PitcherStat Methodology
Three proprietary 0–100 strength scores turn a wide set of MLB performance signals into a fast, consistent read on starting pitchers, hitters and teams. P-Score, H-Score and T-Score are designed to reduce dozens of statistics into context you can understand at a glance — without pretending that one number tells the whole story. P-Score and T-Score share the current composite calibration across leaderboards, profiles and matchups. H-Score uses the established hitter-strength model on the same 0–100 scale.
ONE NUMBER. A MUCH BIGGER PICTURE.
- 01MLB logsBox scores, game logs and season totals.
- 02Inputs12 pitcher · 11 hitter · 10 team signals.
- 03PillarsReadable groups you see on every profile.
- 040–100One strength number, same scale site-wide.
Baseball produces more information than any single box score can comfortably explain. ERA can miss command. Batting average can miss power and patience. A win-loss record can hide how a team actually reached it.
PitcherStat Scores are built to compress that larger picture into three readable numbers.
Each score starts with a broad set of performance signals, organizes them into understandable pillars and places the result on a common 0–100 scale. The number is the shortcut. The pillar breakdown is where you see why.
The goal is not to replace baseball statistics. It is to make them faster to understand.
P-Score and T-Score recently tightened how signals are grouped into pillars and how qualified MLB peers are compared on the 0–100 scale. H-Score continues to use the same hitter-strength framework already on boards and profiles. We do not publish weights, transforms or internal calibration steps.
A P-Score of 82 does not mean an 82% chance to win. An H-Score of 82 is not a projection of tonight's box score. A T-Score of 82 does not guarantee a better record tomorrow. The scores summarize the quality of the performance already in the data. Matchup context, lineups, handedness, venue, rest and recent changes still matter — which is why PitcherStat shows those layers alongside the score instead of hiding them behind it.
TRANSPARENT ABOUT THE IDEA. PROPRIETARY ABOUT THE FORMULA.
We publish what each score is designed to measure, the major performance pillars behind it, how to interpret the scale and where the score should not be used. We do not publish the exact weighting system, normalization logic or internal transformations that produce the final number. Those are proprietary to PitcherStat. That balance is deliberate: readers should understand what a score means without turning the methodology page into a blueprint for reproducing it. When the composite models are recalibrated, we update the live scores and this page at a high level — not the internal recipe.
Score scale
Color follows the same quality engine as boards and profiles — not “red because the number is big”.
The labels make the scale easier to scan, while the underlying color treatment remains continuous. A 79 and an 80 are close numerically; an 80 and a 90 are meaningfully different. The visual system preserves those differences instead of forcing every score inside a band to look identical. Score and pillar colors on this page follow the same quality interpretation used on PitcherStat boards and profiles — green means strong relative to the qualified pool, not “big number equals red.”
P-Score — Starting Pitcher Strength
A starting pitcher is more than his ERA. P-Score is designed to capture the broader shape of the season: how effectively he prevents runs, how well he controls the strike zone, how often he misses bats, how consistently he gives his team competitive starts, how deep he works and whether the recent version of the pitcher is moving away from his season baseline. The result is a 0–100 strength score that makes different pitching profiles easier to compare without pretending they are identical.
A MULTI-LAYER PITCHING PROFILE
P-Score evaluates a broad collection of pitcher-level signals rather than leaning on one headline statistic. The exact formula is proprietary, but the model is organized around several understandable performance families. The current composite generation keeps season body-of-work primary while recent form stays inside the pillar layer rather than replacing the season read.
- Run preventionhow effectively runs and damaging contact are limited.
- Commandcontrol of walks, counts and the strike zone.
- Strikeout qualitythe ability to generate strikeouts and miss bats.
- Consistencywhether performance is repeatable from start to start rather than driven by a few extremes.
- Workloadhow much meaningful starting-pitcher volume the season contains.
- Recent formenough recency to recognize a changing pitcher without turning the score into a short-term hot-streak meter.
WHY NOT JUST ERA?
ERA answers an important question, but not every question. Two starters can own the same ERA with very different command, strikeout ability, workload and start-to-start stability. P-Score is designed to distinguish those profiles. It uses the broader body of work to show whether the headline result is supported by the rest of the pitching profile.
Illustrative example · not a real pitcher comparison
HOW TO USE IT
Use P-Score as a fast measure of season strength, then open the pitcher profile to understand the pillars behind it. On a matchup page, the difference between two P-Scores is useful context — not a prediction. Opponent quality, lineup handedness, rest, venue and current availability still belong in the game-level analysis.
QUALIFICATION
P-Score appears when a starting pitcher has accumulated enough MLB workload for the underlying profile to be treated as stable. Pitchers below that threshold can still have full PitcherStat profiles, game logs and traditional statistics without being forced into a score before the sample is ready. Current qualification threshold: 5 games started and 15 innings.
WHEN IT UPDATES
P-Score is refreshed after new completed-game data is incorporated into PitcherStat. Because it is a season-strength score, one start can move the number, but the score is intentionally more stable than a last-five-start statistic. Recompute runs on a fixed ingestion schedule so leaderboards and profiles update together.
What it does not mean
- Not a win probability.
- Not a prediction of the next start.
- Not a replacement for matchup context.
- Not a claim that two pitchers with the same score have the same strengths.
- Not a public reproduction of the proprietary formula.
Frequently asked
What is a good P-Score?
Is P-Score a prediction?
Why can two pitchers with similar ERA have different P-Scores?
Do you publish the exact P-Score formula?
H-Score — Hitter Strength
Hitters create value in different ways. Some dominate through power, some through contact and on-base skill, and some combine several strengths without leading the league in any one category. H-Score is designed to recognize those different paths. It organizes a wide set of offensive signals into clear pillars, then expresses the overall season on a 0–100 scale that can be read quickly without losing the shape of the hitter underneath it.
WHAT IT LOOKS AT
H-Score is built from dozens of hitter-level signals organized into readable pillars. The exact formula is proprietary. Qualified hitters are compared on one league-relative 0–100 scale across boards and profiles.
- Productionthe overall quality of the offensive results being created.
- Powerextra-base and home-run impact, supported by quality-of-contact signals where available.
- Contactthe ability to put hittable pitches in play and limit empty swings.
- Plate disciplinethe quality of swing decisions, walks and strike-zone control.
- Recent formenough recent performance to recognize meaningful movement without allowing one hot week to define a season.
WHY NOT BATTING AVERAGE?
Batting average is useful, but it treats very different offensive events too similarly and ignores walks entirely. H-Score is designed around the broader question of offensive strength. A hitter who combines power, on-base ability and disciplined plate appearances can produce a stronger offensive season than a player with a higher batting average but less total impact.
QUALIFICATION
H-Score appears when a hitter has accumulated enough MLB plate appearances for the profile to be evaluated responsibly. Players below the threshold still retain their full statistical profiles, splits and game logs; the absence of a score is a sample-size decision, not a missing player. Current qualification threshold: 100 plate appearances.
What it does not mean
- H-Score is primarily an offensive strength score.
- It is not a projection of the next plate appearance.
- It does not turn a tiny batter-versus-pitcher sample into a season-level signal.
- It does not mean two hitters with equal scores create offense in the same way.
- The exact model weights and internal transformations are proprietary.
Frequently asked
What is a good H-Score?
Does H-Score replace OPS?
Does H-Score include BvP?
Is the formula public?
T-Score — MLB Team Strength
The standings tell you who has won the games. T-Score asks a different question: what does the team's broader performance profile look like underneath the record? By combining multiple areas of team performance, it can distinguish clubs that arrive at similar records in very different ways — one through dominant pitching, another through offense, another through balanced strength across the roster.
WHAT IT LOOKS AT
T-Score is built from a broad set of team-level indicators. The pillar names below match the grades on team pages.
- Offensethe club's ability to create runs and sustained offensive pressure.
- Pitchingstarting staff and bullpen run prevention as one pitching picture.
- Defensehow well the club turns balls in play into outs.
- Resultswins, losses and run differential, including an expected-win view.
- Formrecent results without letting a short hot streak rewrite the season.
T-SCORE VERSUS STANDINGS
Standings and T-Score are intentionally shown side by side because they answer different questions. The standings are definitive: they determine who is ahead. T-Score is analytical: it describes the strength of the performance behind that record. A team can therefore rank higher in T-Score than in the standings, or lower. That is not a contradiction — it is the reason both views exist.
What it does not mean
- T-Score does not award playoff spots.
- It is not a simulation of the remaining schedule.
- It does not guarantee that the higher-rated club wins the next game.
- It is a season-strength view, not a substitute for the standings.
- Exact weighting and normalization remain proprietary.