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

Run	Rep	Status	code_quality	defect_rate	idiomatic	maintainability	test_coverage	token_efficiency	cost ($)	duration (s)	tokens	turns	Files	LOC
19	rep 1	completed	1.000	1.000	0.480	0.707	0.000	0.016	$1.2560	218.0s	965,345	24	9	814

src/main/java/com/example/soccer/CsvReader.java (81 loc)

class CsvReader, method countUnescapedQuotes, method parseLine

src/main/java/com/example/soccer/DataStore.java (132 loc)

class DataStore, method matches, method players, method load, method loadDefault, method loadBrasileirao, method loadCup, method loadLibertadores, method loadExtended, method loadHistorical, method loadFifa

src/main/java/com/example/soccer/McpServer.java (185 loc)

class McpServer, method handle, method formatMatches, method formatPlayers, method helpText, method main

src/main/java/com/example/soccer/QueryService.java (152 loc)

class QueryService, method matchesForTeam, method matchesBetween, method matchesByCompetition, method matchesBySeason, method matchesByDateRange, method teamStats, method standings, method playersByName, method playersByNationality, method playersByClub, method topPlayers, method averageGoalsPerMatch, method homeWinRate, method biggestWins

src/main/java/com/example/soccer/TeamNames.java (24 loc)

class TeamNames, method normalize, method matches

src/main/java/com/example/soccer/model/Match.java (36 loc)

class Match, method winner, method totalGoals

src/main/java/com/example/soccer/model/Player.java (27 loc)

class Player

src/main/java/com/example/soccer/model/TeamStats.java (18 loc)

class TeamStats, method points, method goalDifference, method winRate

src/test/java/com/example/soccer/DataStoreTest.java (159 loc, test)

class DataStoreTest

language=java, model=opus, tooling=none

Runs

Code review