Skip to content
Back to skills

2516 Scoring System 8a4825b6

ASecurity

| Component | Weight | Source | |-----------|--------|--------| | Trend Template (Stage 2) | 25% | 7-point Minervini criteria | | Contraction Quality | 25% | VCP pattern detection | | Volume Pattern | 20% | Volume dry-up analysis | | Pivot Proximity | 15% | Distance from breakout level | | Relative Strength | 15% | Minervini-weighted RS vs S&P 500 |

  • 9 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 11, 2026
developmentgoperformance

Works with

  • cli

Security analysis

A100/100

Scanned October 11, 2026

npx -y skills add tools-only/X-Skills --skill 2516-scoring_system_8a4825b6 --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of 2516 Scoring System 8a4825b6?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for 2516 Scoring System 8a4825b6
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/tools-only-2516-scoring-system-8a4825b6/badge)](https://www.skillsdirectory.com/skills/tools-only-2516-scoring-system-8a4825b6)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

SKILL.md
# VCP Screener Scoring System

## 5-Component Composite Score

| Component | Weight | Source |
|-----------|--------|--------|
| Trend Template (Stage 2) | 25% | 7-point Minervini criteria |
| Contraction Quality | 25% | VCP pattern detection |
| Volume Pattern | 20% | Volume dry-up analysis |
| Pivot Proximity | 15% | Distance from breakout level |
| Relative Strength | 15% | Minervini-weighted RS vs S&P 500 |

## Component Scoring Details

### 1. Trend Template (0-100)

Each of the 7 criteria contributes 14.3 points:

| Criteria Passed | Score | Status |
|-----------------|-------|--------|
| 7/7 | 100 | Perfect Stage 2 |
| 6/7 | 85.8 | Pass (minimum threshold) |
| 5/7 | 71.5 | Borderline |
| <= 4/7 | <= 57 | Fail |

**Pass threshold:** Score >= 85 (6+ criteria) to proceed to VCP analysis.

### 2. Contraction Quality (0-100)

| # Contractions | Base Score |
|----------------|-----------|
| 4 | 90 |
| 3 | 80 |
| 2 | 60 |
| 1 or invalid | 0-40 |

**Modifiers:**
- Tight final contraction (< 5% depth): +10
- Good average contraction ratio (< 0.4 of T1): +10
- Deep T1 (> 30%): -10

### 3. Volume Pattern (0-100)

Based on dry-up ratio (recent 10-bar avg / 50-day avg):

| Dry-Up Ratio | Base Score |
|-------------|-----------|
| < 0.30 | 90 |
| 0.30-0.50 | 75 |
| 0.50-0.70 | 60 |
| 0.70-1.00 | 40 |
| > 1.00 | 20 |

**Modifiers:**
- Breakout on 1.5x+ volume: +10
- Net accumulation > 3 days (in 20d): +10
- Net distribution > 3 days (in 20d): -10

### 4. Pivot Proximity (0-100) — Distance-Priority Scoring

Scoring is distance-first. Volume confirmation adds a bonus only within 0-5% above pivot (Minervini: never chase >5% above pivot).

| Distance from Pivot | Base Score | Volume Bonus | Final Score | Trade Status |
|--------------------|-----------|-------------|------------|--------------|
| 0-3% above | 90 | +10 | 100 | BREAKOUT CONFIRMED |
| 3-5% above | 65 | +10 | 75 | EXTENDED - Moderate chase risk (vol confirmed) |
| 5-10% above | 50 | — (none) | 50 | EXTENDED - High chase risk |
| 10-20% above | 35 | — (none) | 35 | EXTENDED - Very high chase risk |
| >20% above | 20 | — (none) | 20 | OVEREXTENDED - Do not chase |
| 0 to -2% below | 90 | — | 90 | AT PIVOT (within 2%) |
| -2% to -5% | 75 | — | 75 | NEAR PIVOT |
| -5% to -8% | 60 | — | 60 | APPROACHING |
| -8% to -10% | 45 | — | 45 | DEVELOPING |
| -10% to -15% | 30 | — | 30 | EARLY |
| < -15% | 10 | — | 10 | FAR FROM PIVOT |

**Volume bonus rules:**
- 0-3% above pivot + volume: +10 points, status = "BREAKOUT CONFIRMED"
- 3-5% above pivot + volume: +10 points, "(vol confirmed)" appended to status
- >5% above pivot: no volume bonus (Minervini: do not chase extended breakouts)
- Below pivot: volume bonus not applicable

**Chase risk rule (Minervini):** Do not buy stocks >5% above their pivot point. Distance determines the base score; volume confirmation is a bonus, not an override.

### 5. Relative Strength (0-100)

Minervini weighting (emphasizes recent performance):
- 40%: Last 3 months (63 trading days)
- 20%: Last 6 months (126 trading days)
- 20%: Last 9 months (189 trading days)
- 20%: Last 12 months (252 trading days)

| Weighted RS vs S&P 500 | Score | RS Rank Estimate |
|-------------------------|-------|------------------|
| >= +50% | 100 | ~99 (top 1%) |
| >= +30% | 95 | ~95 (top 5%) |
| >= +20% | 90 | ~90 (top 10%) |
| >= +10% | 80 | ~80 (top 20%) |
| >= +5% | 70 | ~70 (top 30%) |
| >= 0% | 60 | ~60 (top 40%) |
| >= -5% | 50 | ~50 (average) |
| >= -10% | 40 | ~40 |
| >= -20% | 20 | ~25 |
| < -20% | 0 | ~10 |

## Rating Bands

| Composite Score | Rating | Position Sizing | Action |
|-----------------|--------|-----------------|--------|
| 90-100 | Textbook VCP | 1.5-2x normal | Buy at pivot, aggressive |
| 80-89 | Strong VCP | 1x normal | Buy at pivot, standard |
| 70-79 | Good VCP | 0.75x normal | Buy on volume confirmation |
| 60-69 | Developing VCP | Wait | Watchlist only |
| 50-59 | Weak VCP | Skip | Monitor only |
| < 50 | No VCP | Skip | Not actionable |

### valid_vcp Gate Rule

When the VCP pattern calculator returns `valid_vcp=false` (e.g., contraction ratios exceed 0.75, expanding contractions), the rating is capped regardless of composite score:

- If `valid_vcp=false` AND composite >= 70: rating is overridden to **"Developing VCP"** with guidance "Watchlist only - VCP pattern not validated, do not buy"
- If `valid_vcp=false` AND composite < 70: no override needed (already below actionable threshold)

This prevents stocks with expanding or non-contracting patterns from receiving actionable buy ratings.

## Entry Ready Conditions

A stock is classified as `entry_ready=True` when all of the following conditions are met:

| Condition | Default Threshold | CLI Override |
|-----------|-------------------|--------------|
| `valid_vcp` | `True` | `--no-require-valid-vcp` |
| `distance_from_pivot_pct` | -8.0% to +3.0% | `--max-above-pivot` |
| `dry_up_ratio` | <= 1.0 | — |
| `risk_pct` | <= 15.0% | `--max-risk` |

**Report sections:**
- **Section A: Pre-Breakout Watchlist** — `entry_ready=True` stocks, sorted by composite score
- **Section B: Extended / Quality VCP** — `entry_ready=False` stocks, sorted by composite score

**CLI mode:**
- `--mode all` (default): Shows both sections
- `--mode prebreakout`: Shows only entry_ready=True stocks

## Pre-Filter Criteria (Phase 1)

Quick filter using quote data only (no historical needed):

| Criterion | Threshold | Purpose |
|-----------|-----------|---------|
| Price | > $10 | Exclude penny stocks |
| % above 52w low | > 20% | Roughly in uptrend |
| % below 52w high | < 30% | Not in deep correction |
| Average volume | > 200,000 | Sufficient liquidity |

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.