> Data revision: baf7c2f4307d3a989a8fb9363df8000fda3d9310
> Source files:
- [results/claim-ledger.jsonl](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl)

# CL-SOL-LANG-REPLICATION-POSTHOC-EQUAL-TASK

**Source:** `results/claim-ledger.jsonl`  
**Ledger status:** VALID  
**Evidence status:** Reproducible from the published round-local CSV; equal-task formula is post-hoc because the original combined formula was not recorded.

Values below are transcribed from this claim-ledger entry. The site does not recalculate them. A status such as INTERIM, CONFOUNDED, or WITHDRAWN is not a validated advantage.

## Claim fields

- **metric:** `"posthoc_equal_task_pass_rate"`
- **numerator:** `null`
- **denominator:** `null`
- **unit:** `"unweighted mean of six task-specific pass fractions"`
- **analysis_rule:** `"Within each stack, compute the pass fraction for each of six task identities, then take their unweighted arithmetic mean. Show the raw combined pass totals separately; do not use this post-hoc sensitivity to execute the under-specified four-pass decision threshold."`
- **evidence_status:** `"Reproducible from the published round-local CSV; equal-task formula is post-hoc because the original combined formula was not recorded."`
- **claim_status:** `"VALID"`
- **publication_status:** `"STATUS_ONLY"`
- **reported_values:** `{"go": {"equal_task_rate_percent": 88.9, "raw_cells": 20, "raw_passes": 18, "task_fractions": {"r70-1-go": {"cells": 3, "passes": 2}, "r70-2-go": {"cells": 3, "passes": 3}, "r70-4-go-fe2": {"cells": 3, "passes": 2}, "r70-5-go": {"cells": 4, "passes": 4}, "r70-6-go": {"cells": 4, "passes": 4}, "r70-7-go": {"cells": 3, "passes": 3}}}, "rust": {"equal_task_rate_percent": 90.3, "raw_cells": 20, "raw_passes": 18, "task_fractions": {"r70-1-rust": {"cells": 3, "passes": 3}, "r70-2-rust": {"cells": 3, "passes": 3}, "r70-4-rust-fe2": {"cells": 3, "passes": 2}, "r70-5-rust": {"cells": 4, "passes": 4}, "r70-6-rust": {"cells": 4, "passes": 3}, "r70-7-rust": {"cells": 3, "passes": 3}}}, "ts-bun": {"equal_task_rate_percent": 83.3, "raw_cells": 18, "raw_passes": 15, "task_fractions": {"r70-1-ts-bun": {"cells": 3, "passes": 3}, "r70-2-ts-bun": {"cells": 3, "passes": 3}, "r70-4-ts-bun-fe2": {"cells": 3, "passes": 0}, "r70-5-ts-bun": {"cells": 3, "passes": 3}, "r70-6-ts-bun": {"cells": 3, "passes": 3}, "r70-7-ts-bun": {"cells": 3, "passes": 3}}}}`
- **interpretation:** `"Rates are Rust 90.3%, Go 88.9%, and TS-Bun 83.3%; the raw combined totals remain Rust 18/20, Go 18/20, and TS-Bun 15/18. This selected extension and reused cohort do not establish a general language ranking."`
- **evidence:** `[{"path": "rounds/lang-sol-replication/data/cells.csv", "selector": "all 58 rows; exact cell IDs; grouped by stack and six task identities"}, {"path": "rounds/lang-sol-replication/DECISION-RULE.md", "selector": "registered combined branch and disclosed post-hoc formula"}]`

## Population references

Round: [lang-sol-replication](/rounds/lang-sol-replication/)

### Related hypotheses

- [H91](/hypotheses/h91/)
- [H114](/hypotheses/h114/)

### Related experiments

- [lang-sol-replication](/rounds/lang-sol-replication/)
Exact cell IDs are listed in the [source ledger](https://github.com/KogenAI/kogen-bench/blob/baf7c2f4307d3a989a8fb9363df8000fda3d9310/results/claim-ledger.jsonl); the site does not join outcome-only rows by guessed keys.
