What it is

Stock Intelligence Score (SIS) ek 0–100 composite hai. Yeh seven underlying layer-state outputs ko ek number mein roll-up karta hai. Jaise ek doctor health checkup ka score deta hai — blood pressure, sugar, BMI, ECG sab ko ek roll-up health-index mein convert karte hain.

Sabse important baat shuruaat mein spelt out kar deni chahiye: SIS ek layer-state descriptor hai, directional prediction nahi. SIS 85 ka matlab "is stock ke seven analytical layers ka aggregate state clean / aligned / well-formed hai" — not "yeh stock chadhega." Tietiy ka founding principle hi yeh hai (cockpit CLAUDE.md §5.3).

Doosri equally important baat — har SIS ke saath ek score_breakdown ship hota hai. Aap dekh sakte ho 78 mein se kitna trend se aaya, kitna RS se, kitna setup se. Number anchor karne ka risk hota hai, isi liye breakdown mandatory hai (Phase-revision Option B, locked 2026-06-01).

The 7 components and their weights

Each component is itself a 0–100 score (or normalized to one). Weighted sum gives the final SIS.

  • Trend state (weight 20) — UP states score 100, TRANSITIONAL ~50, CHOPPY ~30, DOWN ~10 (asymmetric — we're not symbol-short-friendly yet).
  • Relative Strength (weight 20) — RS percentile rank directly used (rank 87 → 87 points).
  • Volume state (weight 15) — EXPANSION at trend-direction = 100, NORMAL = 60, DRY = 30, CLIMAX context-dependent.
  • Structure quality (weight 15) — Cleanness of recent pivots, SR zones' touch counts (multi_touch_3p pulls down, fresh zones pull up).
  • Sector rotation (weight 10) — Is this stock's sector currently leading or lagging the broader market?
  • Setup quality (weight 10) — Is a recognized pattern (pullback, flag, breakout) currently in formation? More about formation quality than success odds.
  • Data freshness (weight 10) — Was today's data clean? Freshness gate pass? Min-history (250 bars) satisfied? Any caveats reduce score.

How tietiy computes it

SIS lives in the setup producer (which has expanded over cockpit phases to be a roll-up). It only runs once all the seven ingredient producers have completed for that symbol — its consumes list is long. If any ingredient produced a caveat (e.g., insufficient history, fixture data), SIS publishes its own caveat and the score gets a lower freshness component.

Final SIS = Σ (component_score × weight) / 100. Score breakdown:

score_breakdown: {trend: 18, rs: 17.4, volume: 12, structure: 10, sector: 7, setup: 6, freshness: 10} → SIS = 80

So you can always see which layers carried the score. If 90% of an SIS came from "freshness + sector rotation," it's a weaker descriptive read than if it came from "trend + RS + volume + structure."

What this site shows you

Stock page top par a big SIS chip — number + a small ring showing composition. Hover karne par score_breakdown table dikhta hai (the seven contributions). Markets page pe "today's SIS leaders" aate hain. SIS never appears alone — hamesha breakdown ke saath, taaki anchoring kam ho.

How to read it

  • SIS 80+ with balanced breakdown = textbook well-aligned descriptive state. Multiple layers consistent.
  • SIS 80 with single component dominant = brittle. If 50 of the 80 came from RS alone, RS shift = score collapse. Check breakdown.
  • SIS 40–60 = mixed state. Some layers say one thing, others say another. This is the ambiguity zone — most stocks live here most days.
  • SIS < 40 = several layers misaligned or weak. Watch only descriptive read.
  • Never compare SIS across regimes blindly. SIS 75 in a Bull regime ka meaning alag hai than 75 in a Bear regime — base rates flip.

What this does NOT tell you

  • Price target. SIS is a state descriptor, not a magnitude. SIS 90 doesn't mean "10% higher coming."
  • Probability of trade success. Validation Gate (Phase 8) abhi nahi fire hua. Until ≥30 post-cockpit trades accumulate in operator's journal, SIS-vs-outcome correlation unverified hai.
  • Action. SIS describes; operator decides. Tietiy never auto-suggests a trade based on SIS.
  • Continuation. Yesterday SIS 80, today 78 — yeh predictive nahi hai. SIS daily fresh compute hai, score time-series ko trend ki tarah treat karna risky hai.

The honest disclosure. Composite scores invite anchoring — log ek number dekhke "achha number = good stock" assume karte hain. Tietiy composite scores allow karta hai (Option B revision) only with mandatory score_breakdown sibling, taaki anchoring kam ho. Until Phase-8 Validation Gate fires, SIS ki actual outcome-correlation unverified rehti hai. Use as context only.