Cross-sport / Calibration diagnostics

Brier skill score by sport and game phase

Published sport-phase rows retain Brier scores and Brier skill scores against climatology and market references.

Derived analysis11 published rows7 fieldsSource as of 2026-09-16

Data integrity

These MLB/soccer numbers are revision 2, computed on the segment-clean corpus (2026-09-16). Revision 1 values are withdrawn and kept in the regeneration receipt. Segmentation excluded 51,635 of the 78,986 MLB ticks (65.4 percent); the specifically identified label mismatches were a smaller 27,076 (34.3 percent).

brier_skill_scores

Read the full finding

Published measurements

Choose a measurement

MeasurementPopulationOrder

Whole-corpus estimates and their component phases are not comparable in one ranking. Showing International soccer by default.

Scored-row coverage

How much history has a phase label?

Source as of 2026-09-16

Whole-corpus estimates include scored rows that do not appear in the published phase breakdown.

International soccer

In published phases: 3,383
Outside the phase breakdown: 882
All scored rows: 4,265

Calculation and source fields

Coverage = sum of published phase counts / all scored rows. Outside = all scored rows - sum of phase counts.

sports.soccer_intl.grains.all.n
sports.soccer_intl.grains.0-15.n
sports.soccer_intl.grains.15-30.n
sports.soccer_intl.grains.30-45.n
sports.soccer_intl.grains.45-60.n
sports.soccer_intl.grains.60-75.n
sports.soccer_intl.grains.75-90+.n

These are scored in-game observations, not independent games. The source assigns each observation to at most one phase. Counts describe the full snapshot and do not change with the row search or measurement. Coverage does not measure forecast accuracy.

The snapshot does not publish observation start and end dates here. Its publication date is not an observation window. The counts alone do not explain why a row is outside the phase breakdown.

Inspect phase counts in the published source

6 matching rows; 6 contain model versus market brier skill score.

CSV includes 6 matching rows, all fields, and source context; missing measurements stay blank. Published source dates and row provenance are included when available; structured provenance uses JSON cells.

Measured rows: 6 / 60 unavailable
Median: -0.7478Middle ranked value
Middle 50%: -0.9479 to -0.665125th to 75th percentile
Observed range: -1.1707 to -0.4687Minimum to maximum

Summaries describe the filtered rows with equal row weight. Row counts are not independent game or event counts.

Measurement availability7 / 7 fields have values

Reference population: International soccer. 6 rows published.

Text search does not change this reference population.

Availability describes published values only. It does not establish data quality or independent game counts, and measurement windows can differ.

-1.1707Model versus Market Brier skill score0

Showing 6 of 6 matching numeric rows. Open a row to inspect all its fields. Rankings apply only to the published subset.

Published aggregate

Whole-corpus estimates

These estimates cover the full published sport corpus. They are shown separately from game phases and are not ranked with them. Row search filters the phases; these full-corpus estimates remain available for the selected population.

Brier skill score by sport and game phase: all published measurement fields
Entity	Model versus Market Brier skill score	Model versus Climatology Brier skill score	Market versus Climatology Brier skill score	Model Brier	Market Brier	Climatology Brier	Scored rows
International soccer \| allInternational soccer	-0.7039	-0.8569	-0.0898	0.3325	0.1952	0.1791	4,265

1-1 of 1

Page 1 / 1

01 / Interpretation

What this measures

Compare each Brier skill score only with its named reference and read the three Brier operands with the scored-row count.

02 / Calculation

bss_model_vs_clim = 1 - brier_model / brier_clim. bss_market_vs_clim = 1 - brier_market / brier_clim. bss_model_vs_market = 1 - brier_model / brier_market; n is the published support.

Use the table to inspect available fields and the source JSON to trace the formula inputs. CSV percentage values are stored as fractions.

03 / Population & limits

11 published sport-phase rows. Source artifact generated 2026-09-16.

Brier skill scores inherit the source's climatology and market references. Repeated in-game rows share outcomes, and a positive or negative score describes this published corpus rather than a future result.

04 / Source & method references

Published source JSON Published Brier skill-score method

References explain ingredients and prior methods. They do not validate this derived analysis or establish its originality.