---
name: interpreting-advanced-metrics
description: Use when interpreting or calculating college baseball wOBA, wRC+, FIP, BABIP, ISO, strikeout rates or walk rates, especially when source inputs, league constants or sample coverage are incomplete.
---

# Advanced metrics with explicit inputs

A computed rate is a description of observed events, not a validated talent forecast. Fetch the requested player/team observations, preserve their source, season and cutoff, then check required inputs before calculating.

## Calculation path

Use the installed `baseball-analytics` skill's `scripts/saber.py`; read its `--help` and input contract rather than guessing arguments. For NCAA calculations needing league constants, pass published totals for the requested Division I season with `--totals`. Do not silently use the helper's MLB `--season` defaults for college players. If the helper or totals are unavailable, name what cannot be computed.

The college-total workflow rescales its weights to the supplied run environment, but its relative event values still originate from MLB. Label that methodological limitation; do not call the result an official NCAA metric or a college-trained model.

| Metric | Required context |
|---|---|
| AVG, OBP, SLG, ISO | Exact counting-stat inputs and denominator; ISO is SLG minus AVG |
| BABIP | Hits, home runs, at-bats, strikeouts and sacrifice flies |
| Hitter K% / BB% | Strikeouts / walks and plate appearances |
| Pitcher K% / BB% | Strikeouts / walks and batters faced, not innings |
| FIP | Home runs, walks, HBP, strikeouts, innings and the correct league constant |
| wOBA | Event counts and documented weights for the specified scale |
| wRC+, ERA-, FIP- | Defined league, season and park adjustments; no guessed park factor |

Baseball innings notation is not decimal arithmetic: .1 and .2 denote one and two outs. Use the helper's documented conversion. Missing HBP, sacrifice flies, batters faced or league totals are not zero. A zero denominator is unavailable, not a rate of zero.

## Interpretation

Show the formula or named method, sample size, constants source, and any adjustment. Compare like seasons, leagues, units and selection populations. A list of searched players is not a national leaderboard. Distinguish a provider's statistic from a locally computed statistic.

Treat ERA/FIP divergence and unusual BABIP as questions to investigate, not proof of defense, luck, future regression or hidden talent. No calendar week makes every metric reliable; uncertainty depends on events observed, role, opponents and measurement quality. Do not assign universal All-American or elite-command labels from unsupported cutoffs.

## Deliver and verify

Return supported metrics, missing inputs and a restrained scouting interpretation. Explain whether a metric is descriptive or supplied by an explicitly identified forecast model. Never replace a requested unavailable metric with a different one without labeling the substitution and explaining the limitation.

Method references: [FanGraphs pitching definitions](https://library.fangraphs.com/pitching/complete-list-pitching/) and [park/league-adjusted pitching metrics](https://library.fangraphs.com/pitching/era-fip-xfip/). Use the baseball plugin's source-backed tests for executable formula checks; this interpretation skill has no independent predictive calibration.