← All tasks
T01 · Momentum

Time Series Momentum

Tobias J. Moskowitz et al. (3 authors), 2012 · Journal of Financial Economics

28%
pass@1 across 16 systems
10/16
systems with ≥1 successful attempt
5/5
5/5 attempts passed for the best system
80 attempts · 16 systems × 5
Claude Code + Opus 5
40%
Claude Code + Opus 4.8
100%
Claude Code + Opus 4.7
40%
Codex + GPT-5.6-sol
40%
Codex + GPT-5.5
0%
OpenCode + Kimi-K3
0%
OpenCode + Opus 4.7
0%
OpenCode + GPT-5.5
0%
OpenCode + Sonnet 4.6
20%
OpenCode + Qwen3.7-Max
20%
OpenCode + GLM-5.2
60%
OpenCode + DeepSeek V4-Pro
40%
OpenCode + GLM-5.1
20%
OpenCode + Qwen 3.6 Plus
0%
OpenCode + MiniMax-M3
60%
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.

Each instrument's own past 12-month excess return predicts its next one-month return; trade long winners / short losers, sized inversely proportional to ex-ante volatility, equal-weighted across the diversified futures+FX+index universe.

Type
technical
Implementation
pandas
Rebalance
monthly
Long-short
yes
Declared task setup
Data
Bundled public 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

17 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 (7)
  • A3
    no future data
    §Section 1 “signals based on past returns only”
    stage · feature eng
    78/79
  • A4
    missing data handling
    §Robustness “strategy handles missing data gracefully”
    stage · signal gen
    69/79
  • A5
    temporal order
    stage · portfolio
    76/79
  • A7
    return computation
    §Section 2 “monthly returns from daily prices”
    stage · data loading
    76/79
  • A8
    transaction timing
    §Section 2 “signal at t applied to return from t to t+1”
    stage · execution
    78/79
  • A9
    survivorship bias
    §Data integrity “no retroactive removal of instruments”
    stage · signal gen
    75/79
  • A10
    deterministic
    §Deterministic “same inputs → same outputs”
    31/79
TYPE B · PAPER-SPECIFIC (10)
  • B1
    lookback 12 months
    §Section 2 “12-month lookback period for momentum signal”
    provenance · paper-explicit
    stage · signal gen
    51/79
  • B2
    signal is sign
    §Section 2 “sign of past 12-month return”
    provenance · paper-explicit
    stage · signal gen
    77/79
  • B3
    ewma com 60
    §Section 2 “EWMA with center of mass 60 days”
    provenance · paper-explicit
    stage · feature eng
    73/79
  • B4
    annualization 261
    §Section 2 “annualized using 261 trading days”
    provenance · paper-explicit
    stage · feature eng
    74/79
  • B5
    volatility lagged
    §Section 2 “σ_{t-1} used for position sizing”
    provenance · paper-explicit
    stage · feature eng
    76/79
  • B6
    position size inverse vol
    §Section 2 “position sized at 40% / ex ante volatility”
    provenance · paper-explicit
    stage · portfolio
    76/79
  • B7
    equal weighted portfolio
    §Section 2 “equal-weighted portfolio”
    provenance · paper-explicit
    stage · portfolio
    76/79
  • B8
    tsmom compound formula
    §Section 2 “TSMOM return formula”
    provenance · paper-explicit
    stage · portfolio
    74/79
  • B9
    no skip month
    §Section 2 “no skip month between signal and holding”
    provenance · paper-explicit
    stage · signal gen
    73/79
  • B10
    ewma full formula
    §Section 2 “full ex ante volatility formula”
    provenance · paper-explicit
    stage · feature eng
    75/79
What this task reveals

The lowest-passing observed requirement is deterministic: 31 of 79 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
Next in Momentum →
T02 · Momentum Crashes