parallax-ai-klarman

name

parallax-ai-klarman

description

Applies Seth Klarman's margin-of-safety framework (per 'Margin of Safety', 1991) to a single stock's current financials via Parallax. Four balance-sheet checks: net cash position, debt vs peers, FCF stability, valuation discount. Includes a distinctive 'no position warranted; cash is valid' output when nothing qualifies. Third-person framing, book citation, AI-inferred from public information. NOT financial advice. NOT personalized. Accepts plain tickers or RIC format. NOT for portfolio-level health check (use /parallax-portfolio-checkup). For all five profiles simultaneously use /parallax-ai-consensus.

Parallax AI Klarman Profile

When not to use

• Bottom-up factor scoring → use /parallax-ai-buffett
• Mechanical formula screen → use /parallax-ai-greenblatt
• 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, fallback patterns
• JIT-load _parallax/AI-profiles/profile-schema.md and output-template.md
• JIT-load _parallax/AI-profiles/profiles/klarman.md for the profile spec
• Klarman profile is balance-sheet-first — requires get_financials(balance_sheet) AND get_financials(cash_flow) AND get_financials(ratios), 4 periods each
• 3 statements × 4 periods would be 12 calls, but Parallax returns 4 periods per call → 3 calls + get_peer_snapshot + get_company_info = ~5 tokens
• Compute net cash from balance sheet (cash - total debt), not from the ratios summary
• The "no position warranted" output IS valid output — do not treat it as a failure
• NEVER use first-person impersonation; always "Klarman-style"
• Disclaimer verbatim per output-template.md, substituting "Seth Klarman" for [Investor]
• Value threshold is intentionally loose (≥ 4) per Lev-Srivastava 2022 intangibles caveat — do not tighten without re-anchoring
• 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 Seth Klarman's margin-of-safety framework to a single stock's current balance sheet, cash flow, and peer-relative valuation.

Usage

/parallax-ai-klarman AAPL.O
/parallax-ai-klarman BRKb.N
/parallax-ai-klarman 7203.T

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
2. Load _parallax/AI-profiles/profile-schema.md
3. Load _parallax/AI-profiles/output-template.md
4. Load _parallax/AI-profiles/profiles/klarman.md

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

Step 1 — Resolve ticker

Per shared conventions (RIC suffix table in parallax-conventions.md §1).

Step 2 — Fire data calls in parallel

Tool	Parameters	Purpose
get_company_info	symbol	Sector, market cap, for peer group selection
get_peer_snapshot	symbol	Factor sub-scores (Value used as backup check)
get_financials	symbol, statement=balance_sheet	Cash, total debt, equity (4 periods default)
get_financials	symbol, statement=cash_flow	FCF across 4 periods (default)
get_financials	symbol, statement=ratios	D/E, P/E, peer medians where available (4 periods default)

IMPORTANT — MCP parameter serialization: Do NOT pass numeric parameters like periods=4 as explicit arguments via the :periods=4 syntax — the MCP transport serializes them as strings, causing "Expected number, received string" validation errors. The server default is 4 periods; rely on it or pass as a properly-typed integer at the individual tool-call site.

Step 3 — Pre-render cross-validation gate (MANDATORY per spec §6.4)

After get_peer_snapshot returns, cross-check its target_company field (NOT name — that field on individual peer rows refers to each peer, not the target) against get_company_info's name. On mismatch, refuse to render and emit:

Error: Symbol cross-validation failed for <ticker>.
  get_company_info returned: "<name_a>"
  get_peer_snapshot target_company: "<name_b>"
Cannot render Klarman-style profile — possible wrong-company mapping (see parallax-conventions.md §2).

Step 4 — Compute the four Klarman checks

Check 1: Net cash position

• Net cash = (cash and equivalents) − (total debt), from most recent balance sheet
• Net cash ratio = net cash / market cap (from get_company_info)
• PASS if net cash ratio ≥ 0 (positive net cash)
• PARTIAL if ≥ −0.2 (small, manageable net debt)
• FAIL if < −0.2 (meaningful net debt)

Check 2: Debt vs peer median

• D/E from the ratios call, most recent period
• Compare against peer median D/E from the peer snapshot
• PASS if D/E ≤ peer median × 1.1
• FAIL otherwise
• If peer median D/E is unavailable, flag "Peer comparison unavailable — check computed on absolute D/E < 1.0 as fallback"

Check 3: FCF stability

• Count periods (of 4) where free cash flow was positive
• PASS if ≥ 3 of 4
• PARTIAL if 2 of 4
• FAIL if ≤ 1 of 4

Check 4: Valuation discount

• P/E vs peer median P/E (fallback: P/B if negative earnings make P/E meaningless)
• PASS if stock P/E ≤ peer median P/E × 0.85 (15%+ discount)
• PARTIAL if stock P/E ≤ peer median × 1.0 (at peer)
• FAIL if stock P/E > peer median × 1.0

Backup: Parallax Value sub-score

• Sanity check, not a primary criterion
• Threshold: ≥ 4 (loose per Lev-Srivastava 2022 intangibles caveat)
• If Value < 4, flag in output: "⚠️ Parallax Value sub-score suggests valuation may not be attractive even if peer-relative metrics pass"

Step 5 — Compute verdict

Let N be the count of PASSing balance-sheet checks (0-4).

• N ≥ 3 AND Value ≥ 4 → match
• N ≥ 3 AND Value < 4 → partial_match — balance sheet qualifies but Parallax Value backup fails. Flag in output: "strong balance sheet but absolute Parallax Value below backup threshold — intangibles-era calibration caveat applies."
• N = 2 → partial_match
• N ≤ 1 → no_match

If N = 0 AND Value < 4, append the distinctive footer: "No position warranted on this ticker per margin-of-safety principles. Cash is a valid stance." (This is the explicit "do nothing" output from the profile body.)

Step 6 — Render through output template

Klarman-style profile applied to <ticker>

Source: Klarman, S. (1991). Margin of Safety: Risk-Averse Value Investing Strategies for the Thoughtful Investor.

Klarman's framework prioritizes survival over return: balance-sheet strength, downside protection via valuation discount, and willingness to hold cash when nothing qualifies.

| Check                        | Target                    | Actual              | Result  |
|------------------------------|---------------------------|---------------------|---------|
| Net cash position            | Net cash ratio ≥ 0        | <value>             | PASS/PARTIAL/FAIL |
| Debt vs peer median          | D/E ≤ peer median × 1.1   | <value> vs <peer>   | PASS/FAIL |
| FCF stability                | Positive ≥ 3 of 4 periods | <count> of 4        | PASS/PARTIAL/FAIL |
| Valuation discount           | P/E ≤ peer median × 0.85  | <value> vs <peer>   | PASS/PARTIAL/FAIL |

Parallax Value sub-score (backup check): <score> / 10

Profile fit: <match / partial_match / no_match> — <count> of 4 checks passed.

Verdict sensitivity: the verdict-relevant input nearest its cutoff is <check or Value backup> (<actual> vs <target>); state the arithmetic crossing that would change the pass count or Value backup status, and state the verdict change only when that crossing changes the overall match / partial / no-match tier.

[Optional: "No position warranted on this ticker per margin-of-safety principles. Cash is a valid stance."]

Workflow derived from: Klarman, S. (1991). Margin of Safety.
Last anchor-tested: 2026-04-07 (BRKb.N, NVDA.O)
Tool sequence: get_company_info, get_peer_snapshot, get_financials(balance_sheet/cash_flow/ratios, 4 periods)
Token cost: ~5 tokens

---
This output is an AI-inferred interpretation of Seth Klarman'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 Seth Klarman 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>** Klarman-style profile. 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 balance-sheet data is unavailable for 4 periods, compute the checks on whatever is available (minimum 2 periods) and flag the coverage loss. If peer-median data is unavailable, skip the debt-vs-peers check and note "Peer comparison unavailable — absolute thresholds applied." If all four checks fail due to missing data, return DATA_UNAVAILABLE rather than a false no_match.