders the zone as a lopsided, rounded blob, but much of that shape reflects <em>where pitches are thrown</em> rather than how the umpire calls, and it exaggerates. The truer asymmetry is quieter: not the blob’s outline but the width of the <em>edge</em> itself — sharp in some places, soft in others — which those pictures don’t isolate. More than one perceptual force shapes it, and this paper follows the one with the cleanest signature, the one that tracks something specific: <strong>where the umpire is looking.</strong></p>