The active decision path: spot direction + conviction drive entries; this is what's happening right now
trigger
Direction
—
Conviction
—
Chop eff
—
EMA
—
Waiting for data…
Why no trade — today's stand-aside timeline
No stand-aside events logged yet
Live Candidates
Highest-scoring signals right now — what's close, what's blocking entry
Floor —5s refresh
Waiting for data…
Open Trades
No open positions
System Health
Pivot Levels
No pivot data
Index · 5-min Candles · last 1 hour
Intraday P&L
Realized: —Unrealized: —
Market News AI · Nifty impact only
Loading…
Near Miss Log
Candidates within 15 pts of entry floor — sorted by regret score (missed move × closeness)
30s
9:15
15:30
Loading…
🤖
Live Trading AI
Ask why you're not trading · settings · near misses
Why no trade?Top candidate blockerCooldown statusMost active gateNear miss reviewScore gap fix
Live Score Breakdown
Top candidates — every component that built or dragged the score
Waiting for data…
EOD Controls
Manual run with guardrails
Choose a trading date, set time range (default 00:00-23:59), then load the report or run AI analysis.
Decision Logs
-
Cycles: -
Trades Entered
-
Closed: -
Realized P&L
-
Win rate: -
AI Status
-
Model: -
Best Trade
—
Worst Trade
—
Per-Trade Critical Analysis
Based on spot price movement vs option direction
Select a date and click Refresh to load analysis.
Trades
Load a report to see trades.
Top Blockers
Why signals were rejected today
Load a report.
CE vs PE
Load a report.
Activity by Time Slot
Load a report.
Decision Funnel
How many candidates survived each gate
Load a report.
AI Analysis
No analysis loaded
Run AI Analysis to see insights for this date.
Saved Analyses
Latest stored EOD analyses
🤖
Trading AI
EOD Analysis Assistant
Why no trade?Top candidate blockerNear miss analysisMost common gateWinners vs losers
Trade Performance Window
All trades entered in selected IST dates
Choose a date range to track trades and P&L.
Trades
-
Closed: - | Open: -
Realized P&L
-
Unrealized: -
Total P&L
-
Avg closed trade: -
Win Rate
-
W/L/F: -
Score Filter Context
Configured threshold and custom score filter
No data loaded
Trade Performance Detail
No data loaded
Live vs Expected PnL
—
Is the live paper track behaving like the backtest? Actual realized PnL vs the backtest's expected range (±σ noise band). Inside the band = normal.
Actual cumulative
-
live realized
Expected (backtest)
-
mean × days
Deviation
-
inside ±1σ = on track
Sample size
-
need ~45+ to confirm edge
Cumulative PnL vs expected range
ActualExpected mean±1σ / ±2σ band
Daily realized PnL
Each day's swing vs the +₹283/day the model expects on average.
Read before trusting the lines
The backtest edge is +₹283/day but daily noise is ±₹4,266/day — randomness is ~15× the signal. Over a short window the actual and expected totals are statistically the same number. This chart catches a broken system (actual falling out the bottom of the band); it cannot confirm the edge works until ~45+ trading days.
The numbers
Capital Ledger
Epoch: -
Ledger restarted clean — every stat below only counts trades and adjustments from the epoch date forward.
⚠ Reset Capital Adjustments
This will permanently delete all manual capital adjustments from the database.
Trade P&L history and the ledger epoch/starting capital are not affected. This cannot be undone.
Current Capital
-
Total return: -
Starting Capital
-
Since -
Realized P&L (life)
-
Adjustments: -
Peak / Drawdown
-
Max drawdown: -
Win Rate
-
W/L/F: -
Profit Factor
-
Avg win / loss: -
Expectancy / Trade
-
Trades (life): -
Day Win Rate
-
Streak: -
Runway: -
Equity Curve
No data loaded
Solid line = closing capital · dashed line = running peak · red fill = drawdown from peak · blue line = replay backtest with current settings (spot pts × lot size)
Daily Result Calendar
-
Daily P&L Distribution
Realized P&L per trading day
Capture & Utilization
How much of the favorable move gets kept
MFE Capture
-
Sample: -
Avg Capital Utilization
-
Peak trade: -
By Exit Reason
By Option Type
By Weekday
By Entry Hour (IST)
Daily Ledger
No data loaded
Capital Adjustments
Auditable manual cash movements or corrections
Use adjustments instead of editing the running balance directly, so history stays auditable.
Recent Adjustments
Latest 50
Parameter Backtest
Select a window where you believe a trade should have happened. The engine tests 192 parameter combinations and shows what settings would have entered and what profit was achievable.
Run Summary
WHAT BLOCKED ORIGINAL ENTRY
AI ANALYSIS
Top Parameter Combinations
#
Threshold
Pivot Req
Mkt Ctx
Penalty Wt
Multi-Strike
Entry Time
Entry LTP
Strike
Hypo Score
Peak PnL%
Stopped PnL%
Duration
VIX Regime Accumulator
Collect 3+ backtest runs per regime to unlock AI suggestions for that regime.
AI Parameter Suggestion
Suggested Settings — Preview & Apply
Trade Replay
Select a past trade to see the 1-min Nifty EMA state at entry and whether the Spot EMA Gate would have blocked it.
SIGNAL SCORING — WHAT WOULD HAVE ENTERED (±30 min window)
Threshold
Pivot
Mkt Ctx
Entry Time
LTP
Score
Peak PnL%
Stopped PnL%
Spot Directional Engine — Replay
range:
Measured in Nifty spot points — no premium, no spread. Replay = live behavior. On first load, inputs are filled from your saved Settings. Change any value and replay to test; click Save to Settings to apply.
PARAMETERS (all 13 knobs)
Conv ≥ entry gate
Confirm bars hold N bars
Min body pts
Chop eff ≥ 0–1
Chop lookback bars
Break buffer pts
Fractal K swing width
Exit bars flip confirm
EMA fast period
EMA slow period
Min swing pts
Trend lookback bars
Wick frac 0–1
Hover any label for explanation · Replay uses current values · Save pushes to live
Chart — engine calls on Nifty 1-min (▲ long / ▼ short entry · ● exit)
Scoreboard
Directional Calls
✓ Auto-pick — most robust combo
Parameter Sweep — ranked by Train net (Validate = held-out check)
Trust Validate (out-of-sample) + Profit days, not just Train net. A combo big on Train but weak on Validate is overfit. Apply pushes the 7 params live.
Blank values are normal. Levels never carry over automatically to another day.
Pre-Session Pattern Profile
Auto-runs at 08:50 IST daily
Analyses last 30 days of trade outcomes by VIX regime, score bucket, MultiStrike, time slot, level penalty, and session position. Suggests and applies targeted setting adjustments via AI. Requires OpenAI to be enabled. Always fails safe — errors never block trading.
Trade Lifecycle Settings
State-machine entry, validation, expansion, profit lock, and re-entry
These settings control the new deterministic trade-state flow separately from the older generic threshold set.
Volatility Settings
India VIX regime engine, hysteresis, and per-regime profile overrides
These settings control how India VIX adjusts confirmation, stop loss, minimum hold, cooldown, and defensive daily trade limits for new trades.
Slow Trend Mode
ranked_v2 persistence bonus for drift moves
These settings reward repeated aligned index drift, option premium rise, and mild OI compression without turning slow trend into a hard entry rule.
Runtime Settings
Settings are editable and apply immediately. Secrets are not exposed here.
Nifty Chart
● LIVE
Overlays
The nearest two calculated supports and resistances, VWAP, EMAs and active-trade risk remain visible by default.
No-trade zone
EMA 20
EMA 50
Trade entry / exit
Structural stop
Waiting for the latest engine evaluation…
Market diagnostics —
15-bar directional efficiency and EMA spread
Signal Review
Look at each signal as if the chart stopped at that bar — would you take this trade? Mark Good (strong setup, clear institutional flow), Skip (unclear, no edge), or Fade (wrong direction, this is the trade to take the other side). These labels train the ML model — 10 minutes a night improves signal quality every day.