Breadth
IMPROVING
Breadth suffered a catastrophic collapse on 2026-03-13 (advance ratio 8.2%, 117 stocks down >4%, 252 stocks down >4% on weekly basis) and has since staged a sharp single-day recovery on 2026-03-17 (advance ratio 64.4%, only 2 stocks down >4%). However, structural breadth indicators remain deeply damaged: %>EMA20 is only 14.6%, %E20>E40 is 23.6%, and 52-week new highs are just 1 vs 63 new lows on 2026-03-17, confirming the recovery is narrow and not yet broad-based.
⚠ 52-week new highs collapsed to 0-1 on 2026-03-13/16/17 vs 63-84 new lows — structural damage not repaired
⚠ %>EMA20 at 14.6% and %E20>E40 at 23.6% on 2026-03-17 — majority of stocks still below medium-term averages
⚠ 2026-03-13 breadth collapse (8.2% advance, 117 stocks -4%+) leaves technical scars; single-day recovery insufficient to confirm trend reversal
Risk Flags
⚠ OI cap MEDIUM (score=+0.00) — reduce sizing; no HIGH conviction allowed until PCR>0.80 and VIX<18
⚠ VIX 19.79 (ELEVATED) — elevated fear; all conviction capped MEDIUM; BUY requires VIX below 18
⚠ TRIPLE_WALL @ 24000 [CE_RESISTANCE] — weekly+nxtweek+monthly+nxtmonth expiries stacked
⚠ TRIPLE_WALL @ 23500 [PE_SUPPORT] — weekly+nxtweek+monthly+nxtmonth expiries stacked
⚠ 52-week new highs collapsed to 0-1 on 2026-03-13/16/17 vs 63-84 new lows — structural damage not repaired
⚠ %>EMA20 at 14.6% and %E20>E40 at 23.6% on 2026-03-17 — majority of stocks still below medium-term averages
⚠ 2026-03-13 breadth collapse (8.2% advance, 117 stocks -4%+) leaves technical scars; single-day recovery insufficient to confirm trend reversal
🧠 STAGE 3 REASONING ▶ (click to expand)
Let me work through this systematically.
**1. OI Conviction Cap Check FIRST**
- cap=MEDIUM, score=+0.00
- divergence=NO
- This means: ALL HIGH conviction must be downgraded to MEDIUM. No HIGH conviction allowed.
- No AVOID_ALL, so watchlist can have candidates.
**2. Regime + VIX Impact**
- Regime signal: NEUTRAL (verbatim from Stage 1)
- Regime confidence: MEDIUM (verbatim from Stage 1)
- VIX: 19.79, ELEVATED zone (18-22), FALLING trend, ABOVE_20DMA (+2.48pts)
- VIX > 18 → no BUY conviction allowed per rules. Max sector view = HOLD.
- Wait, the sector rotation shows "BUY" for Exchange and Data Platform and Computers Hardware. But VIX rule says ELEVATED VIX (18-22): max sector view = HOLD. BUY only allowed if VIX < 18. So I need to downgrade those BUY views to HOLD.
**3. Candidates - Conviction Assignment**
After cap=MEDIUM check:
- TATAPOWER: CONFLUENCE=4, CLEAN, industry PASS (NEUTRAL), RS_5D=8.0 > 0 → Would be HIGH but cap=MEDIUM → MEDIUM
- AEROFLEX: CONFLUENCE=3, CLEAN, industry PASS (NEUTRAL), RS_5D=17.0 > 0 → Would be HIGH but cap=MEDIUM → MEDIUM
- CHENNPETRO: CONFLUENCE=3, CLEAN, industry PASS (NEUTRAL but SM=MARKDOWN), RS_5D=17.4 > 0 → Would be HIGH but cap=MEDIUM → MEDIUM. Wait, industry_sm=MARKDOWN for Refineries & Marketing. But industry_pass=true. DISTRIBUTION/MARKDOWN phase → view = AVOID per sector rules. But this is the stock's industry, not the sector rotation view. The industry_pass=true means it passed the industry gate. Let me keep it as MEDIUM but note the MARKDOWN industry.
- PFC: CONFLUENCE=3, CLEAN, industry PASS (NEUTRAL), RS_5D=4.3 > 0 → MEDIUM
- AUROPHARMA: CONFLUENCE=3, CLEAN, industry PASS (NEUTRAL), RS_5D=2.8 > 0 → MEDIUM
- COALINDIA: CONFLUENCE=3, CLEAN, industry PASS (NEUTRAL), RS_5D=7.0 > 0 → MEDIUM
- KTKBANK: CONFLUENCE=3, CLEAN, industry PASS (NEUTRAL), RS_5D=5.0 > 0 → MEDIUM
- PAISALO: CONFLUENCE=3, SUSPECT, industry PASS but SM=DISTRIBUTION → MEDIUM (confluence=2 threshold for MEDIUM, but SUSPECT + DISTRIBUTION industry → should be LOW → avoid)
Actually: MEDIUM — CONFLUENCE=2, SUSPECT or partial industry. PAISALO has CONFLUENCE=3 but SUSPECT classification and industry SM=DISTRIBUTION. The conviction rules say MEDIUM for CONFLUENCE=2, SUSPECT or partial industry. PAISALO has CONFLUENCE=3 but SUSPECT + DISTRIBUTION industry. I'll assign MEDIUM but note the risks.
Wait, re-reading: "MEDIUM — CONFLUENCE=2, SUSPECT or partial industry". PAISALO has confluence=3 but SUSPECT. The rules say HIGH requires CLEAN classification. SUSPECT → can't be HIGH. With confluence=3 and SUSPECT, it would be MEDIUM. But industry SM=DISTRIBUTION is a concern. I'll include as MEDIUM with caveats.
- TICL: CONFLUENCE=3, SUSPECT, industry PASS (NEUTRAL) → MEDIUM
- JINDALPOLY: CONFLUENCE=2, CLEAN, industry FAIL (MARKDOWN) → LOW → avoid
- DCMFINSERV: CONFLUENCE=2, CLEAN, industry PASS (NEUTRAL) → LOW → avoid (confluence=2 is MEDIUM threshold but LOW conviction → avoid)
Wait, re-reading conviction rules:
- HIGH: CONFLUENCE=3 or 4, CLEAN, passing industry, RS_5D > 0 (only when cap=HIGH)
- MEDIUM: CONFLUENCE=2, SUSPECT or partial industry
- LOW → avoid
So CONFLUENCE=3 CLEAN with passing industry and RS_5D > 0 would normally be HIGH, but cap=MEDIUM downgrades to MEDIUM.
CONFLUENCE=2 with CLEAN and passing industry → MEDIUM (since it meets the MEDIUM criteria)
CONFLUENCE=2 with failing industry → LOW → avoid
JINDALPOLY: confluence=2, industry FAIL → LOW → avoid
DCMFINSERV: confluence=2, industry PASS → MEDIUM? But the conviction rules say MEDIUM is for CONFLUENCE=2. Let me include it.
Actually wait - DCMFINSERV has RSI=81.1 (overbought) and 5D_RET=65.2% (extreme momentum). This is suspicious. But the rules don't explicitly exclude it. I'll include as MEDIUM with strong caveats.
Actually, let me reconsider PAISALO - industry SM=DISTRIBUTION. The sector rules say DISTRIBUTION/MARKDOWN phase → view = AVOID. So PAISALO's industry is DISTRIBUTION → should be avoided.
Let me finalize watchlist:
- TATAPOWER: MEDIUM (downgraded from HIGH due to cap)
- AEROFLEX: MEDIUM
- CHENNPETRO: MEDIUM (note MARKDOWN industry)
- PFC: MEDIUM
- AUROPHARMA: MEDIUM
- COALINDIA: MEDIUM
- KTKBANK: MEDIUM
- TICL: MEDIUM (SUSPECT, missing 200D)
- PAISALO: avoid (DISTRIBUTION industry)
- JINDALPOLY: avoid (confluence=2, industry FAIL)
- DCMFINSERV: MEDIUM or avoid? confluence=2, industry PASS (NEUTRAL), CLEAN → MEDIUM
Actually I'll keep the watchlist to the stronger candidates. DCMFINSERV with 65% 5D return and RSI 81 is extremely overbought - I'll include it but with strong caveats.
**4. Risk Flags**
- VIX 19.79 in ELEVATED zone (94.8th percentile)
- FII consecutive selling streak: 2 days (not ≥ 3, so no flag for that)
- TRIPLE_WALL at 24000 (CE resistance) across all 4 expiries - within... 24000 vs spot 23581 = 419 pts away = 1.78% away. The rule says "within 1% of spot". 1% of 23581 = 235.81. 24000 - 23581 = 419 pts > 235 pts. So NOT within 1%. 23500 - 23581 = -81 pts, which IS within 1% (81 < 235). So the PE support at 23500 is within 1% of spot. That's a TRIPLE_WALL within 1%.
- cap=MEDIUM (not AVOID_ALL, so no specific flag needed for that)
- Breadth: March 13 showed only 8.2% advance ratio (41 advances vs 459 declines) - severe breadth deterioration
- FII 2-day selling streak (not ≥ 3)
**5. Breadth Reading**
- March 13: catastrophic breadth (8.2% advance, 117 stocks down >4%)
- March 17: recovery (64.4% advance, only 2 stocks down >4%)
- But structural indicators remain weak: %>EMA20 only 14.6%, %E20>E40 only 23.6%
- Trend: IMPROVING (from 8.2% to 64.4% over the week)
**Action Plan**
Since cap=MEDIUM (not AVOID_ALL), the action plan should note the MEDIUM cap, re-entry conditions for full BUY conviction (PCR > 0.80 already met at 0.984, VIX < 18 needed), and risk management.
Let me now construct the JSON.
For sector rotation - VIX is ELEVATED so BUY views must be downgraded to HOLD:
- Exchange and Data Platform: SM=MARKUP, would be BUY but VIX > 18 → HOLD
- Computers Hardware: SM=ACCUMULATION, would be BUY but VIX > 18 → HOLD
For delivery signal:
- CLEAN candidates: TATAPOWER (spike_ratio=2.3), others without spike_ratio
- TRAP candidates: ANGELONE (4.7), SPARC (3.19), HINDPETRO (2.72), etc.
- Smart money: NEUTRAL to mild accumulation given mixed signals
Let me write the JSON now.
For stacked walls: TRIPLE_WALL at 23500 (PE support) is 81 pts below spot = 0.34% away → within 1% → risk flag.
TRIPLE_WALL at 24000 (CE resistance) is 419 pts above spot = 1.78% away → NOT within 1%.
Risk flags:
1. "TRIPLE_WALL PE support at 23500 is 81 pts (0.34%) below spot 23581 — pinning risk, breach could accelerate selling"
2. "VIX 19.79 at 94.8th percentile (52W) — ELEVATED zone caps all conviction to MEDIUM; BUY conviction requires VIX < 18"
3. "FII net short -240,775 index futures with 2-day selling streak; cash selling -4,219 cr — institutional-vs-retail divergence"
4. "Breadth collapse on 2026-03-13: only 8.2% advance ratio (41/500), 117 stocks down >4% — structural damage not yet repaired (%>EMA20 = 14.6%)"
5. "OI cap MEDIUM (score=+0.00): mixed OI structure prevents HIGH conviction across all candidates"
For the action plan - since cap=MEDIUM (not AVOID_ALL), I should still mention re-entry conditions for full conviction:
"NEUTRAL regime with MEDIUM OI cap (score=+0.00) limits all positions to MEDIUM conviction. Spot at 23581 is pinned between TRIPLE_WALL support at 23500 (81 pts below) and TRIPLE_WALL resistance at 24000 (419 pts above), creating an asymmetric range. Prefer MEDIUM conviction longs in CLEAN stocks (TATAPOWER, AEROFLEX, COALINDIA) with strict stops below 23500. For full BUY conviction to be restored, weekly PCR must remain above 0.80 (currently 0.984 — met), VIX must drop below 18 (currently 19.79 — not met), and FO quadrant must not be SHORT_BUILDUP (currently LONG_BUILDUP — met). Until VIX < 18, maintain reduced position sizing and avoid chasing overbought names (JINDALPOLY RSI=89.2, DCMFINSERV RSI=81.1)."
Let me finalize the watchlist. I'll include the strongest candidates:
1. TATAPOWER - MEDIUM (confluence=4, CLEAN, best setup)
2. AEROFLEX - MEDIUM (confluence=3, CLEAN, at 200D high)
3. COALINDIA - MEDIUM (confluence=3, CLEAN)
4. KTKBANK - MEDIUM (confluence=3, CLEAN)
5. PFC - MEDIUM (confluence=3, CLEAN)
6. AUROPHARMA - MEDIUM (confluence=3, CLEAN, RSI=75.7 elevated)
7. CHENNPETRO - MEDIUM (confluence=3, CLEAN, but MARKDOWN industry)
8. TICL - MEDIUM (confluence=3, SUSPECT)
For avoid list, I'll include all Stage 2 avoided stocks plus PAISALO (DISTRIBUTION industry), JINDALPOLY (confluence=2, industry FAIL), DCMFINSERV (confluence=2, extreme overbought).
Actually DCMFINSERV has confluence=2, industry PASS (NEUTRAL), CLEAN → technically MEDIUM. But RSI=81.1 and 5D_RET=65.2% is extreme. I'll include it in watchlist with strong caveats or avoid. Given the extreme overbought nature, I'll put it in avoid.
Let me finalize:
Watchlist (MEDIUM only):
1. TATAPOWER
2. AEROFLEX
3. COALINDIA
4. KTKBANK
5. PFC
6. AUROPHARMA
7. CHENNPETRO
8. TICL
Avoid:
- All Stage 2 avoided stocks
- PAISALO (DISTRIBUTION industry)
- JINDALPOLY (confluence=2, industry FAIL/MARKDOWN)
- DCMFINSERV (confluence=2, extreme overbought RSI=81.1)
For delivery signal top picks, I'll use TATAPOWER (spike_ratio=2.3, GREEN MA) as the main clean delivery spike.
Let me now write the full JSON carefully.