Experiment 2 — brazil-bench (24 cells) · language=typescript, model=sonnet, tooling=none

Run	Rep	Status	code_quality	defect_rate	idiomatic	maintainability	test_coverage	token_efficiency	cost ($)	duration (s)	tokens	turns	Files	LOC
13	rep 1	completed	0.733	1.000	0.680	0.553	0.961	1.000	$0.7092	274.5s	797,512	24	5	1495

src/dataLoader.ts (195 loc)

function readCsv, function parseGoals, function normalizeDate, function normalizeTeamName, function teamMatches, function loadBrasileiraoMatches, function loadCupMatches, function loadLibertadoresMatches, function loadBRFootballMatches, function loadHistoricalMatches, function loadFifaPlayers, function getAllMatches, function clearCache

src/index.ts (365 loc)

function formatMatch, function formatPlayer, function formatStats, function main

src/queries.test.ts (409 loc, test)

src/queries.ts (425 loc)

function searchMatches, function getTeamStats, function getHeadToHead, function getStandings, function getOrCreate, function getBiggestWins, function getCompetitionStats, function searchPlayers, function getBestTeams, function getOrCreate, function getExtendedStats

src/types.ts (101 loc)

language=typescript, model=sonnet, tooling=none

Runs

Code review