The MLB Player Index

Click any column header to sort. Click a row (or search a unique name) to spotlight a pitcher. Default sort: ERA ascending.

Percentile is relative to all 216 pitchers. For ERA/WHIP/BB%/HR/9, lower is better so the percentile is inverted (100th = lowest). Adding a comparison pitcher overlays both on the stuff radar, timeline, and arsenal table.

What the pitcher's completed 2026 is projected to look like. Under the hood the ERA projection blends our full rule-based engine (Marcel + xFIP/xERA/SIERA + BABIP regression + Statcast iz_whiff/chase + velocity trend + age curve + park + OAA + injury) with a gradient-boosted ML model (trained on 1.4M Statcast pitches, validated R² 16.0% on 2025) and the public ZiPS + Steamer projections. Current partial 2026 stats shown below each tile for comparison.

Composite pitcher value score combining 5-year cumulative WAR projection, ERA-quality bonus (xFIP/xERA/SIERA skill baseline), age curve, injury risk, and youth bonus. Higher = more expected future value. Ceiling/floor are ±1.96σ bounds — wide gap = boom-or-bust, narrow = stable. Blue tick = league average FV. Surplus Value = projected production × $10M/WAR (current MLB FA market rate) minus remaining contract cost — positive = bargain, negative = overpay.

Current contract terms from Spotrac. Total value = the deal's maximum payout; on contracts with options it's labelled "(max)" and a separate Guaranteed figure shows the team's firm commitment (club / mutual / vesting option years excluded). AAV (avg annual value) = luxury-tax hit. 2026 payroll = this year's cash salary. $/FV = total value ÷ Future Value (dollars per projected future win). Lower = better value for the team.

Payroll salary for each remaining year of the pitcher's current contract. Status flags indicate contractual state each year: ARB = arbitration-eligible, FA = free-agent winter, Opt-Out = player option, Pre-Arb = league-minimum team control. Pitchers on multi-year extensions show the agreed annual breakdown; arb-eligible guys show this year only + future arbitration dates.

Pearson correlation. Drag the slider to change who counts as "successful." Too few pitchers (n < 10) won't render. When a pitcher is selected, their z-score vs the group appears as a translucent gold bar on the right chart.

Pitchers on X-axis sorted best ERA → worst. The line shows how the selected stat trends with pitcher quality.

Team defense = sum of qualified fielders' Outs Above Average (Baseball Savant / Statcast). OAA is the modern Statcast equivalent of "fielders' combined defensive value" — MLB Stats API doesn't expose true dWAR, so this is the closest available proxy. Positive OAA = above-average defense.

Select a pitcher to project their career ERA under a different defense. Projection = career ERA + slope × (hypothetical OAA − actual 2025 OAA). Each dot = one pitcher. Trendline shows if pitchers with better defenses behind them actually post lower ERAs.

Every starter plotted against their Statcast xERA (expected ERA based on quality of contact allowed). The diagonal is where ERA equals xERA — performing at skill level. Dots above the line have pitched unluckier than their contact suggests (primed to rebound). Dots below the line are getting better results than their skill suggests (primed to regress).

Hover a dot for the luck gap in ERA runs. Selected pitcher highlighted as a gold star.

Each axis is a percentile rank against the 216-pitcher pool. Outer edge = elite, center = bottom of the league. "Command" and "HR suppress" invert BB% and HR/9 so up is always better. Shape at a glance tells you the pitcher type: balanced ace, velocity-first strikeout guy, soft-tossing contact manager, or command-driven crafty vet.

Every pitch they throw at ≥2% usage, with Statcast-measured performance. Velo = avg speed. Whiff% = swing-and-miss rate (higher better). xwOBA = expected weighted on-base against this pitch (lower better; .320 ≈ league avg). Hard Hit% = % batted balls ≥ 95 mph (lower better). RV/100 = run value per 100 pitches (negative saves runs, positive costs them). Colored cells: green above avg, red below.

League-wide top-10 leaderboards for the new Round-3 metrics. Extension = ft in front of rubber (≥6.5 elite). Perceived velo = real velo adjusted for extension. Arm angle = degrees off horizontal (90 over-the-top, 0 sidearm). LOB% = strand rate (lg avg ≈73; >76 = lucky). RV/100 = run value per 100 pitches (positive = runs prevented vs avg). CSW% = called-strike + whiff rate (best per-pitch quality marker).

Five most stylistically similar MLB starters. Distance computed across K%, BB%, HR/9, FB mph, xBA, Hard Hit%, WHIP, K/BB (each z-scored so every dimension weighs equally). Age is intentionally excluded — so a 25-year-old rookie may match up with a 35-year-old ace if their stuff profile is similar. Useful for projecting "this rookie looks like young-Verlander" or "this vet is aging into a Kershaw template."

Side-by-side comparison. Pick a second pitcher from the similar-pitcher cards below the selected pitcher, or from the table (hold Shift + click).

How the projected ERA for next year was built. Each signal's contribution is shown as a bar — positive values push the projection up (worse), negative values push it down (better). The final number at the bottom is what the chart displays.

Single stat = raw values. Multiple stats = z-scores (σ from career avg) so different scales compare. Sign flipped for ERA/WHIP/HR/9/BB% so up = better for every line. Marker size scales with IP that year. Projection uses a Marcel-inspired model: (1) weighted avg of last 3 seasons (5/4/3), (2) regressed to league mean with weight based on IP vs. stat stabilization point (ERA stabilizes at ~130 IP, K% at ~60, etc. — short samples regress harder), (3) per-year age adjustment using per-stat peak-age curves. ERA projection is upgraded: uses xFIP (HR/FB-luck-neutralized FIP) as the skill baseline instead of raw ERA, then adds BABIP regression (70% toward league mean of .289, ~1 ERA run per 100 BABIP points) and a velocity-trend signal (lost velocity last year → higher projected ERA). Shaded band = ±1.96σ uncertainty widening √t. Hover projection points to see age + reliability %. Injury adjustment: if most recent completed season's IP is ≥20% below recent peak, model flags as likely injured and discounts projected IP/GS by 5-35% (ramps back to full over ~2 years) plus adds 0.05-0.25 ERA to the return year. Flagged years appear as "likely injured" on the chart.