MIS March 5 Formula Restoration Reference (Session 2 Brain Restore Prep)

Source: Copy of MIS_v7.1_gsheet - March 5, 3:01 PM (Drive ID 17SRUho-vIQ1J2Rb_xh6Uk7-PUD1orJeQIr64ctSSeEw)
Captured: 2026-04-28 ~12:15 AM EDT (overnight Session 17 prep, auto-mode subagent)
Sheet 1 (current/regressed): MIS_v7.1_gsheet (the live one)
Sheet 2 (this doc): the "March 5" backup

NOTE — Drive MCP returns evaluated values, not raw formulas. To get the actual
formula strings (e.g., the SACS weighted formula in col AZ, the Flow Strength formula
in col AC, the Snapshot gate logic in col S/T) Sam must open Sheet 2 in the browser →
click the cell → copy from the formula bar. This doc gives the structural target
+ observed value patterns that prove formulas are intact.

Critical caveat — this "backup" is NOT frozen. The Drive export shows live
2026-04-27 data (VIX=18.02, Total_Equity=16,238). That means Sheet 2's formulas
are pulling current GOOGLEFINANCE / live-data feeds — same as Sheet 1. The
"regression" Sam saw between Sheet 1 vs. Sheet 2 is not a formula corruption
in Sheet 2; it's that Sheet 1 has had columns/formulas deleted or replaced
while Sheet 2 still has the full original column set. Restoration target = copy
column structure + formulas FROM Sheet 2 INTO Sheet 1.

1. Momentum_Engine (col-by-col observed state)

Tab confirmed present in Sheet 2. Header found at line 677 of the markdown export.
Total observed columns: 56 data columns (matches the "wide" engine described in mobile-Claude doc).

Note on the AC vs AD column letter discrepancy: the original spec/mobile-Claude
doc names "Flow Strength Index" as col AC. In the actual Sheet 2 export Flow Proxy
sits at index 29 (AC) and Flow Strength Index sits at index 30 (AD). When Sam opens
Sheet 2 in browser, read the actual column letter from the header bar before
assuming AC vs AD. This doc preserves the order observed in the export.

Cols not visible to MCP that mobile-Claude flagged as relevant:
- col BG (30D Avg Volume) — not present in this 56-col export
- col BH (Volume vs Avg) — not present in this 56-col export

Either (a) those cols exist in Sheet 2 but the markdown exporter truncated them,
or (b) mobile-Claude's column-letter mapping was approximate. Verify in browser.

2. Reference_Rules (full key-value table)

CRITICAL FINDING — Reference_Rules tab IS NOT VISIBLE in the Drive markdown export.

The Drive MCP tool flattens all visible-data tabs into one markdown stream. Hidden
tabs, named-range-only tabs, and sheets with no traditional row structure (i.e.,
a key-value lookup table laid out as A: name | B: value) may not render in
the export. The export from Sheet 2 contains 944 lines, and zero of them match
"VIX_Low", "VIX_High", "Lean In", "Reversal Watch", "Macro Bias", "Trend Bonus",
or "Earnings Penalty" (the parameter names mobile-Claude listed).

Indirect evidence Reference_Rules exists in Sheet 2:
- Snapshot col 21 header reads literally: "Dynamic ATR multiple based on VIX bands in Reference_Rules (Regime Multiplier)". Observed value: 1.00 (VIX=18.02, regime=NORMAL).
- Snapshot col 22 header: "Visual tag for regime" — observed: NORMAL.
- Snapshot col 23: "ATR-Regime Stop (auto)" — pulls multiplier × ATR from Reference_Rules.
- Snapshot col 19: "Trade OK?" vs max % loss per position (your rule, e.g., 2.5% in Settings!B4) — references a Settings tab too (named-range Settings!B4).

What we KNOW from observed values about the rules:
| Observed parameter | Observed value | Source line |
|---|---|---|
| Regime threshold (VIX=18.02) | NORMAL | snapshot row col 22 |
| Regime multiplier @ NORMAL | 1.00 | snapshot col 21 sample |
| Max % loss per position | 2.5% (gate trips at >2.5%) | "REJECT (too loose / >2.5%)" appears 30+ times |
| HIGH regime threshold | VIX > ~19 (observed: 4/20–4/23 with VIX 18.87–19.5 all flipped to HIGH) | VIX history table at lines 65-72 |
| NORMAL → HIGH transition | "Regime change: NORMAL→HIGH" at VIX=19.5 | line 79 |
| HIGH → NORMAL transition | at VIX=18.71 | line 80 |

Restoration action: Sam must open Sheet 2 → click Reference_Rules tab → screenshot
or copy-paste the full A:B table (parameter name + value). MCP cannot reach this tab.
Same for the Settings tab (B4 = 2.5% confirmed by inference).

3. Snapshot tab (target column structure)

Tab confirmed present in Sheet 2. Header at line 362.
Observed columns: 28 data columns (NOT 57 as mobile-Claude doc claimed — possible the doc was conflating Snapshot with Momentum_Engine, OR the markdown exporter flattened helper cols).

Full column header list in order:

Sample row (AAPL):

Ticker=AAPL · Last=267.62 · Chg=−1.27% · Close_10dAgo=266.43 · Ret_10d=+0.45% · Close_30dAgo=252.62 · Ret_30d=+5.94% · ATR14=3.22 · Stops 1×/1.3×/2×: 264.40/263.43/261.18 · %Risk 1.20%/1.56%/2.41% · Shares 7/5/3 · $Risk_deployed=20.93 · Gate=OK (within 2.5%) · One-liner=AAPL: ATR14=3.22 | Stops(1x/1.3x/2x): 264.40/263.43/261.18 | %Risk: 1.2%/1.6%/2.4% | $Risk: $21 | Regime: NORMAL (1.00×) | AutoStop: 264.40 | Gate: OK (within 2.5%) · Regime_mult=1.00 · Regime=NORMAL · AutoStop=264.40 · Sector=Technology · Sector%=+0.27% · Risk_Per_Trade=0.01

Gate (col 19) logic — observed values:
- OK (within 2.5%) — when % Risk @ 1.3× ATR ≤ 2.5%
- REJECT (too loose / >2.5%) — when % Risk @ 1.3× ATR > 2.5%

Sheet 1 currently downgraded to ~14 cols (per the briefing prompt). Restoration
target = recreate cols 8–28 above, in this exact order, with the formulas Sam pulls
from Sheet 2's formula bar.

4. execution_playbook (decision-field structure)

Tab confirmed present in Sheet 2 — multiple instances of the "Execution Shortlist" header found (lines 771, 937 — likely the main exec table appears at top of tab and is repeated in views).

Column list (line 771):
- [x] Ticker ✓
- [x] Stock Name
- [x] Current % Change
- [x] Current Price
- [x] SACS (Alpha Conviction Score) ✓
- [x] VSM (Vol Stress Meter) ✓
- [x] Flow (Flow Proxy) ✓
- [x] Flow Strength Index ✓
- [x] RS_SPY_30D ✓
- [x] RS_Sector_30D ✓
- [x] RiskReward_13W ✓
- [x] IPQ (Intraday Price Quality) ✓
- [x] Action Flag ✓
- [x] Sector

Mobile-Claude doc fields → presence check:
- [x] Entry Signal — NOT explicitly named in this header — but implied by Action Flag column ("✅ In Sync – Momentum" = entry signal)
- [x] Stop — NOT in this 14-col header — but IS in the related Quiet Alpha tab (Final % Risk @ 1.3× ATR, line 902 col 11) and Snapshot (cols 9-11)
- [x] Risk% — present as "Final % Risk @ 1.3× ATR" in Quiet Alpha view (col 11)
- [x] Target% — implied by Stop (6%) / Target (10%) on Momentum_Engine cols AI/AJ
- [x] R:R — present as "RiskReward_13W" ✓
- [x] Shares — NOT in execution_playbook header — sits in Snapshot cols 15-17
- [x] SACS ✓
- [x] VSM ✓
- [x] Flow ✓ (Flow Proxy, but Trend not explicit — see note)
- [x] Trend — NOT a separate col in this header — but Momentum_Engine col Z (Momentum Regime) covers it
- [x] IPQ ✓
- [x] Action Verdict — "Action Flag" ✓ + "Trade Conclusion" column on Quiet Alpha (col 14) which carries the verdict text "✅ HIGH CONVICTION BUY (Risk: 1.56%)" / "🛑 REJECT: Risk Too High (4.28% > 2.5%)"
- [x] Reason — present in Momentum_Engine col AX
- [ ] Action Notes — not explicitly observed as a separate col; the "One-liner" in Snapshot col 20 may serve this purpose

Conclusion: all 14 mobile-Claude decision fields ARE represented in Sheet 2, but
they're spread across Momentum_Engine + Snapshot + execution_playbook + Quiet Alpha.
Restoration plan must rebuild those 4 tabs together — not just one.

5. Sector_Map

Anchor "Proxy ETF" CONFIRMED PRESENT in Sheet 2 — found at line 217 of export.

Header observed:
| Col | Header |
|---|---|
| A | Proxy ETF ✓ |
| B | Sector Group |
| C | Stock Avg % |
| D | 3-Day Momentum |
| E | Flow Strength |

Sectors aggregated (16 rows observed):

Sheet 1 is missing the "Proxy ETF" anchor label per the briefing. Restoration =
re-add col A header text + formulas mapping each sector → its proxy ETF ticker.
The "Flow Strength" col E was observed as blank in this export — possibly because
its formula references the Flow Strength Index from Momentum_Engine which has not
yet been re-evaluated, OR the col is genuinely empty and gets populated post-close.

6. Daily_Snapshot

Tab confirmed present — this is the main tab the Drive export landed on, lines 0-235.

Structure (observed top-to-bottom):

1. Regime header line (line 0):
   Regime_Label --> NORMAL | VIX Range (7d): 2.02 pts | Last Regime: NORMAL
2. Date/time stamp (line 2): 4/27/2026 | 10:20 PM
3. VIX block (lines 3-4): VIX price + VIX Day Hi
4. Index block (lines 5-7): DJI / IXIC / INX with % change
5. Equity block (lines 8-9): Total_Equity=16,238 | Daily_PnL=0 | Exposure%=blank
6. 🧠 Smart Reversal Opportunity table (lines 11-22, 10 rows): Ticker, Name, Price, %DayLow, 10D%, %DayHigh, RR_13W, Stop(6%), Target(10%)
7. Top 10 table (lines 24-35): Ticker, % Change, Live_Price, Sector, Name
8. Bottom 10 table (lines 37-48)
9. 3D % gainers + Compression flags + Momentum Regime (lines 50-61, 10 rows)
10. VIX history mini-table (lines 63-72): Date, VIX_Close, Max regime touched, Δ vs Prev Day
11. Regime Change log (lines 74-83)
12. Top 10 Strongest Upward Momentum view (lines 85-96)
13. Top 10 High-Scoring High Risk/Reward (lines 98-109)
14. Top 10 High-Scoring with outlier flags (lines 111-122)
15. Top 10 highest Vol % (30D) (lines 124-135)
16. 💥 Final Breakout Formula (lines 137-148)
17. 🚀 Sector Alpha Breakout (lines 150-161)
18. 🧘 Quiet Alpha Accumulation (lines 163-174)
19. 🥇 The Ultimate View (lines 176-187)
20. Low Volatility Reversal (lines 189-200)
21. Quiet Alpha (executable trade conclusion) table (lines 202-215)
22. Sector_Map block embedded (lines 217-235) — YES, sector table IS embedded inline in Daily_Snapshot
23. Compression Watch (lines 237-255)
24. … more views below …

Sector table embedded? YES. The Sector_Map content sits inline within
Daily_Snapshot at lines 217-235 as a ranked-by-perf table.

Top 10 / Bottom 10 structure: present (lines 24-35 / 37-48), 5-col layout:
Ticker, % Change, Live_Price, Sector, Name.

7. Restoration plan for Session 2

Suggested order:
1. Reference_Rules (manual, from screenshot) → unblocks Snapshot col 21-23 + regime logic everywhere
2. Universe tab integrity check (lines 466-642 export shows it intact w/ 175+ tickers)
3. Sector_Map (5-col rebuild, 16 sectors)
4. Momentum_Engine col-by-col formula port from Sheet 2 (the big one)
5. Snapshot col 8-28 formulas (ATR + risk gate)
6. execution_playbook & Quiet Alpha views (mostly QUERY())
7. Daily_Snapshot → mostly auto-rebuilds once 1-6 done

8. Open questions Sam will need to answer in Session 2 (formula bar required)

These cannot be answered from the Drive MCP markdown export — Sam must open Sheet 2
in browser and copy from the formula bar:

Source trail

• Sheet 2 (READ-ONLY source): Copy of MIS_v7.1_gsheet - March 5, 3:01 PM · Drive ID 17SRUho-vIQ1J2Rb_xh6Uk7-PUD1orJeQIr64ctSSeEw
• Sheet 1 (regressed, restoration target): MIS_v7.1_gsheet
• Companion docs:
• docs/MIS_FSE_ARCHITECTURE.md — root-cause table for the regression
• docs/MIS_PHASE2_BLUEPRINT.md — ATR-based restore architecture
• docs/MIS_AUDIT_2026-04-26.md — most recent audit findings
• docs/MIS_REVIEW_2026-04-27.md — reviewer notes
• Local artifact: C:\Users\ztrei\OneDrive\2. Hook Street\05. 2026 BH\docs\MIS_MARCH5_FORMULAS.md
• Generated: 2026-04-28 ~12:15 AM EDT (overnight Session 17 prep)
• Method: mcp__claude_ai_Google_Drive__read_file_content on Sheet 2 → markdown stream parsed locally; no writes attempted.

Generated by Claude Code subagent (general-purpose) for Session 17 overnight prep · Opus 4.7 (1M)