/ search · ? ASKB · 1-6 timeframe · A analytics · G signals · I india · R refresh · T toggle TV
MONITOR…
ALERTSBSV · FXPOS · WIRE · GEO · DHAN
NEXThigh / medium
DHAN OFF
RELATED
loading desk…
DIAG · CONNECTIVITY SELF-TESTrelays · CDN · streams · widgets · suppressed errors · R to re-run
TV · LIVECNBC-TV18 · Bloomberg · C = next channel · T = back
loading GEO desk…
HELP · FUNCTIONS · KEYS · SOURCESone registry drives the menubar, the palette and this page
S&P 500
CAPITALCOM:US500
◈ TECHNICALS
◈ PROFILE
◈ FINANCIALS
◈ SCREENER
◈ HEATMAP
◈ CALENDAR
◈ OVERVIEW
MARKET MONITOR 0 INSTRUMENTS · CLICK TILE
▲ GAINERS · LOSERS · ACTIVE
◆ US MARKET SNAPSHOT
NEWS WIRE · US LIVE · 0 SRC
LOADING NEWS…
NIFTY 50
BSE:NIFTYBEES
◈ TECHNICALS
◈ PROFILE
◈ FINANCIALS
◈ SCREENER
◈ HEATMAP
◈ CALENDAR
◈ OVERVIEW
MARKET MONITOR 0 INSTRUMENTS · CLICK TILE
▲ BSE GAINERS · LOSERS · ACTIVE
◆ INDIA MARKET SNAPSHOT
NEWS WIRE · INDIA LIVE · 0 SRC
LOADING NEWS…
EUR/USD
FX:EURUSD
◈ TECHNICALS
◈ FX HEATMAP
◈ COT POSITIONING
◈ FX SCREENER
◈ CALENDAR
◈ FX STORIES
◈ OVERVIEW
FX & METALS MONITOR 0 PAIRS · CLICK TILE
◆ CROSS-RATE MATRIX
◆ FX & METALS SNAPSHOT
NEWS WIRE · FX LIVE · 0 SRC
LOADING NEWS…
◈ OPTIONS POSITIONING · MAX PAIN —
SPOT (DERIVED)—
MAX PAIN—
PUT/CALL OI—
LOADING CHAIN…
READ: price is drawn toward max-pain into expiry as dealers hedge.
Big call OI above spot = resistance. Big put OI below = support.
◈ SHORT VOLUME · SQUEEZE FUEL —
SYMBOL SHORT%—
SHORT VOL—
TOTAL VOL—
LOADING FINRA…
READ: FINRA daily short volume, all US venues. Sustained >60%
= aggressive shorting; forced covering is the squeeze fuel.
◈ INSIDER FILINGS · FORM 4 —
LOADING EDGAR…
READ: real time from SEC EDGAR — the same instant institutions get it.
Clusters of insider buys at one company are the highest-signal free data there is.
◈ ACTIVIST STAKES · 13D / 13G —
LOADING EDGAR…
READ: 13D = >5% stake with intent to influence, filed within 10 days.
Far timelier than 13F, which is 45 days stale by law.
◈ OFF-EXCHANGE · DARK POOL PROXY —
OFF-EXCH SHARE—
UNIVERSE—
SESSION—
COMPUTING…
READ: high off-exchange share = size worked away from the lit tape.
Quiet chart + heavy dark volume often precedes a repricing.
◈ MATERIAL EVENTS · 8-K WIRE —
LOADING EDGAR…
READ: 8-K is filed before the press picks it up. Same primary
source the terminals read — you are not behind here.
COVERAGE: COT / Short-Squeeze / Insider panels are
US-market only — verified: CFTC lists 0 India contracts,
FINRA carries only US-listed ADRs. Correlation & Session
analytics include USD/INR. For Indian equities use the INDIA DESK (charts, technicals,
screener, news all work).
◈ COT POSITIONING EXTREMES —
LOADING 239 WEEKS OF CFTC…
READ: z-score = distance from its own 3-yr norm. Beyond ±2σ = crowded.
Backtested: 150 extreme signals since 2022 → 65% reverted within 12 weeks.
The % badge is each contract's own measured hit-rate — BITCOIN is 22%, so ignore its flags.
◈ SHORT-SQUEEZE RADAR —
LOADING 10 DAYS OF FINRA…
READ: short% high AND rising AND volume expanding.
Backtested (45 sessions, 4 walk-forward windows): flagged names saw short interest fall
96% of the time vs 48% baseline — covering is real and predictable. But forward
price edge was ≈0 on median (one +2830% outlier faked a big mean).
Use as a watchlist filter, not a buy signal.
◈ INSIDER CLUSTERSUNVALIDATED—
SCANNING EDGAR…
NOT BACKTESTED — this panel is descriptive. The clustering logic is sound in principle but its forward edge has never been measured here, so treat it as research, not signal. READ: one insider filing is noise. Multiple insiders at one company
in the same window is the strongest free equity signal. Click to open the filings.
◈ CROSS-ASSET CORRELATIONUNVALIDATED—
COMPUTING CORRELATIONS…
NOT BACKTESTED — correlation breakdowns are descriptive. No study here shows a breakdown predicts returns. READ: when a stable correlation breaks, something structural is
happening. Δ column = change vs the prior window of equal length.
◈ FX SESSION ANALYTICSUNVALIDATED—
COMPUTING SESSIONS…
NOT BACKTESTED — session statistics are descriptive; they describe past behaviour, not future edge. READ: tells you when a pair actually moves. Most traders have no
idea which session sets the daily high/low or which weekday carries the range.
◈ WATCHLIST & ALERTS 0 ACTIVE
NO ALERTS SET
SAVED LOCALLY in your browser — survives reload. Alerts are checked on
every data refresh and fire a desktop notification plus an on-screen toast.
NSE EOD · all data end-of-day (~6:30 PM IST), routed via public relay—
◈ FII vs CLIENT POSITIONING —
FII NET (IDX FUT)—
Z-SCORE—
CLIENT NET—
LOADING NSE PARTICIPANT DATA…
READ: FII net short + Client net long = institutions positioned
against the crowd. Daily & same-day — better than CFTC COT (weekly, 3-day lag).
Caveat: FII index shorts are often hedges, not directional bets.
◈ FII / DII CASH FLOWS —
FII NET—
DII NET—
COMBINED—
LOADING FLOWS…
READ: the defining Indian dynamic — FII selling absorbed by
SIP-funded DII buying. Both buying = strong bullish. FII sell + index holds = DII absorbing.
◈ DELIVERY ANALYTICS —
MARKET (10D MEDIAN)—
REGIME—
CHURN RELIABILITY—
LOADING 10 SESSIONS OF DELIVERY DATA…
RE-VALIDATED 25-Aug-2026 on 150 fresh NSE sessions (14-Jan → 25-Aug 2026,
118 walk-forward windows).CHURN came back WEAKER than the
original study — median −0.81 pts @20d with 57% of windows negative, versus the
original −1.82 pts / 74%. Net of ~0.37% round-trip costs the edge is marginal.
The signal still leans the right way, but treat it as a filter, not a trade trigger.
All are 20-day signals — none work intraday.
◈ BULK & BLOCK DEALS —
LOADING DEALS…
READ: named counterparties. Repeat buyers = same entity
accumulating across sessions — India's version of insider-cluster logic.
◈ F&O BAN & MWPL —
LOADING BAN LIST…
READ: banned stocks allow only position reduction — forced
unwinding regardless of view. Watch names approaching 95% MWPL before the ban hits.
◈ MARKET BREADTH & SECTORS —
LOADING INDICES…
READ: 30-day sector leadership shows where institutional money is
rotating. PE/PB vs own range gives valuation context, not just price.
HOW IT WORKS: US markets close 01:30 IST —
eight hours after India closes and before it reopens. This regresses NIFTY's
overnight gap ln(Open/PrevClose) on the last completed US session.
It predicts direction far better than magnitude: on quiet nights
(<0.3% move in EPI) the hit rate is only 66%, but above 1.5% it reaches 82.7%.
The badge greys out when the signal is not worth acting on. GIFT Nifty would
beat this comfortably but is not on any free feed. Not investment advice.
◈ CORRELATION LAB — STOCKS · EQUITIES · ETFs
—
UNIVERSEWINDOWMETHODidle
HIGHLIGHT |ρ| ≥OFF
INSTRUMENTS—
COMMON SESSIONS—
AVG PAIR ρ—
PC1 (MARKET FACTOR)—
ABSORPTION RATIO—
MARKET STATE—
Pick a universe and press ▶ COMPUTE.
Prices are pulled per symbol and joined on dates every instrument traded —
never by row position, which is the error that silently corrupts most
correlation tables.
METHOD: log returns
r = ln(P_t / P_t-1), date-aligned across the full universe.
Pearson for linear co-movement, Spearman on ranks to expose outlier-driven
readings, EWMA (λ=0.94, RiskMetrics) for the current regime, and
downside-conditional ρ because correlations rise exactly when you need them low.
Full derivations in CORRELATION-METHODOLOGY.md. Not investment advice.
DECISION ENGINE · combines every signal into a ranked shortlistCAPITALRISK %IDLE
◆ RANKED SHORTLIST — WHAT TO DO
—
UNIVERSE SCANNED—
PASSED FILTERS—
REGIME—
SIGNAL STATE—
DATA DATE—
Press ▶ RUN ENGINE to score the market.
Engine idle. It will scan the full liquid universe, score each name across
every signal this terminal collects, and return only names that clear the bar —
with direction, entry zone, stop and position size.
HOW IT DECIDES: each name is scored on the signals that
were actually backtested (delivery churn, delivery accumulation, volume anomaly,
short-volume pressure) and penalised for regime conflict and F&O ban status.
Conviction is the weighted sum. Only names above the cutoff are shown. Size comes
from your capital × risk% ÷ ATR-based stop distance — so a wider stop gets a
smaller position automatically.
◆ LIVE SCORECARD — IS THIS ENGINE ANY GOOD?
—
CALLS LOGGED0
RESOLVED0
HIT RATE—
AVG EDGE—
No calls logged yet. Every time you run the engine, its shortlist is saved with
today's price. Re-open the terminal later and press ⟳ GRADE OPEN CALLS —
it marks each one win or loss against the live price. Nothing is hidden.
WHY THIS EXISTS: a signal with no track record is an opinion.
This panel is the terminal grading itself in public, losses included. 20-day horizon.
◆ WHAT CHANGED SINCE LAST SESSION
—
Run the engine once to take a baseline snapshot. From the next run onward this
panel diffs today against your last session and reports only what is new:
names entering or leaving the shortlist, direction flips, and fresh F&O bans.
READ: the story is in the change, not the level.
A stock that just entered churn matters more than one that has been there a week.
◆ MY PORTFOLIO — RISK ON MY ACTUAL MONEY
—
One holding per line — SYMBOL, QTY, AVGPRICE
(NSE names for India, tickers for US). Saved locally in your browser only.
BOOK VALUE—
OPEN P&L—
CONCENTRATION—
PORTFOLIO BETA—
1-DAY VaR 95%—
-5% INDEX DAY—
Enter your holdings above. The terminal will value the book, compute real
correlation between your positions (to catch the case where five holdings are
secretly the same bet), estimate portfolio VaR from actual return history, and
flag any holding that today's signals are firing against.
NOT ADVICE. VaR is historical-simulation on ~6 months of
daily returns and assumes tomorrow resembles the recent past — it fails precisely
in the crashes you most want protection from. Treat it as a floor, not a ceiling.
S&P 500 CAPITALCOM:US500
—
◈ RISK LENS
—
ANN. VOLATILITY—regime
VaR 95% · 1 DAY—historical sim
VaR 99% · 1 DAY—worst 1-in-100
MAX DRAWDOWN—1-year peak→trough
CURRENT DRAWDOWN—from 1y high
BETA vs INDEX——
POSITION SIZER. ATR-based: risking
%
of
→ — units,
stop —
(— ATR×1.6)
◈ FACTOR EXPOSURE
—
READ: 6-month daily-return correlation to each macro
driver. High |ρ| means the name is really trading that factor, not its own story.
◈ STRESS SCENARIOS
—
METHOD: beta-implied move plus the worst actual
drawdown observed in the last year. Beta-scaling assumes the relationship holds
in a crash — historically it strengthens, so treat these as a floor.
◈ CROWD SENTIMENT · StockTwits
—
——
MESSAGES SCANNED—latest stream
TAGGED BULL/BEAR—explicit sentiment
CONTRARIAN CAVEAT: retail sentiment is not a
timing signal. Extreme readings (>85% either way) have historically marked
local exhaustion more often than continuation. Bot spam is filtered. Never
trade this alone.
The engine pulls NSE exchange filings, live movers, 8 Google News topic searches, 13 newswires and 14 X accounts — then detects
which words are bursting versus their baseline, clusters near-duplicate
stories into single events, and then counts how many independent sources
confirm each one.
A story carried by four outlets is evidence. The same story from one account
is a rumour — and it will be labelled as such.
◉ BURSTING NOW
Run a scan to see which terms are
spiking against their own baseline.
◈ TICKERS IN THE NEWS
Detected company mentions mapped to NSE
symbols. Click to open the stock.
⚙ SOURCE HEALTH
Which feeds answered, which failed.
URGENCY (0–100) = min(2.2, burst z ÷ 5) × log₂(1 + sources) × event weight × recency × 42.
Every term is visible and arguable — unlike a vendor black box. Corroboration is the
core mechanic: Dataminr's real edge is confirming an event across independent
sources, and that is exactly what the ⚑ badge counts. Single-source items are shown
but explicitly marked UNCONFIRMED. This is a news-latency tool, not a
backtested signal — it contributes to COMMAND at low weight. Not investment advice.
▦ GLOBAL MAPcross-asset transmission · world mapidle
positive ρnegative ρbeta only (spurious)overnight → next sessionregime break● pulsing = market open · dimmed = closed · grey band = night sessionhalo = |z| ≥ 1.5σ
LIVE
PRESS ▶ LOAD MAP
📋 AUTO ROUNDUP
Press ▶ LOAD MAP to pull 29 assets
(25 drawn + 4 that feed the India strip) with one year of history — 2 batch calls, about a second.
🌡 RISK APPETITE
—
⚡ BIGGEST DISLOCATIONS z vs own 60d vol
—
🔥 REGIME SHIFTS & DIVERGENCE
—
⇢ STRONGEST OVERNIGHT EDGES
—
ℹ METHOD
Log returns, date-joined (never array position).
ρ over 120 sessions; regime check compares 120d vs 20d.
MST = minimum spanning tree on √(2(1−ρ)) — thins ~70 raw edges to 24
while keeping every node connected.
PARTIAL controls for S&P 500 to expose links that are just market beta.
Overnight arrows use lag-1: yesterday's US close → today's Asia session.
Node colour is z-score vs each asset's own 60-day vol, so a 0.8% bond move
outranks a 2.5% crude move.
ASSOCIATION, NOT PROVEN CAUSE.
Correlations break exactly when you need them — that is what the regime-shift panel is for.
EOD daily bars. Not investment advice.
🇮🇳 TOMORROW'S NIFTY OPEN———
4-factor model · OOS R² 0.26 · not advice
LIVE WIREall markets · auto-refetchidle
LOADING LIVE WIRE · FREE RSS · AUTO-REFRESH 60s
SESSION BIAS · 12H
decoding headlines…
NEXT CATALYSTS
fetching economic calendar…
DATA DESKofficial sources · every pane shows SOURCE + AS-OFFA
SECURITYdigits 1-9 jump · Backspace = DES · ‹GO› = Enter
MARKETSworld indices · index futures · sectors · commodities · money markets
ECO · ECONOMIC EVENTScalendar · actual vs consensus vs previous · revisions · surprise index · what moved · central banksECO · WECO · ECSU · ECST ‹GO›
loading calendars…
DHAN · BROKER LIVE FEEDMCX gold · silver · crude · NSE currency futures · indices · 5-level depth · tick capture around releasesDHAN · DOM · TAPE ‹GO›
…
RETAIL · SENTIMENT & POSITIONINGsocial trending · velocity alerts · retail FX & gold long/short · fear & greed composites · NSE client vs FIIRETAIL · TREN · BSV · FXPOS · CROWD ‹GO›
…
◆ ASKBconversational research · rules-based agents over the live wire, bias, calendar & prices · no external AI
›ENTER
▣ COMMANDevery panel, one verdict — pulls live state from all 9 desks—idle
AWAITING DATA
Press ▶ RUN FULL SWEEP to load every desk
HOW THE SCORE WORKS: each desk contributes a signed
vote weighted by how much evidence backs it — the India delivery CHURN signal
(backtested, p<0.0001) carries far more weight than StockTwits sentiment
(no validated edge). Votes are summed, normalised, and shown as a
−100…+100 bias. Conflicting signals reduce the score rather than cancel
silently, and the confidence read drops when desks disagree. Anything
unloaded is excluded, never assumed neutral. Not investment advice.
RELIANCE NSE · vs NIFTY 50WINDOWidle
◉ WHAT ACTUALLY MOVED THIS STOCK — RETURN ATTRIBUTION
—
METHOD: two-factor decomposition
r = α + β_mkt·r_market + β_sec·r_sector + ε. The sector factor is
orthogonalised against the market before fitting, so the same move is never
counted twice. Contributions sum exactly to the total return — verified to
9 decimal places. This is Aladdin's signature visual: it answers
"was this the market, the sector, or the company?"
◉ HISTORICAL SCENARIO REPLAY
—
WHY THIS BEATS BETA-SCALING: instead of assuming
"if the index falls 10%, this falls 10×β", it shows what the stock
actually did in real crises. Windows were validated against index data —
any candidate that wasn't genuinely a stress event was dropped.
◉ RETURN DISTRIBUTION & VaR TAIL
—
VaR 95%—
VaR 99%—
CVaR 95%—
SKEW—
READ: the shaded left tail is where your losses live.
CVaR (expected shortfall) is the average loss given you're already
in the worst 5% — the number that matters when things go wrong. Negative skew
means crashes are sharper than rallies.
◉ RISK THROUGH TIME — ROLLING VOL & BETA
—
60d annualised volatility60d rolling beta—
WHY IT MATTERS: a single beta over a year hides the fact
that it may have moved from 0.7 to 1.4. Risk is a time series, not a
constant — expanding vol means your position size should be shrinking.
◉ TIME TO INSIGHT — PLAIN-ENGLISH SUMMARY
—
BlackRock's own stated goal is reducing "time to insight" —
the system tells you what matters instead of making you hunt. These bullets are
generated from the computed metrics above, not from an opinion.
Not investment advice.
CNBC-TV18 · LIVE
CONNECTING…
Q…
SELF-DIAGNOSTIC · CONNECTIVITY
Running…
WIRE…
SOURCE—RELAYS…WIRE—DHANOFFDIAG…LAYOUTDEFAULT/ cmd ? ask ESC back 1-9 menu W dock B rail F1 helpINITIALISING