parallax-ai-greenblatt

name

parallax-ai-greenblatt

description

Applies Joel Greenblatt's Magic Formula (per 'The Little Book That Beats the Market', 2006; academic replication Gray & Carlisle 2012) to Parallax data. Two modes: universe mode returns a top-decile ranked basket; ticker-check mode reports whether a single stock falls in the top decile of its peer universe by combined ROC + earnings yield rank. Third-person framing, book citation, AI-inferred from public information. NOT financial advice. NOT personalized. Accepts no args (universe mode) or a single ticker. NOT for portfolio-level health check (use /parallax-portfolio-checkup). For all five profiles simultaneously use /parallax-ai-consensus.

Parallax AI Greenblatt Profile

When not to use

• Bottom-up factor scoring without the mechanical formula → use /parallax-ai-buffett
• Balance-sheet-first special situations → use /parallax-ai-klarman
• Top-down macro analysis → use /parallax-ai-soros
• Cross-profile consensus → use /parallax-ai-consensus
• Full due diligence → use /parallax-due-diligence
• Running backtests → use /backtest

Gotchas

• JIT-load _parallax/parallax-conventions.md for RIC resolution, parallel execution, fallbacks
• JIT-load _parallax/AI-profiles/profile-schema.md for the dispatcher workflow and cross-validation gate
• JIT-load _parallax/AI-profiles/output-template.md for the required output structure and verbatim disclaimer
• JIT-load _parallax/AI-profiles/profiles/greenblatt.md for the profile spec
• Universe mode is default when no ticker is provided; ticker-check mode activates with a single ticker
• build_stock_universe is ~5 tokens; budget accordingly
• Exclude financials and utilities from default universe per Greenblatt's original rule
• NEVER use first-person impersonation — always "Greenblatt-style" or "Magic Formula"
• Disclaimer verbatim; substitute "Joel Greenblatt" for [Investor] in the disclaimer block
• Profile is derived from public book + academic replication only — no get_assessment, no score_total
• JIT-load _parallax/white-label/integration-pattern.md before the Pre-Render step. Loader call is load_visual_branding() (7-key visual subset; voice structurally excluded — branding["voice"] raises KeyError). Apply §5 (Branding Header) and §7 (About This Report) in Output Format.

Applies Joel Greenblatt's Magic Formula (ROC rank + earnings yield rank, combined, top decile) to Parallax data.

Usage

/parallax-ai-greenblatt                    # universe mode — default US equity universe
/parallax-ai-greenblatt AAPL               # ticker-check mode — is AAPL in the top decile?
/parallax-ai-greenblatt AAPL.O             # RIC format also accepted

Workflow

Execute using mcp__claude_ai_Parallax__* tools.

Step 0 — JIT-load dependencies

Before the first Parallax tool call:

1. Load _parallax/parallax-conventions.md — RIC resolution, parallel execution, fallback patterns.
2. Load _parallax/AI-profiles/profile-schema.md — dispatcher workflow and cross-validation gate.
3. Load _parallax/AI-profiles/output-template.md — required output structure and disclaimer.
4. Load _parallax/AI-profiles/profiles/greenblatt.md — profile spec.

Call ToolSearch with query "+Parallax" to load the deferred MCP tool schemas.

Step 1 — Determine mode

• No ticker → universe mode
• Exactly one ticker → ticker-check mode
• Multiple tickers → reject: "Greenblatt profile takes zero or one ticker. For multi-ticker checks use /parallax-ai-consensus."

Step 2 — Universe mode workflow

IMPORTANT — query scoping: build_stock_universe is async and broad queries consistently time out in practice (e.g., "US large-cap equities with high ROIC" — confirmed timeout). The default query MUST be sector-scoped or otherwise narrow. Valid defaults:

• "US large-cap consumer staples" ← DEFAULT if no theme provided
• "US large-cap industrials"
• "US large-cap healthcare"
• User-provided theme passed via --universe "<theme>"

Do NOT pass broad queries like "US large-cap and mid-cap equities excluding financials and utilities" — they time out. If a user requests a broad screen, execute sector-by-sector and merge the rankings.

1. Call build_stock_universe with the sector-scoped query. Cost: 5 tokens. If the call times out, retry ONCE with a narrower query (e.g., drop "large-cap and mid-cap" → "large-cap"). If the retry also times out, return INSUFFICIENT_UNIVERSE and decline to render a verdict.
2. Cap the universe at top 30 names by composite_score from the build_stock_universe response (which already ranks candidates internally). This bounds the token cost for Step 3.
3. For each candidate in the top-30 cap, call get_financials(statement=ratios) in parallel batches to pull ROC and earnings yield. Substitute Parallax's return_on_invested_capital if direct ROC is not available. Derive earnings yield as 1 / enterprise_value_ebit. Cost: 1 token per name, so ~30 tokens max.
4. Rank each candidate on ROC and earnings yield independently. Sum the two ranks. Sort by combined rank ascending.
5. Take the top 10% as the Magic Formula basket (top 3 names out of 30).
6. For the top 3 basket members, call get_peer_snapshot (1 token each) to surface Parallax factor scores as pedagogy.

Step 3 — Ticker-check mode workflow

1. Resolve ticker per shared conventions.
2. Call get_company_info to identify the ticker's sector/industry.
3. Call build_stock_universe with a sector-based peer universe query derived from the ticker's sector (e.g., if AAPL is in "Technology Hardware," query "US large-cap technology hardware"). Sector-scoped queries succeed where broad queries time out.
4. Cap the peer universe at top 30 names by composite_score.
5. Run Step 2 sub-steps 3-4 on the capped peer universe.
6. Check where the target ticker ranks in the combined distribution.

Step 4 — Cross-validation gate

For ticker-check mode, after get_peer_snapshot on the target ticker, cross-check target_company (the top-level field — NOT name on individual peer rows) against get_company_info's name. Refuse to render on mismatch per profile-schema.md §2 Step 2.

For universe mode, cross-validation is per-name on the top-3 basket members that get get_peer_snapshot calls.

Step 5 — Compute verdict (ticker-check mode only)

• Top 10% of combined rank → match
• Top 25% but below top 10% → partial_match
• Below top 25% → no_match

Step 6 — Render through output template

Universe mode output:

Greenblatt-style basket (Magic Formula)
Source: Greenblatt, J. (2006). The Little Book That Beats the Market.
Secondary: Gray, W., Carlisle, T. (2012). Quantitative Value.

Universe: <query>
Universe size: <N>

Top decile by combined ROC + earnings yield rank:

| Rank | Ticker  | ROC rank | EY rank | Combined | Sector          |
|------|---------|----------|---------|----------|-----------------|
| 1    | XXX.N   | 2        | 5       | 7        | Consumer Disc.  |
| 2    | YYY.O   | 4        | 3       | 7        | Technology      |
| ...  | ...     | ...      | ...     | ...      | ...             |

Workflow derived from: Greenblatt, J. (2006). The Little Book That Beats the Market; Gray & Carlisle (2012).
Last anchor-tested: 2026-04-06 (CSCO.O, MSFT.O, NVDA.O)
Tool sequence: build_stock_universe, get_peer_snapshot × N, get_financials(ratios) × N
Token cost: ~10-30 tokens (universe mode) / ~10-15 tokens (ticker-check mode)

---
This output is an AI-inferred interpretation of Joel Greenblatt's approach, derived solely from publicly available information — the cited source, Parallax factor data, and Parallax's public methodology. It is produced by the Parallax AI Investor Profiles framework. It is not financial advice, not personalized, not endorsed by Joel Greenblatt or his representatives, and not a recommendation to buy or sell any security. For illustrative and educational use only. Past characterization does not guarantee future relevance. Please consult a qualified financial advisor before making investment decisions.

Ticker-check mode output:

Greenblatt-style profile applied to <ticker>
Source: Greenblatt, J. (2006). The Little Book That Beats the Market.

Peer universe: <sector peers>
Universe size: <N>

<Ticker> ranking:
  ROC rank:            <R> of <N> (<percentile>th percentile)
  Earnings yield rank: <R> of <N> (<percentile>th percentile)
  Combined rank:       <R> of <N> (<percentile>th percentile)

Profile fit: <match/partial_match/no_match> — <ticker> is in the [top 10% | top 25% | below top 25%] of the peer universe by Greenblatt combined rank.

Verdict sensitivity: combined rank sits at the <percentile>th percentile, <D> positions from the [top-10% match | top-25% partial] cutoff; the verdict flips to the adjacent tier if the combined rank crosses that boundary.

Workflow derived from: Greenblatt, J. (2006). The Little Book That Beats the Market; Gray & Carlisle (2012).
Last anchor-tested: 2026-04-06 (CSCO.O, MSFT.O, NVDA.O)
Tool sequence: get_peer_snapshot, get_financials(ratios) × N
Token cost: ~10-15 tokens

---
This output is an AI-inferred interpretation of Joel Greenblatt's approach, derived solely from publicly available information — the cited source, Parallax factor data, and Parallax's public methodology. It is produced by the Parallax AI Investor Profiles framework. It is not financial advice, not personalized, not endorsed by Joel Greenblatt or his representatives, and not a recommendation to buy or sell any security. For illustrative and educational use only. Past characterization does not guarantee future relevance. Please consult a qualified financial advisor before making investment decisions.

Step 7 — Emit

Output additions (white-label branding + §9.2 disclosure)

These additions apply to the rendered output ABOVE in addition to the persona-specific disclaimer shown in the output example. They are required regardless of view state.

Pre-Render — Load white-label branding

Load _parallax/white-label/integration-pattern.md §2 and compute white_label_active + client_name per that section. Apply §5 (Branding Header) and §7 (About This Report) when composing the Output Format.

• Branding Header (only if white_label_active AND client_name != "") — single line at the very top of the rendered output: **<client_name>** Greenblatt-style screen. Logo handling per integration-pattern.md §5.
• About This Report (always present): one line stating branding state per integration-pattern.md §7. If a logo was skipped, append Logo on file: <basename> as a second About This Report line.

AI-interaction disclosure (required regardless of view state): Render parallax-conventions.md §9.2 immediately above the disclaimer below. The persona-specific disclaimer in the output example characterizes the source of the framing; the §9.2 banner characterizes the LLM-generated synthesis itself.

Render the standard disclaimer verbatim from parallax-conventions.md §9.1.

Graceful fallback

If build_stock_universe returns fewer than 10 names, the top-decile calculation is unreliable. Expand the universe query once; if expansion fails, return INSUFFICIENT_UNIVERSE and decline to render a verdict. In ticker-check mode, this means the /parallax-ai-consensus meta-skill should treat the profile as skipped.

If get_financials(ratios) fails for a subset of universe members, drop those names from the ranking and note the coverage loss in the methodology footer.