| name | research-progression-docs |
| description | Creating PROGRESSION.md documents for research experiment series in Volatio. Captures the evolution of ideas, learnings, and breakthroughs across numbered experiments (001, 002, ...). Use when documenting experiment series or reviewing research evolution. |
research-progression-docs
When to Use
- Documenting a completed experiment series
- Reviewing the evolution of research ideas
- Onboarding to understand how a strategy developed
- Extracting cross-experiment learnings
Purpose
Research experiments are numbered sequentially (001, 002, ...) where each builds on previous learnings. A PROGRESSION.md captures:
- Why each experiment was run (the hypothesis)
- What was learned (the finding)
- How it influenced the next experiment (the progression)
- Which breakthroughs changed direction
Document Structure
# [Series Name]: Experiment Progression
> **Series:** [Full description]
> **Goal:** [What we're trying to achieve]
> **Status:** ✅ ACTIVE | ⚠️ DEPRECATED | ✅ COMPLETED
---
## Progression Summary
| Phase | Experiments | Focus | Outcome |
|-------|-------------|-------|---------|
| 1 | 001-005 | Signal Discovery | [Key outcome] |
| 2 | 006-010 | Validation | [Key outcome] |
| 3 | 011-015 | Optimization | [Key outcome] |
---
## Phase 1: [Phase Name]
### 001 - [Experiment Name]
**Idea:** [What hypothesis are we testing?]
**Finding:** [What did we learn?]
**Progression:** [How does this inform the next experiment?]
---
### 002 - [Experiment Name] ✅ (BREAKTHROUGH)
**Idea:** [Hypothesis]
**Finding:** 🎯 **[Major discovery in bold]**
**Progression:** [Next steps]
---
## Key Learnings
1. **[Learning 1]** — [Explanation]
2. **[Learning 2]** — [Explanation]
---
## Evolution of Best Strategy
| Experiment | Strategy | Sharpe |
|------------|----------|--------|
| 001 | Baseline | 0.50 |
| 005 | +Optimization | 1.20 |
| 010 | +Filter | **2.50** |
---
## Production Configuration
```python
SIGNAL = "..."
ENTRY = "..."
EXIT = "..."
---
## Entry Format for Each Experiment
Each experiment entry follows this pattern:
```markdown
### NNN - Experiment Name [Status Emoji]
**Idea:** One sentence describing the hypothesis or question being tested.
**Finding:** What was discovered. Use 🎯 for breakthroughs, tables for data.
**Progression:** How this informs the next experiment or changes direction.
---
Status Emojis
| Emoji | Meaning |
|---|
| ✅ | Successful, confirmed |
| ❌ | Failed, disproven |
| ⚠️ | Partial success, needs caveat |
| 🎯 | Major breakthrough |
| (none) | Standard experiment |
Marking Breakthroughs
When an experiment changes the direction of research, mark it clearly:
### 010 - Validated SHORT Signal ✅ (BREAKTHROUGH)
**Idea:** Lock in the best SHORT signal with validation.
**Finding:** 🎯 **CORE SIGNAL DISCOVERED**
| Metric | Value |
|--------|-------|
| Sharpe | 12.67 |
| Win Rate | 83% |
**Progression:** Core signal validated. Now optimize.
Capturing Failed Experiments
Failed experiments are as important as successes — they prevent repeating mistakes:
### 003 - Earnings Reversal
**Idea:** Big moves revert the next day (mean reversion hypothesis).
**Finding:** Mean reversion appeared to work with high returns.
❌ *Later found to have look-ahead bias (see experiment 023)*
**Progression:** Invalidated — do not use.
Cross-Experiment Insights
When insights from one series apply to others, capture this:
## Key Learnings
| Insight | Discovered In | Applied To | Result |
|---------|---------------|------------|--------|
| Signal Exit | Laplace 016 | Extrinsic | ❌ Doesn't work |
| Signal Exit | Laplace 016 | Risk-Rotation | ✅ Works |
| Regime Filter | Extrinsic 036 | Risk-Rotation | ❌ Hurts |
Evolution Table
Track how the best strategy improved across experiments:
## Evolution of Best Strategy
| Experiment | Change | Sharpe | Δ |
|------------|--------|--------|---|
| 008 | Baseline | 0.50 | — |
| 027 | +Production params | 0.70 | +0.20 |
| 031 | +Net Put-Call | 1.62 | +0.92 |
| 033 | +DTE 5-14 | 1.99 | +0.37 |
| 035 | +Trail50 | **4.72** | +2.73 |
Production Configuration
End with actionable configuration for the final strategy:
## Production Configuration
```python
# Signal
SIGNAL = "Net Put-Call z-score > 1.0"
DTE_RANGE = (5, 14)
# Entry
ENTRY = "T+1 Open"
# Exit
EXIT = "trail50_7d"
MAX_HOLD = 7 # days
# Filters
REGIME_FILTER = "STABLE only"
---
## Examples in Codebase
| Series | Location | Experiments |
|--------|----------|-------------|
| SEC 8-K Original | `research/experiments/sec-8k-signals/PROGRESSION.md` | 001-023 |
| SEC 8-K V3 | `research/experiments/sec-8k-signals/v3/PROGRESSION.md` | 001-023 |
| Extrinsic Flow | `research/experiments/options/extrinsic/PROGRESSION.md` | 001-038 |
| Risk-Rotation | `research/experiments/options/extrinsic-risk-rotation/PROGRESSION.md` | 001-017 |
| Laplace | `research/experiments/options/laplace/PROGRESSION.md` | 001-021 |
---
## Checklist for New Progression Doc
- [ ] Header with series name, goal, and status
- [ ] Progression summary table with phases
- [ ] Individual entry for each experiment
- [ ] Breakthroughs clearly marked with 🎯
- [ ] Failed experiments marked with ❌ and explanation
- [ ] Key learnings section with cross-experiment insights
- [ ] Evolution table showing strategy improvement
- [ ] Production configuration if strategy is ready
