Market closed·Opens 9:15 AM
Brutal honesty · transparency framework

We publish what hurts.

This page is the opposite of marketing. It shows every loss, every drawdown, every regime where the engine has negative edge, and every failure mode we are still chasing. If a number on the dashboard looks better than what you see here, the dashboard is wrong — not this page.

Sample dataThe numbers on this page are illustrative until the live broker ledger is wired in. We're publishing the shape of brutal honesty first — losses, drawdowns, failed signals, calibration — so you can audit the framework before the live data replaces it. Live wire-up is tracked in the public roadmap.
Append-only ledgerOut-of-sample onlyNo survivorship biasCalibration audited weekly
01

Loss reporting

Loss ledger · last 7 closed losersnet -6.8R · append-only

Every loss is recorded with its reason code before the P&L tile updates. The engine cannot delete or re-sort this list — overrides and stop hits get equal billing.

IDDateSymbolPlaybookRReasonEngine note
L-22842026-05-28NIFTY 24800 CEVWAP reclaim-1.0Rstop hitBreadth flipped negative at 09:52, exit at hard SL.
L-22812026-05-28BANKNIFTY 51200 PEOpening drive fade-0.7Rtime stopTTL expired flat; no follow-through within 4 bars.
L-22772026-05-27RELIANCE FUTORB long-1.4RslippageMarket order, queue jumped 9bps vs intended.
L-22742026-05-27NIFTY 24750 PETrend pullback-1.2Rregime flipVIX spiked +8%, classifier switched mid-trade.
L-22702026-05-26HDFCBANK FUTMean reversion-0.9Rspread widenedBid/ask widened to 41bps; exit took 3 prints.
L-22662026-05-26NIFTY 24900 CEPower-hour squeeze-1.0Rstop hitTextbook stop, no excuses.
L-22612026-05-23INFY FUTVWAP reclaim-0.6Rmanual overrideUser closed early at -0.6R; setup would have hit T1.
02

Drawdown analytics

Drawdown analyticspeak-to-trough · published, not pruned
Worst DD (R)
-11.2R
Worst DD (%)
-8.9%
Avg days underwater
11d
Recovery rate
4/4
Underwater curve · last 4 drawdowns (stitched)
Y-axis: % below prior equity peak. The chart never lies — it goes down before it goes up.
DD-072026-04-142026-04-29-6.2%11dElection-week volatility; classifier mis-tagged regime for 2 sessions.
DD-062026-02-192026-03-04-4.7%9dFOMC week — playbooks degraded; risk auto-throttled at -4R.
DD-052025-12-092026-01-06-8.9%19dWorst drawdown on record. Holiday liquidity + thin tape.
DD-042025-10-212025-10-28-3.1%5dRoutine variance — within expected band.
03

Signal grade history

Signal grading · post-trade audit675 graded · last 90 sessions

Every closed signal is auto-graded A→F by an independent post-trade evaluator — not the same model that armed the trade. D and F grades are surfaced to the user with the same prominence as A grades.

A
142 · 21%+184.6R
B
211 · 31%+96.2R
C
188 · 28%+12.4R
D
96 · 14%-41.3R
F
38 · 6%-67.8R
A· All confirmations green, regime aligned, vol in band
B· Minor confirmation miss or 1 stale factor
C· Mixed signals, marginal edge, often break-even
D· Forced setup, weak conviction, post-hoc rationalised
F· Override of engine veto or chase outside playbook
04

Performance breakdown

Performance breakdownnegative cells shown — not hidden

Both losing playbooks and losing time-of-day buckets are kept visible. We do not cherry-pick the best slice and call it the strategy.

By playbook
VWAP reclaim
184t · 58% · +86.1R
Opening drive fade
142t · 51% · +34.2R
Trend pullback
121t · 56% · +47.8R
ORB long/short
96t · 49% · +8.1R
Power-hour squeeze
81t · 44% · -12.6R
Mean reversion
63t · 41% · -18.4R
By time of day (IST)
09:15–10:00
142t · 47% · -4.2R
10:00–11:30
218t · 61% · +94.3R
11:30–13:30
96t · 44% · -11.1R
13:30–14:45
138t · 54% · +38.7R
14:45–15:30
93t · 49% · +7.4R
05

Confidence calibration

Confidence calibration · reliability diagramECE 5.1pts · target < 8

If the engine says 70% confidence, did it actually win 70% of the time? The diagonal is the truth; bars under the diagonal mean we were overconfident in that bucket.

025507510014%28%47%64%76%predicted win prob (%)
BucketPredictedActualΔn
0-20%12%14%+2pt41
20-40%31%28%-3pt88
40-60%50%47%-3pt173
60-80%71%64%-7pt142
80-100%88%76%-12pt56
The 80–100% bucket is currently 12pt overconfident — flagged for model review on the next training run.
06

Risk-adjusted metrics

Risk-adjusted metricswith caveats, not without
Sharpe (annualised)
1.42
Sortino
1.98
Calmar
1.07
Profit factor
1.34
Expectancy / trade
+0.18R
Tail ratio (95/5)
1.21

Sharpe (annualised):Computed on daily R; assumes 252 sessions. Drops to 0.91 if you include weekends as zero-return days.

Sortino:Penalises downside vol only; flatters strategies with rare large losses. Read alongside max DD.

Calmar:Net return / max drawdown. < 1 means the worst DD exceeded your annual return.

Profit factor:Gross win / gross loss. Above 1.30 is acceptable for a high-frequency intraday system, not great.

Expectancy / trade:Mean of all R-multiples after costs. Net of typical 6bps round-trip slippage.

Tail ratio (95/5):Right-tail vs left-tail magnitude. Close to 1 means wins and losses are symmetric.

07

Market condition attribution

Market regime attributionwhere the edge lives — and where it doesn't

Total P&L decomposed by regime classifier label. Two of four regimes have negative edge and are either disabled or force-throttled — we publish them anyway.

Trending — high RVOLedge
184 trades61% win+142.3R
Engine sweet spot. Most A-grades originate here.
Range-bound — normal volneutral
246 trades52% win+38.1R
Marginal edge; size scaled down by 35%.
Chop — low RVOLnegative-edge
142 trades41% win-47.6R
Disabled by default. Manual override only.
Event/spike — VIX > 22negative-edge
81 trades38% win-52.2R
Hard veto when VIX prints > 25.
08

When NOT to trade

When NOT to tradehard vetoes · enforced in code

These are not suggestions. Each rule is a server-side gate — the engine refuses to arm a signal, and the manual ticket is disabled, until the condition clears.

VIX > 25 — circuit-breaker risk

Implied vol regime invalidates intraday playbook win rates by ~14pts. Engine refuses to arm.

Expiry day after 14:30 IST

Theta decay and OI unwind dominate price; the playbooks have negative edge in this window.

Spread > 50 bps

Execution slippage eats > 0.3R on average. Hard veto, no override.

First 5 minutes of the open

Tick noise > signal; classifier confidence falls below 55%. Wait for the opening-drive window.

After 3 consecutive losses

Mandatory 60-second cool-off. Revenge-trade detection halts new entries.

Latency > 400ms

Quote staleness invalidates stops. Engine pauses arming until probe recovers.

09

Failure transparency

Top failure modes · last 30dpublished with mitigation

The most common ways DyadScalp has failed, ranked by occurrence. Each row carries the mitigation we shipped — or admits we have not fixed it yet.

Failure mode30d countTrendMitigation shipped
Slippage on market orders > 8bps31Default order type switched to LIMIT-with-2bps-tolerance.
Stop hit within 2 bars of entry24Added micro-structure filter for false breakouts.
User override of engine veto17Friction dialog with cool-off; logged to audit chain.
Regime flip mid-trade11Trailing reclassifier now runs every 15s vs 60s.
Data gap > 2s6Failover to secondary feed; positions held with widened stop.
10

Trust reinforcement

Our commitments to youhow we keep ourselves honest
Every fill is logged

Order ID, timestamp, intended price, fill price, slippage — written to an append-only ledger before P&L is shown.

Losses are published with reasons

We do not curate the loss feed. Stop hits, time stops and manual overrides all appear with the same prominence as wins.

Walk-forward, not back-fit

All published win rates are out-of-sample. In-sample numbers are watermarked and never used in marketing.

Drawdowns are not deleted

Worst-case drawdown history is permanent. We show DD-05 (-8.9%) on the home page, not just the recovered curve.

Confidence is calibrated

When the engine says 70%, it should win 70% of the time. Calibration error > 8pts triggers a model review.

When we don't know, we say so

Uncertainty bands are shown on every projection. No point estimates without a confidence interval.

If any line on this page is later contradicted by reality, we change the line — not the data.
If we lied on this page, the audit chain would catch us before you did · DyadScalp
Risk Notice

DyadScalp is a decision-support terminal. It does not custody funds, manage portfolios, or offer investment advice. Orders are routed to your registered broker under your own credentials. Index-options scalping carries the risk of total loss on each trade — server-side risk gates cap the bleed but do not prevent it. Past simulation and replay performance is not indicative of live results.