← All tasks
T05 · Momentum

Comomentum: Inferring Arbitrage Activity from Return Correlations

Dong Lou and Christopher Polk, 2022 · Review of Financial Studies

0%
pass@1 across 16 systems
0/16
systems with ≥1 successful attempt
0/5
no system produced a successful attempt
80 attempts · 16 systems × 5
Claude Code + Opus 5
0%
Claude Code + Opus 4.8
0%
Claude Code + Opus 4.7
0%
Codex + GPT-5.6-sol
0%
Codex + GPT-5.5
0%
OpenCode + Kimi-K3
0%
OpenCode + Opus 4.7
0%
OpenCode + GPT-5.5
0%
OpenCode + Sonnet 4.6
0%
OpenCode + Qwen3.7-Max
0%
OpenCode + GLM-5.2
0%
OpenCode + DeepSeek V4-Pro
0%
OpenCode + GLM-5.1
0%
OpenCode + Qwen 3.6 Plus
0%
OpenCode + MiniMax-M3
0%
OpenCode + MiniMax M2.7
0%
all checkers pass Type-A clean, Type-B fail fail checking timeout no result / incomplete
Strategy

What the paper asks for.

Cross-sectional 12-1 momentum (long top decile / short bottom decile, value-weighted) is timed by Comomentum: average pairwise partial correlation of weekly Fama-French 3 residuals across the loser decile measured over a trailing 52-week window. High Comomentum quintile predicts low subsequent momentum returns.

Type
cross_sectional
Implementation
pandas
Rebalance
monthly
Long-short
yes
Declared task setup
Data
One-command fetcher for WRDS-sourced task data
Operational scope
The paper-defined methodological core is retained; the fixed instruction instantiates only the data mappings, interfaces, and evaluation conventions needed for execution.
Six-stage quantitative strategy pipeline
  1. 01 Data loading
  2. 02 Feature engineering
  3. 03 Model fitting
  4. 04 Signal generation
  5. 05 Portfolio construction
  6. 06 Execution
Checker suite

18 properties · 80 attempts.

Each row is one requirement the strategy must satisfy. The bar reports the share of observed checker verdicts that pass; missing checker output is not converted into a property verdict.

TYPE A · DOMAIN-LEVEL (8)
  • A1
    no backfill
    stage · data loading
    39/69
  • A3
    no future data access
    stage · feature eng
    54/69
  • A4
    causal consistency
    stage · signal gen
    60/69
  • A5
    no survivorship bias
    stage · portfolio
    67/69
  • A6
    no lookahead in features
    stage · model fit
    43/69
  • A7
    no target leakage
    stage · data loading
    54/69
  • A8
    realistic execution delay
    stage · execution
    65/69
  • A10
    e2e metric consistency
    4/69
TYPE B · PAPER-SPECIFIC (10)
  • B1
    momentum lookback 12m skip 1m
    §Section 3 “At the end of each month, we sort all stocks into deciles based on their previous 12-month return (skipping the most recent month).”
    provenance · paper-explicit
    stage · signal gen
    23/69
  • B2
    momentum decile sort
    §Section 3 “we sort all stocks into deciles based on their previous 12-month return”
    provenance · paper-explicit
    stage · signal gen
    68/69
  • B3
    universe price filter
    §Section 3 “To mitigate the impact of microstructure issues, stocks with prices below $5 a share and/or are in the bottom NYSE size decile are excluded from the sample.”
    provenance · paper-explicit
    stage · data loading
    65/69
  • B4
    universe nyse size filter
    §Section 3 “stocks with prices below $5 a share and/or are in the bottom NYSE size decile are excluded from the sample”
    provenance · paper-explicit
    stage · data loading
    65/69
  • B5
    comomentum partial correlation ff3
    §Section 3 “We then compute pairwise partial correlations using 52 weekly returns for all stocks in each decile in the portfolio ranking period. We control for the Fama-French three factors in computing these partial correlations to purge out any comovement in stock returns in the same momentum decile induced by known risk factors.”
    provenance · paper-explicit
    stage · feature eng
    60/69
  • B6
    comomentum pairwise average excludes diagonal
    §Section 3 “We then compute pairwise partial correlations using 52 weekly returns for all stocks in each decile in the portfolio ranking period.”
    provenance · paper-explicit
    stage · feature eng
    32/69
  • B7
    momentum portfolio value weighted
    §Section 4 (Table III description) “we then form a zero-cost portfolio that goes long a value-weight portfolio of the stocks in the top decile and short a value-weight portfolio of stocks in the bottom decile”
    provenance · paper-explicit
    stage · portfolio
    67/69
  • B8
    comomentum quintile classification
    §Section 4 (Table III description) “All months are then classified into five groups based on their loser comomentum.”
    provenance · paper-explicit
    stage · signal gen
    68/69
  • B9
    comomentum uses pairwise not leave one out
    §Section 3 “We then compute pairwise partial correlations using 52 weekly returns for all stocks in each decile in the portfolio ranking period.”
    provenance · paper-explicit
    stage · feature eng
    59/69
  • B10
    nyse breakpoints for momentum sort
    §Section 3 “stocks with prices below $5 a share and/or are in the bottom NYSE size decile are excluded from the sample”
    provenance · benchmark-specified
    stage · data loading
    35/69
What this task reveals

The lowest-passing observed requirement is e2e metric consistency: 4 of 69 observed checker verdicts pass. The full property list above shows whether errors concentrate in domain-level validity or the paper-specific strategy logic.

Open paper → All tasks
← Previous in Momentum
T04 · Value and Momentum Everywhere