Experiment 2 — brazil-bench (24 cells) · language=typescript, model=sonnet, tooling=beads

Run	Rep	Status	code_quality	defect_rate	idiomatic	maintainability	test_coverage	token_efficiency	cost ($)	duration (s)	tokens	turns	Files	LOC
5	rep 1	completed	0.733	1.000	0.670	0.578	0.000	1.000	$0.9246	361.9s	1,556,688	43	8	1073

Run

Rep

Status

code_quality

defect_rate

idiomatic

maintainability

test_coverage

token_efficiency

cost ($)

duration (s)

tokens

turns

Files

LOC

rep 1

completed

0.733

1.000

0.670

0.578

0.000

1.000

$0.9246

361.9s

1,556,688

1073

Code review

Per-replicate file listing extracted from the archived workspace. For semantic analysis (architecture, interfaces, control flow), invoke the run-summary skill via retort evaluate <run_dir>.

rep 1 — 8 file(s), 1073 loc (2 test file(s), 255 test loc)

src/__tests__/normalize.test.ts (36 loc, test)

src/__tests__/queries.test.ts (219 loc, test)

src/dataLoader.ts (141 loc)

function parseDate, function toInt, function toFloat, function readCsv, function loadBrasileirao, function loadCopaDoBrasil, function loadLibertadores, function loadBrFootball, function loadHistoricalBrasileirao, function loadFifaPlayers, function loadAllData

src/index.ts (280 loc)

src/normalize.ts (100 loc)

function normalizeTeamName, function teamsMatch, function teamMatchesQuery

src/queries.ts (233 loc)

function searchMatches, function getHeadToHead, function getTeamStats, function getStandings, function searchPlayers, function getStatistics, function getBestTeams

src/types.ts (57 loc)

vitest.config.ts (7 loc)

Stack signature (sorted JSON): {"language": "typescript", "model": "sonnet", "tooling": "beads"}

language=typescript, model=sonnet, tooling=beads

Runs

Code review