Choose a strategy to explore
Each card shows native PF from its saved report. Open a card for rules, exits, a hypothetical price example and the recorded closed-position curve.
Saved native backtests
All figures below are simulated, net of their modeled costs. PF and drawdown use the native reports; TLBS counts partial exit slices, while the other five use full positions.
| Strategy / TF | Window | Capital | Native PF | Native count | Net P&L | Return | Equity DD |
|---|
KST: exact cutoff changes the accounting
Native eventual report: 38 positions, PF 1.997, +$2,142.30. One exit occurred at 02:00 HK on 10 Nov, after the exclusive cutoff.
Before cutoff: 37 closed positions, PF 2.034, +$2,180.82. Add the open position’s saved mark of −$6.17 → strict-year marked P&L +$2,174.65. This does not turn native PF 1.997 into a PF>2 result.
TLBS: partial exits are not extra positions
Older Arm B: 462 native exit slices, PF 1.909. Reconstructed entry groups: 231 positions, PF 1.992. Grouping changes PF because wins/losses can net within a position.
Grouping assumes no pyramiding and identical entry time, side and entry price. Native P&L $29,064.04 versus rounded ledger $29,064.07 is a $0.03 rounding difference.
The five-sleeve accounting sum
+$9,284.66 / +9.28% on five separate $20k sleeves ($100k total), including the KST cutoff mark. Closed trades alone total $9,290.83; pooled closed-trade PF 2.357 over 137 positions.
This is not a native netted single-account portfolio test. Full portfolio equity drawdown is unavailable. The recorded closed-balance DD of $607.30 omits open-position excursions. No six-strategy total is calculated.
Earlier/later halves of the November year
Exit-date split at 10 May 2026 HK. Reused history, not walk-forward validation. DD in this table is reconstructed closed-position balance DD, not native intrabar equity DD.
| Strategy | Earlier n | PF | Net P&L | Balance DD $ | Later n | PF | Net P&L | Balance DD $ |
|---|
Walk-forward: what is actually available?
A genuine test selects rules on training data, freezes them, then evaluates the next unseen period. Repeating that cycle can assess how the selection process travels through time.
Three settings per strategy; already inspected during earlier research.
Already inspected too. Can support a retrospective exercise, not a fresh holdout claim.
No future trades or unverified delivery logs are filled into this report.
Retrospective one-step train/test exercise
Declared for this report on12Nov 2026: among each strategy’s three saved settings, require training PF>2, ≥30positions and positive net P&L; choose the highest training PF (then net P&L to break ties). Read the saved next-year result at that exact setting. No new native reruns.
| Strategy | Training result | Next-year result at selected setting | Interpretation |
|---|
What this exercise tells us
Only FRAMA pivot width 1 meets the declared training gate. Its next-year PF falls from 3.269 to1.803 (34 positions, +$1,567.79, +7.84%, native equity DD5.61%). The next year stays profitable but fails the PF>2 target. This does not justify changing the saved baseline pivot width 2.
WICK, KST, FLUX and CLIP have no qualifying training candidate under this gate. TLBS has only same-window news-boost A/B evidence here; a sequential train/test result is unavailable.
All settings are retained
No hidden winner selection. WICK parameter = target R; KST = ATR stop multiple; FLUX = directional volume threshold%; FRAMA = pivot wing length; CLIP = ATR stop multiple.
| Strategy | Setting | Period | n | PF | Net P&L | Return | Equity DD |
|---|
Actual forward-validation status: all six pending verification
Saved alert activation and synthetic Telegram tests are configuration/delivery evidence, not forward trading returns. This HTML has no verified current forward ledger, forward PF or broker-executed P&L. Do not label missing data as zero profit or zero trades.
Next protocol: align current source/inputs/feed and costs; freeze all settings and selection gates; retain every paper signal, fill, partial and exit; reconcile complete new periods; report counts, net P&L, intrabar equity DD and cost stress before drawing conclusions.
How to read these numbers
Profit factor
PF = sum of positive net trade P&L ÷ absolute sum of negative net trade P&L. PF2 means twice as many winning dollars as losing dollars in that sample. It does not mean a200%return or a guaranteed future edge.
Changing partial-exit grouping changes the unit of observation. Do not compare TLBS slice PF to whole-position PF without that distinction.
Three different P&L views
Native eventual: includes recorded exits after the entry window.
Closed by cutoff: only positions finished before the boundary.
Strict-year marked: closed P&L plus the saved mark of open exposure, not an executable liquidation quote.
Curves show completed-position net accumulation; native equity DD additionally reflects modeled open excursions.
Quick glossary: ATR, EMA, R and drawdown
ATR14: average true range over14bars, a volatility measure. EMA200: exponential moving average over200bars, used here as a direction filter. R: planned entry-to-stop price distance; a3R target is three times that distance before fill/cost effects. ROC: rate of price change; KST combines several ROC horizons. FRAMA: fractal adaptive moving average; its responsiveness changes with price structure. Order block: this indicator’s algorithmic zone, not proof of an institution’s resting orders. Equity drawdown: peak-to-trough decline including modeled open-position excursions. POC: volume-profile point of control; not required by these six saved trading overlays.
Five-sleeve execution and risk assumptions
VANTAGE:XAUUSD,1h/2h/4h standard bars; $20k each; planned0.5%equity risk capped$100,95%notional cap,1×leverage,0.1native chart-unit increments. Units are not independently verified MT5 lots. Assumed$.10/unit/side commission and30ticks market/stop slippage; no separate variable spread, swap, latency or executable bid/ask spread. Standard OHLC emulator, no Bar Magnifier; CLIP one-tick target penetration. Warmup calculation history requested from2Jan2024; actual entry window is the November365days. Gaps and slippage can exceed planned risk; signal-close brackets can alter actual RR.
TLBS assumptions and current-source caveat
Older1m retest24Sep 2025–24Sep 2026, $100k, commission$.10/unit/side,1tickslippage. ArmA news boostOFF: nativePF 1.819,462slices, net+$25,829.56(+25.83%),equityDD3.05%/$3,762.60. ArmB ON is the headline sample. Exact export clock timezone, full-year warmup and intrabar-fill-setting parity are not independently established by this audit. Current Telegram source and historical retest hashes differ. Rules/defaults shown are current saved source descriptions, not certification of the alert snapshot or old-test parity.
Evidence limits and failed research
All displayed market history was previously inspected; small samples, parameter searches and repeated selection matter. The five-sleeve sweep includes15settings/30native tests;10alternatives did not pass every original advancement condition. No latest-year native doubled-cost result is established here for the existing five, nor a matched current-window stress for TLBS. Mechanical fixed-fill friction deductions in an older audit are not equivalent native reruns. Subsequent DeepCharts-inspired continuation tests had0PF>2passes in8runs; they are separate research and not a seventh deployed method.
Arithmetic reconciliation establishes recorded counts and P&L consistency. It does not establish data authenticity, realistic fills, robustness, profitability in unseen data or live readiness. Volume estimates in WICK/FLUX are not actual exchange bid/ask order flow. Legacy ACTIVE/DEFAULT labels are historical configuration labels.
Source manifest and downloads
Built12Nov 2026 from saved observations through10Nov 2026 and the older TLBS report. No new backtests or live service checks for this guide.
Download report data Five-year accounting Recorded-ledger audit 30 native sweep records