soph-sophisticated-analysis

name	soph - Sophisticated Analysis
description	Apply maximum analytical sophistication to a problem — multi-layered reasoning, cross-domain synthesis, second-order effects, hidden assumptions, and structural insights that simple analysis misses.
output	{"format":"prose"}

Sophisticated Analysis

Input: $ARGUMENTS

---

Core Principles

1. Sophistication is seeing structure that others miss. A simple analysis sees the surface: "X causes Y." A sophisticated analysis sees the structure: "X causes Y through mechanism M, but only when condition C holds, and the same mechanism also causes Z which undermines Y over time." Sophistication means identifying the hidden structural features — feedback loops, phase transitions, composition effects, boundary conditions.

2. Sophistication is not complexity for its own sake. Adding layers of analysis that don't change the conclusion is theater, not sophistication. Every additional layer must either (a) change the recommended action, (b) identify a risk that simple analysis misses, or (c) reveal an opportunity that simple analysis overlooks. If it doesn't do one of these, it's decorative.

3. The most sophisticated move is sometimes the simple one. True sophistication includes knowing when complexity is unnecessary. If the problem genuinely is simple, saying so (with evidence for why the simple reading is correct) IS the sophisticated analysis. Forced complexity is the opposite of sophistication.

4. Second-order effects are where sophistication earns its keep. First-order: "If we cut prices, we sell more." Second-order: "If we cut prices, competitors respond, margins compress industry-wide, and the equilibrium shifts." Third-order: "And that shifts customer expectations permanently, making price increases impossible." Simple analysis stops at first order. Sophisticated analysis follows the chain.

5. Cross-domain pattern recognition is the hallmark. When you recognize that a pricing problem has the same structure as an arms race, or that a team dysfunction mirrors a coordination game, that's sophisticated analysis. The pattern transfer reveals insights that domain-internal analysis can't reach.

---

Phase 1: Surface Analysis

Start with the obvious reading. This isn't the answer — it's the baseline.

[S1] PROBLEM: [the problem or question as stated]
[S2] SURFACE_READING: [what a quick, simple analysis would conclude]
[S3] SURFACE_ACTION: [what action the simple reading suggests]
[S4] CONFIDENCE_IN_SURFACE: [high | medium | low — is the simple reading likely sufficient?]

If confidence is high, state so explicitly. Not everything needs sophisticated analysis.

→ If the surface reading is likely sufficient: say so, explain why, and offer the sophisticated path as optional.

---

Phase 2: Structural Decomposition

Look beneath the surface for hidden structure.

[S5] HIDDEN_VARIABLES: [what factors are not mentioned but are driving the situation?]
[S6] MECHANISM: [through what mechanism does the cause produce the effect?]
[S7] BOUNDARY_CONDITIONS: [under what conditions does this analysis hold? Where does it break?]
[S8] FEEDBACK_LOOPS: [does the effect feed back into the cause? Positive or negative feedback?]
[S9] PHASE_TRANSITIONS: [are there thresholds where behavior changes qualitatively?]
[S10] COMPOSITION_EFFECTS: [does this interact with other factors in non-obvious ways?]

Structural Questions

Question	What It Reveals
What's the mechanism?	HOW the cause produces the effect (not just THAT it does)
What are the boundary conditions?	WHERE the analysis stops being true
What feeds back?	Whether effects amplify or dampen over time
What's the timescale?	Whether short-term and long-term effects diverge
Who else is affected?	Stakeholders and systems not mentioned in the problem
What's the equilibrium?	Where things settle after all responses play out

---

Phase 3: Multi-Order Effects

Trace consequences beyond the first order.

[S-N] ORDER_[1/2/3/N]:
     EFFECT: [what happens at this order]
     MECHANISM: [how the previous order produces this one]
     TIMESCALE: [when this effect manifests]
     CHANGES_ACTION: [does this change what you'd recommend? yes/no]
     CHANGES_RISK: [does this reveal a risk the surface missed? yes/no]

Stop Rules for Order Tracing

• Stop when effects become speculative (low confidence)
• Stop when effects no longer change the recommendation
• Stop when effects become too diffuse to be actionable
• But always go to at least order 2 — that's where sophistication begins

---

Phase 4: Cross-Domain Pattern Match

[S-N] PATTERN:
     THIS_PROBLEM_RESEMBLES: [known pattern from another domain]
     STRUCTURAL_SIMILARITY: [what features match]
     INSIGHT_FROM_PATTERN: [what the pattern tells you about THIS problem]
     TRANSFERRED_PREDICTION: [what the pattern suggests will happen]
     TRANSFER_RISK: [how might the analogy mislead?]

Common Cross-Domain Patterns

Pattern	Structure	Appears In
Arms race	Competitive escalation with no stable equilibrium	Pricing, features, regulation, talent wars
Tragedy of the commons	Individual rationality → collective failure	Resource use, technical debt, meeting culture
Principal-agent	Misaligned incentives between delegator and executor	Management, consulting, outsourcing
Network effects	Value increases with adoption	Platforms, standards, tools
Coordination game	Everyone benefits from same choice but no one moves first	Standards adoption, team practices
Selection effects	What you see is filtered by what survived	Survivorship bias, talent pools, market data

---

Phase 5: Synthesis

[S-N] SOPHISTICATED_READING: [what the problem actually looks like after deep analysis]
     KEY_INSIGHT: [the single most important thing that simple analysis missed]
     REVISED_ACTION: [what to do — how does it differ from the surface action?]
     RISKS_REVEALED: [risks invisible to simple analysis]
     OPPORTUNITIES_REVEALED: [opportunities invisible to simple analysis]
     CONFIDENCE: [high | medium | low — with reasoning]

SOPHISTICATED ANALYSIS
======================

SURFACE: [what simple analysis says]
SOPHISTICATED: [what deep analysis reveals]
KEY_INSIGHT: [the single most important structural feature]

ACTION CHANGE: [how the recommended action differs from simple]
RISKS_ADDED: [risks the surface missed]
TIMELINE: [short-term vs long-term divergence if any]

PATTERN: [cross-domain pattern if applicable]
CONFIDENCE: [level with reasoning]

---

Failure Modes

Failure	Signal	Fix
Complexity theater	Added layers but conclusion didn't change	Every layer must change the action, reveal a risk, or reveal an opportunity
Forced sophistication	Problem is genuinely simple but got over-analyzed	Say "the simple reading is correct" — that IS sophisticated when it's true
Pattern forcing	Applied a cross-domain pattern that doesn't actually fit	Check structural similarity specifically. Metaphors aren't patterns
Lost the thread	So many orders of effect that the core insight is buried	The synthesis must name ONE key insight. If you can't, the analysis is unfocused
Sophistication as hedge	"It depends on many factors" without specifying which	Name the factors, their values, and what changes at each value
No actionability	Deep analysis but no different recommendation	If the analysis doesn't change the action, say so. Maybe simple was right
Infinite regress	Third and fourth order effects that are pure speculation	Apply the stop rules. Go to order 2 minimum, but stop when confidence drops

---

Depth Scaling

Depth	Orders Traced	Cross-Domain	Structural Analysis	Synthesis
1x	2	No	Key variables only	One insight
2x	3	1 pattern	Full structural decomposition	Full synthesis
4x	4	2-3 patterns	Full + feedback loops + phase transitions	Full + scenario analysis
8x	All traceable	All applicable	Full + formal modeling if applicable	Full + decision tree

Default: 2x. These are floors.

---

Pre-Completion Checklist

• Surface analysis completed as baseline (not skipped)
• At least one hidden variable or mechanism identified
• Effects traced to at least order 2
• Every added layer of analysis changes the action, risk, or opportunity
• Cross-domain pattern checked (even if none applies)
• Synthesis names ONE key insight
• Confidence level stated with reasoning
• If simple reading was sufficient, said so explicitly

---

Integration

• Use from: complex problems, strategic decisions, situations where first-order thinking is dangerous
• Routes to: /ar (explore what follows from the sophisticated reading), /cmp (compare simple vs sophisticated recommendations)
• Complementary: /smpl (deliberately simple analysis — opposite end of the dimension)
• Differs from /smpl: smpl intentionally avoids depth; soph intentionally pursues it
• Differs from /dd: dd does deep dives into a topic; soph applies structural analysis techniques
• Differs from /certainty: certainty pursues maximum confidence; soph pursues maximum structural insight