← All tasks
T24 · Event-Driven Trading

Earnings Extrapolation and Predictable Stock Market Returns

Hongye Guo, 2025 · 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.

The aggregate market return in newsy months (Jan/Apr/Jul/Oct, when earnings are announced) is predicted by past returns with the opposite sign of non-newsy months; investors extrapolate announced earnings, so a sign-flipped, expanding-mean-demeaned, vol-scaled time-series market-timing strategy earns a Sharpe near 0.50.

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)
  • A2
    no global normalize
    66/70
  • A3
    no future data access
    64/70
  • A4
    causal consistency
    63/70
  • A6
    train test temporal order
    48/70
  • A8
    transaction timing
    37/70
  • A9
    signal delay test
    69/70
  • A10
    e2e metric consistency
    0/70
TYPE B · PAPER-SPECIFIC (10)
  • B1
    newsy month definition
    §§4 / Appendix C.1 “Group 1 contains January, April, July, and October... These newsy months largely correspond to the so-called 'earnings months'”
    provenance · paper-explicit
    69/70
  • B2
    sum past four newsy returns window
    §Appendix C.1 “I take the total returns of the aggregate market in the past four newsy months (potentially include month t−1)”
    provenance · paper-explicit
    50/70
  • B3
    expanding window mean computation
    §Appendix C.1 “compute its expanding window mean, which is the average from the beginning of the sample to month t−1”
    provenance · paper-explicit
    68/70
  • B4
    demeaned signal sign flip
    §Appendix C.1 “I then take the difference, and flip sign if month t is newsy to arrive at the demeaned signal x_{t−1}”
    provenance · paper-explicit
    54/70
  • B5
    constrained regression specification
    §Appendix C.1 “I then run a constrained time-series regression mkt_t = β x_{t-1} + 1·mkt_{t-1} + ε_t, where... the coefficient before mkt_{t-1} is constrained to be 1”
    provenance · paper-explicit
    53/70
  • B6
    portfolio weight ct times xt
    §Appendix C.1 “The portfolio weight in my strategy is c_t x_t, which roughly has a mean of zero over time”
    provenance · paper-explicit
    51/70
  • B7
    volatility scaling
    §Appendix C.1 “the portfolio is scaled by a constant so that it has the same volatility as the aggregate market, which is 5.34% per month”
    provenance · paper-explicit
    35/70
  • B8
    monthly rebalance frequency
    §Appendix C.1 “at the end of each month t−1 I take the total returns of the aggregate market in the past four newsy months”
    provenance · paper-explicit
    53/70
  • B9
    regression dependent variable total return
    §Appendix C.1 “constrained time-series regression mkt_t = β x_{t-1} + 1·mkt_{t-1} + ε_t, where mkt_t is the market return in month t”
    provenance · paper-explicit
    66/70
  • B10
    expanding window regression reestimation
    §Table C1 note “regress the aggregate market return on lag one month signal and the expanding window mean aggregate market return based on data available in real time at each month, and with the coefficient on the expanding window mean aggregate market return constrained to 1”
    provenance · paper-explicit
    49/70
What this task reveals

The lowest-passing observed requirement is e2e metric consistency: 0 of 70 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 Event-Driven Trading
T23 · Days to Cover and Stock Returns
Next in Event-Driven Trading →
T25 · Short selling efficiency