investment-autoresearch:report

Autoresearch Report

Generates a structured markdown report from archive/{ticker}-autoresearch-v{N}/autoresearch_result.json.

Input

Read the JSON file for the requested ticker:

archive/{ticker}-autoresearch-v{N}/autoresearch_result.json

If no version is specified, use the latest v{N} directory. If the JSON doesn't exist, run the investment-autoresearch-parse skill first.

Report Template

Fill each section from the JSON fields below. Omit a row/field only if the value is null.

# {ticker} Autoresearch Report

## Strategy Identity

| Field | Value |
|---|---|
| Ticker | {ticker} |
| Champion Strategy | {current_best.strategy_name} |
| Parameters | {current_best.parameters as key=value pairs} |

## Performance Summary

| Metric | Value |
|---|---|
| CAGR | {current_best.cagr}% |
| Max Drawdown | {current_best.max_drawdown}% |
| Sharpe | {current_best.sharpe ?? —} |
| Trades | {leaderboard[0].trades} |
| Trades / Parameter | {leaderboard[0].trades_per_param ?? —} |

## Risk-Adjusted Validation

Walk-forward RA = CAGR / |MaxDD| across rolling periods:

| Period | CAGR | RA |
|---|---|---|
| 5y | {walk_forward.5y.cagr ?? —} | {walk_forward.5y.ra ?? —} |
| 3y | {walk_forward.3y.cagr ?? —} | {walk_forward.3y.ra ?? —} |
| 2y | {walk_forward.2y.cagr ?? —} | {walk_forward.2y.ra ?? —} |
| 1y | {walk_forward.1y.cagr ?? —} | {walk_forward.1y.ra ?? —} |

**Min RA across periods: {current_best.min_ra_across_periods}**
_(Lower bound on risk-adjusted return; guards against period-specific overfitting)_

## Research Process

- **Rounds completed:** {research_summary.rounds_completed}
- **Agents run:** {research_summary.agents_run}
- **Strategies tested:** {research_summary.strategies_tested ?? unknown}

### Verified Insights

{verified_insights as numbered list}

### Rejected Approaches

| Approach | Why It Failed |
|---|---|
{rejected_approaches as table rows: approach | reason}

## Top Strategies Leaderboard

| Rank | Strategy | CAGR | MaxDD | Calmar | Trades |
|---|---|---|---|---|---|
{leaderboard rows}

## Recommendations

- **Graduate to production:** {recommendation.graduate}
- **Confidence:** {recommendation.confidence}

### Open Hypotheses

{open_hypotheses as numbered list}