Accuracy audit fixes: alerts carry validated per-TF trail plans + 13 more

Five-agent audit of the signal path; all fixes are delivery/measurement
side. Regression-proven behavior-neutral on validated signals: the 5yr
M15 replay reproduces the pinned book exactly (+210.65R PF 1.35
WR 57.8% MaxDD 33.90R). Tests 215 -> 221.

- TradeSetup carries trail_arm_r/trail_dist_r stamped from the producing
  config; Telegram trail plan reads them (was showing default 0.6/0.75
  instead of the validated M15 arm 0.70 / M5 dist 0.85)
- Alert dedup: fingerprint = (signal, quality, direction); entry movement
  compared against last-ALERTED entry/risk (old SL-dist bucket churned on
  ATR drift -> near-duplicate alerts every candle)
- Fresh fetches enforce the closed-bar boundary check with one 3s grace
  retry, then warn (feed lag no longer silently analyzes the prior bar)
- Backtest max DD includes the zero baseline (opening losing streaks
  were undercounted; pinned M15 DD unchanged)
- Replay loads the regime-filter TF when the H4-ADX gate is on (was a
  silent no-op for M5/M15 primaries)
- clear_cache under _cache_lock; dead ROC zero-cross branches removed
  from both engines (provably unreachable); RSS dates via
  parsedate_to_datetime (named zones no longer stamped UTC); event risk
  grades from the soonest event; .env inline comments stripped;
  auto-scan skips NEUTRAL like the breach path; report_to_dict honors
  the weight-0 skip rule and emits strategy/trail params; PB funnel
  counts spread_gate/accepted; session-filter fail-open paths log loudly
- Docs corrected to match validated code (H&S neckline = lower peak,
  exhaustion anchors farthest tracked swing) with re-validation warnings

Claude-Session: https://claude.ai/code/session_01R5soRUeRBi8kHMB99csrJy

CLAUDE.md

@@ -130,7 +130,7 @@ Results land in `TechnicalAnalysis.indicators: Dict[str, IndicatorSignal]` (fiel
 | `macd` | 1.5 | Histogram = MACD line − signal line. Cross above 0 → BUY (conf 55+); accelerating (hist>prev>0) → STRONG_BUY; mirrored down; positive-but-fading / negative-but-fading → NEUTRAL (35); flat → NEUTRAL (30). Confidence scaled by |hist| vs its 20-bar mean |
 | `bollinger` | 1.5 | Position = (close−lower)/(upper−lower). >1.0 SELL (65), <0.0 BUY (65), >0.8 SELL (45), <0.2 BUY (45), else NEUTRAL (30). Squeeze = width < 0.75 × 20-bar avg width → +10 conf. Emits `bb_upper/lower/mid` |
 | `adx` | 0.5 | Wilder ADX/DI (ties zero BOTH DMs — fixes a systematic bearish nudge). ADX ≥ 25 → BUY/SELL by DI dominance, else NEUTRAL. Conf 40/65/80 by ADX band. ADX slope (5-bar diff): <25 with slope > 3.0 → "TREND EMERGING" +15 conf; also strengthening/weakening adjustments. Sets `_adx_slope` |
-| `roc` | 1.5 | 10-period %change. Positive & rising → BUY (STRONG if >2× threshold=2%); negative & falling → SELL; zero-cross → BUY/SELL (65); else NEUTRAL. **Directionally redundant with MACD/Keltner/EMA (0.8–0.94 corr) but its weight is load-bearing: zeroing it backtested 13R WORSE — composite thresholds are calibrated with it. Do not remove without retuning thresholds** |
+| `roc` | 1.5 | 10-period %change. Positive & rising → BUY (STRONG if >2× threshold=2%); negative & falling → SELL (zero-crosses land in these branches by definition — a separate cross branch was provably dead and removed 2026-07-23); else NEUTRAL. **Directionally redundant with MACD/Keltner/EMA (0.8–0.94 corr) but its weight is load-bearing: zeroing it backtested 13R WORSE — composite thresholds are calibrated with it. Do not remove without retuning thresholds** |
 | `keltner` | 1.5 | EMA20 mid ± 1.5×ATR(10). TTM squeeze = BB fully inside KC. Squeeze released (was→not) + MACD hist sign → STRONG_BUY/STRONG_SELL (80); squeeze active → NEUTRAL (60); close above/below channel → BUY/SELL continuation (65); inside → NEUTRAL (30). Emits `kc_upper/lower/mid` |
 | `divergence` | 1.5 | RSI + MACD-hist vs price over 40 bars. Swing points (window 3) must align within ±2 bars to pair up; last two pairs: price LL + indicator HL = bullish, price HH + indicator LH = bearish. Any found → BUY/SELL/NEUTRAL-if-both, conf 70, value = count |
 | `candlestick` | 1.0 | Last bar: Doji (body<10% range), Hammer/Hanging Man (lower wick>2×body), Shooting Star/Inverted Hammer, Engulfing (vs prev bar), Marubozu (body>85%). Majority vote → signal, conf 50; +20 if pattern sits within 0.5 ATR of a swing S/R level |
@@ -185,7 +185,7 @@ Two entry points:
 6. **`_compute_final_signal`** (bonuses → event cap → penalties → thresholds → grade):
    - Bonuses: MTF agreement → +10 conf and `score = 0.6×primary + 0.4×mean(all TF scores)`; news sentiment aligned (|sent|>0.2) → +5 (gated by `news.sentiment_affects_signals`, False by default); in a *directional* trend, RSI in pullback zone → +5, BB position favorable → +5 (mixed `trending` regime gets NO directional bonus).
    - Event risk `danger`: warning always; confidence hard-capped at 30 ONLY if `news.event_risk_caps_signals` (False by default — XAU-tuned).
-   - **Penalty system (diminishing, replaces old additive stacking):** collect, sort by size desc, apply with decay ×1.0, ×0.5, ×0.25…, total capped at 30. Penalties: `mtf_disagreement` 15; `news_conflict` 10 (gated by `sentiment_affects_signals`, off by default — warning still emitted); `event_caution` 15 (config-gated); `counter_trend` 15; `extreme_vol` 10; `ema200_near` 5 (price within 0.5% of EMA200 — env `FB2_EMA_LO` overrides); `ema200_far` 5 (>2.5%); `exhaustion` 8 (price >3.5 ATR beyond nearest swing); `no_pullback` 8 (>2.5 ATR from EMA21 in trend) / `weak_pullback` 5 (>2.0); `rsi_extended` 5 (RSI>65 with BUY / <35 with SELL in any strong-ADX regime, both directions in mixed `trending`); `bb_overextended` 8 (BB pos >0.85 BUY / <0.15 SELL).
+   - **Penalty system (diminishing, replaces old additive stacking):** collect, sort by size desc, apply with decay ×1.0, ×0.5, ×0.25…, total capped at 30. Penalties: `mtf_disagreement` 15; `news_conflict` 10 (gated by `sentiment_affects_signals`, off by default — warning still emitted); `event_caution` 15 (config-gated); `counter_trend` 15; `extreme_vol` 10; `ema200_near` 5 (price within 0.5% of EMA200 — env `FB2_EMA_LO` overrides); `ema200_far` 5 (>2.5%); `exhaustion` 8 (price >3.5 ATR beyond the FARTHEST of the 3 tracked swing levels — fires earlier than a nearest-swing rule would; calibrated as-is in every validated book); `no_pullback` 8 (>2.5 ATR from EMA21 in trend) / `weak_pullback` 5 (>2.0); `rsi_extended` 5 (RSI>65 with BUY / <35 with SELL in any strong-ADX regime, both directions in mixed `trending`); `bb_overextended` 8 (BB pos >0.85 BUY / <0.15 SELL).
    - **Exhaustion hard block**: ≥2 exhaustion-tagged penalties ({exhaustion, no_pullback, rsi_extended, bb_overextended} → `report.exhaustion_count`) AND momentum fading (ROC sign against score, or MACD signal against score) → confidence = 0.
    - Confidence clamped [0, 95]. **Final signal**: score ≥35 & conf ≥55 STRONG_BUY; ≥25 & ≥50 BUY; mirrored sells; else NEUTRAL.
    - **Quality grade**: |score|≥40 & conf≥65 & MTF → **A+**; ≥33 & ≥55 & MTF → **A**; ≥28 & ≥50 → **B**; ≥25 & ≥45 → **C**; ≥15 → **D**; else **F**.
@@ -218,7 +218,7 @@ Two entry points:
 
 **D. `_try_chart_pattern` (XAU_CP)**: detection in `chart_patterns.py` — 6 classical patterns (double_bottom/double_top, inv_hs/hs_top, bull_flag/bear_flag), swing-point based (strict local extrema, window `cp_swing_window`=3), all thresholds in ATR units, scanned over the last `cp_lookback`=72 closed bars. Trigger = FIRST close beyond the neckline/boundary since the pattern formed (one breakout → one signal, no repaint); reversal patterns are invalidated by a new extreme beyond the pattern. No composite-score gate BY DESIGN. Gates: MTF non-opposition (6f mirror); SELL patterns respect `sell_vol_regimes`, `sell_blocked_hours_utc` (bar-close hour), M5 RSI floor; M5 blocks `m5_blocked_vol_regimes` both ways; ATR ≥ $1. Geometry: SL = pattern anchor ∓ 0.5 ATR (min $8); TP1 = measured move (pattern height), TP2 = 1.5×height; RR ≥ 0.5; spread gate $9. Sets `strategy="XAU_CP"` (global trailing params), grade min B, marker `"XAU CHART PATTERN: …"` at reasons[0]. Funnel in `SignalAggregator.cp_funnel`.
 
-`TradeSetup` fields: pair, direction, entry, SL, TP1, TP2, RR, risk/reward pips, pip_value, `strategy` ("GENERIC", "XAU_PB", or "XAU_CP" — selects trailing params; XAU_CP uses the global pair).
+`TradeSetup` fields: pair, direction, entry, SL, TP1, TP2, RR, risk/reward pips, pip_value, `strategy` ("GENERIC", "XAU_PB", or "XAU_CP" — selects trailing params; XAU_CP uses the global pair), `trail_arm_r`/`trail_dist_r` (stamped from the PRODUCING config at generation so per-TF factory overrides reach formatters; `None` = legacy unstamped → display falls back to platform defaults).
 
 ---
 
@@ -235,7 +235,7 @@ Two entry points:
 
 - Twelve Data SDK (`TDClient.time_series(...).as_pandas()`), symbols normalized to `"XAU/USD"` form; renames lowercase columns → capitalized; sorts oldest-first; localizes UTC.
 - **Rate limit**: 55 credits/min sliding window (`_throttle`), blocks with sleep when exhausted.
-- **In-memory cache**: key `{symbol}_{tf}`, TTL = candle duration (M5 300 s … H4 14400 s); a cached frame is evicted and refetched if it fails the freshness check **or the boundary check** (`_has_latest_closed_bar`: the frame must contain the most recent fully closed bar, weekend-exempt — a wall-clock TTL alone would serve mid-candle snapshots at the next boundary and scans would run one bar behind). Cache and throttle are lock-protected (one provider is shared across the bot's thread pool).
+- **In-memory cache**: key `{symbol}_{tf}`, TTL = candle duration (M5 300 s … H4 14400 s); a cached frame is evicted and refetched if it fails the freshness check **or the boundary check** (`_has_latest_closed_bar`: the frame must contain the most recent fully closed bar, weekend-exempt — a wall-clock TTL alone would serve mid-candle snapshots at the next boundary and scans would run one bar behind). Fresh fetches enforce the same boundary guarantee: a response missing the just-closed bar (feed latency at the boundary) gets one grace retry after `FRESH_FETCH_RETRY_SECONDS` (3 s), then proceeds with a loud warning. Cache and throttle are lock-protected (one provider is shared across the bot's thread pool).
 - **`_drop_forming_candle`**: Twelve Data includes the current partial candle; the scanner fires seconds after boundary, so the last bar is dropped if its close time is in the future. Without this every indicator would read a seconds-old stub and signals would repaint.
 - **Freshness** (`_is_data_fresh`): last bar older than 3× candle duration → stale (warns; live path suppresses signal if the PRIMARY TF is stale). Weekends exempt (Fri 22:00→Sun 22:00 UTC).
 - **`_validate_and_clean`**: drop all-NaN rows → ffill (limit 3) → drop NaN closes → sort + dedupe (keep last) → drop future timestamps (>now+5 min) → repair OHLC inconsistencies (High=max, Low=min of O/C/existing) → remove >10σ return outliers (spike bar only — the opposite-direction snap-back bar right after it is kept, else every spike punched a two-bar hole) → warn if < ~210 bars.
@@ -270,11 +270,11 @@ Run: `python telegram_bot.py` (credentials resolve env `TELEGRAM_TOKEN`/`TELEGRA
 - **Owner lock**: every command/button is wrapped in `_owner_only` — only `TELEGRAM_CHAT_ID` gets responses; others get a refusal. The bot is the owner's control panel; the product is the channel.
 - **Subscriber channel** (`TELEGRAM_CHANNEL_ID`, `-100…`; `""` disables): auto-scan signal alerts + breach alerts also post there (bot needs channel-admin). Monetization = Telegram Stars subscription invite links; Telegram handles membership, the bot has zero payment code. The channel feed keeps running even when the owner toggles DM auto-scan off (`_feed_active()` = DM chats OR channel).
 - **Auto-scan**: boot-starts if `TELEGRAM_CHAT_ID` set (initial M15 scan populates the level cache first). Candle-boundary loops for M5/M15/H1/H4, firing at boundary + 10 s. Each scan runs all `scan_pairs` concurrently, updates the level cache, then broadcasts.
-- **Alert gating**: quality threshold (`_alert_min_quality`, default "C", menu cycles C→B→A) + fingerprint dedup per (symbol, TF) on (signal, quality, direction, entry bucket) with 1-hour re-arm — a persisting identical signal doesn't re-send every candle, but a setup that moved ≥ half its risk distance (new pullback/swing anchor) re-alerts. The fingerprint is committed only **after** a message reaches ≥1 target (`_commit_alert`), so a failed Telegram send doesn't consume the alert.
+- **Alert gating**: quality threshold (`_alert_min_quality`, default "C", menu cycles C→B→A) + fingerprint dedup per (symbol, TF) on (signal, quality, direction) with 1-hour re-arm — a persisting identical signal doesn't re-send every candle, but a setup whose entry moved ≥ half the last-ALERTED risk distance (new pullback/swing anchor; compared against the stored last-alert entry/risk, a stable reference immune to per-candle ATR drift) re-alerts. The fingerprint is committed only **after** a message reaches ≥1 target (`_commit_alert`), so a failed Telegram send doesn't consume the alert.
 - **WebSocket monitor**: subscribes every `scan_pairs` symbol to TD quotes WS; 10 s heartbeat; exponential reconnect (≤60 s). On tick crossing any cached key level (`swing_high/low_1..3`, `ema_trend`, `bb_upper/lower`, `kc_upper/lower`; cache TTL 1 h) → immediate M15 rescan → "LEVEL BREAK" alert if actionable+changed (same gates as auto-scan), 5-min per-pair cooldown.
 - **Commands**: `/start` (menu), `/scan` (buttons: per pair — M5/M15/H1/H4 + Scan All TFs), `/autoscan` (toggle DM copy), `/status` (loops, WS, live prices, level cache, cooldowns), `/cache` (age per pair/TF), `/clearcache`, `/help`.
 - **Tiingo spread monitor** (optional; on only when a Tiingo key resolves — see Environment variables): a 5-min sampler task fetches top-of-book bid/ask per scanned pair (`tiingo_provider.py`), logs to `logs/spread_log.csv` (skipping the weekend-closed window), and caches the latest quote; alert sends refresh the quote first, and signal messages then show live bid/ask + a warning when spread ≥ 2× the pair's assumed round-trip cost (`WIDE_SPREAD_MULTIPLE`). **Warn-only BY DESIGN** — no historical bid/ask exists, so a suppressing gate could never be validated (live == backtest rule); signals never consume bid/ask. `/status` shows the live spread per pair. **License (free tier is internal-use-only):** raw bid/ask/spread numbers go to owner DMs always; the channel gets them ONLY while config.json `telegram.channel_spread_detail=true` (the channel is internal/testing as of 2026-07-22 — **flip to false before the channel goes public**), otherwise a number-free "spread unusually wide" caution via `format_signal_message(report, spread_detail=False)` — routing in `_spread_detail_for`, test-enforced, committed default false. **Rate limits (free tier 50 req/h, 1000/day, checked 2026-07-22):** sampler 12/h + alert-batch refreshes ≤17/h + breach refreshes ≤12/h = worst case ≈41/h with ONE pair; adding pairs multiplies this — recheck before growing `scan_pairs`.
-- Signal messages include the full indicator breakdown, warnings, trade setup, **trailing-stop management plan** (arm price + trail distance, from Signal/pb params by `strategy`), and a risk disclaimer on every outbound message.
+- Signal messages include the full indicator breakdown, warnings, trade setup, **trailing-stop management plan** (arm price + trail distance, read from the `trail_arm_r`/`trail_dist_r` stamped on the setup by the producing per-TF config), and a risk disclaimer on every outbound message.
 - Blocking scan work runs in `ThreadPoolExecutor(max_workers=4)`; one shared `ForexDataProvider` across all scans (cache reuse).
 - **Logging (hardened 2026-07-22)**: root-logger handlers (`force=True`) → console + `logs/telegram.log` (3-hour rotation, 240 backups = 30 days), full-date timestamps; every module logger (data_provider, tiingo_provider, news_engine, websockets, telegram.ext) lands in the same file. Failure surfaces all covered: candle loops and the spread sampler catch-and-continue per cycle (one bad scan can't kill a loop), breach tasks log exceptions via done-callback, PTB `add_error_handler` registered, and `sys.excepthook`/`threading.excepthook` (installed in `main()`) route uncaught crashes to the log. WS monitor keeps its exponential-backoff reconnect + `logger.exception`.
 
@@ -297,7 +297,7 @@ Replays history bar-by-bar through the **real `SignalAggregator.analyze_from_dat
 - Same-bar SL+TP tie → SL first (conservative). Gap through SL/TP → filled at the bar open.
 - **Trailing exits** (default on, from `signals.trail_arm_r/trail_dist_r`; XAU_PB trades carry per-trade 0.8/0.75 overrides): stop ratchets at bar close only, so a new stop can only fill from the next bar (conservative). Exit reasons: SL / TRAIL / TP1 (static mode) / TIMEOUT / NO_DATA.
 - Session + weekend + weekday filters evaluate the **decision time = bar close** (`ts + primary_duration`) — exactly matching the live bot's wall-clock check.
-- The optional H4-ADX regime filter (`config.backtest.regime_filter`, default off) is applied in the replay by calling the live `_apply_regime_filter_live` method itself under the same guard — live and backtest cannot diverge if it's enabled (2026-07-21 parity fix; the unregistered-pair config fallback also applies the config.json overlay now).
+- The optional H4-ADX regime filter (`config.backtest.regime_filter`, default off) is applied in the replay by calling the live `_apply_regime_filter_live` method itself under the same guard, and the replay loads/slices the regime-filter TF whenever the gate is on (like live `analyze_pair`) — previously M5/M15 primaries silently no-opped the filter because H4 wasn't in their confirm maps. Live and backtest cannot diverge if it's enabled (2026-07-21 parity fix; the unregistered-pair config fallback also applies the config.json overlay now).
 - Confirmation TFs are sliced to bars **fully closed** by decision time (no look-ahead).
 
 **Performance:** `precompute_indicators` builds every indicator series once per TF (CuPy on GPU or NumPy — check the log line `Backend: CuPy GPU` vs `NumPy CPU fallback`; CuPy needs the `nvidia-*-cu12` pip wheels or it silently falls back). `GPUTechnicalEngine` subclasses the live engine and overrides only the `_compute_*` methods with cache lookups — outputs are byte-identical to the live engine (regression-tested). Bar replay: pure per-bar evaluation fans out over forked worker processes (`--workers`, 0 = auto: cores−2 capped 12; <5000 bars forces 1); the entry **cooldown (5 bars) is the only sequential rule**, applied in an ordered pass afterwards — parallel and sequential books are identical. ~300 bars/s/core, ~1700 bars/s at 10 workers.

app_config.py

@@ -65,6 +65,13 @@ def load_dotenv(path: str = ENV_PATH) -> None:
         key, value = key.strip(), value.strip()
         if len(value) >= 2 and value[0] == value[-1] and value[0] in "\"'":
             value = value[1:-1]
+        elif value[:1] not in "\"'":
+            # Unquoted value: whitespace-then-# starts an inline comment
+            # (KEY=abc  # prod). Quote the value to keep a literal " #".
+            for sep in (" #", "\t#"):
+                cut = value.find(sep)
+                if cut != -1:
+                    value = value[:cut].rstrip()
         if key:
             os.environ.setdefault(key, value)
 

backtest/gpu_backtest.py

@@ -851,15 +851,10 @@ class GPUTechnicalEngine(TechnicalEngine):
                 signal = SignalStrength.SELL
                 conf = min(70, 50 + roc_abs * 10)
                 desc = f"ROC {roc_val:.2f}% — Falling momentum"
-        elif roc_prev <= 0 < roc_val:
-            signal = SignalStrength.BUY
-            conf = 65
-            desc = f"ROC crossed above zero (+{roc_val:.2f}%)"
-        elif roc_prev >= 0 > roc_val:
-            signal = SignalStrength.SELL
-            conf = 65
-            desc = f"ROC crossed below zero ({roc_val:.2f}%)"
         else:
+            # Zero-crosses land in the branches above by definition (a
+            # cross up is positive and rising, a cross down negative and
+            # falling) — only flat/fading momentum reaches here.
             signal = SignalStrength.NEUTRAL
             conf = 30
             desc = f"ROC {roc_val:+.2f}% — No clear momentum"
@@ -1690,6 +1685,13 @@ def run_backtest(
 
     # ── Load all required timeframes ──
     timeframes = [bt_cfg.primary_tf] + list(bt_cfg.confirmation_tfs)
+    # Live parity: analyze_pair also fetches the regime-filter TF when the
+    # optional H4-ADX gate is enabled. Without it here, M5/M15 primaries
+    # (whose confirm maps lack H4) would make _apply_regime_filter_live
+    # silently no-op in the replay while live applies the gate.
+    flt = platform_cfg.backtest
+    if flt.regime_filter and flt.regime_filter_timeframe not in timeframes:
+        timeframes.append(flt.regime_filter_timeframe)
     # Pad lookback so earliest TF has enough warmup for EMA200, ADX, etc.
     pad_days = max(60, bt_cfg.warmup_bars // 4)
     fetch_start = bt_cfg.start - timedelta(days=pad_days)
@@ -1879,7 +1881,9 @@ def compute_metrics(trades: List[Trade]) -> Metrics:
     gross_loss = float(-rs[losses_mask].sum()) if losses_mask.any() else 0.0
     pf = gross_profit / gross_loss if gross_loss > 0 else float("inf")
 
-    cum = np.cumsum(rs)
+    # Prepend the flat starting equity so an opening losing streak counts
+    # as drawdown from 0, not from the first trade's already-negative cum.
+    cum = np.concatenate(([0.0], np.cumsum(rs)))
     high_water = np.maximum.accumulate(cum)
     drawdown = high_water - cum
     max_dd = float(drawdown.max()) if len(drawdown) else 0.0

chart_patterns.py

@@ -12,9 +12,10 @@ Chart Patterns*) that are unambiguous from OHLC at bar close:
                                       the interim peak (neckline)
     double_top     bearish reversal   mirrored
     inv_hs         bullish reversal   inverse head & shoulders, break of
-                                      the neckline (max of the two peaks
+                                      the neckline (the LOWER of the two
-                                      flanking the head — conservative
+                                      peaks flanking the head — the earlier
-                                      horizontal line, no sloped fits)
+                                      horizontal trigger, no sloped fits;
+                                      validated as-is, see detector note)
     hs_top         bearish reversal   head & shoulders top, mirrored
     bull_flag      bullish continuation  strong pole, shallow ordered
                                       retrace, close at a new high
@@ -166,8 +167,10 @@ def _detect_hs(o, h, l, c, atr, cfg, bullish: bool) -> Optional[Dict]:
             return None
         if abs(s1 - s2) > cfg.cp_hs_shoulder_tol_atr * atr:
             return None
-        # Neckline: the LOWER of the two interim peaks — the price must
+        # Neckline: the LOWER of the two interim peaks — the EARLIER of
-        # clear both, so this is the conservative horizontal trigger.
+        # the two horizontal triggers (max() would demand clearing both
+        # peaks). The whole validated CP book (M5 BUY-only) ran on this
+        # trigger — do not flip to max() without full re-validation.
         p1 = float(h[i1 : i2 + 1].max())
         p2 = float(h[i2 : i3 + 1].max())
         neckline = min(p1, p2)

data_provider.py

@@ -67,6 +67,9 @@ class ForexDataProvider:
 
     RATE_LIMIT = 55
     RATE_WINDOW = 60.0
+    # Grace wait before the one retry when a fresh response at a candle
+    # boundary is missing the just-closed bar (feed latency).
+    FRESH_FETCH_RETRY_SECONDS = 3.0
 
     def __init__(self, config: PlatformConfig):
         self.config = config
@@ -130,6 +133,24 @@ class ForexDataProvider:
         df = self._drop_forming_candle(df, timeframe)
         df = self._validate_and_clean(df, symbol, timeframe)
 
+        # The closed-bar guarantee must hold for FRESH responses too, not
+        # just cache hits: the scanner fires seconds after the boundary,
+        # and a lagging feed can answer without the just-closed bar — the
+        # scan would silently analyze (and session-check) the previous
+        # bar. Give the feed one short grace retry, then proceed loudly.
+        if (not df.empty and not self._has_latest_closed_bar(df, timeframe)):
+            time.sleep(self.FRESH_FETCH_RETRY_SECONDS)
+            retry = self._fetch_twelve_data(symbol, timeframe, outputsize)
+            retry = self._drop_forming_candle(retry, timeframe)
+            retry = self._validate_and_clean(retry, symbol, timeframe)
+            if retry is not None and not retry.empty:
+                df = retry
+            if not self._has_latest_closed_bar(df, timeframe):
+                logger.warning(
+                    f"{symbol} {timeframe.value}: feed lagging — just-closed bar "
+                    f"still missing after retry; this scan may run one bar behind"
+                )
+
         if not self._is_data_fresh(df, symbol, timeframe):
             logger.warning(
                 f"STALE DATA from API: {symbol} {timeframe.value} — "
@@ -198,6 +219,7 @@ class ForexDataProvider:
 
     def clear_cache(self) -> int:
         """Wipe all cached data. Returns number of entries cleared."""
+        with self._cache_lock:
             count = len(self._cache)
             self._cache.clear()
             self._cache_timestamps.clear()

engine.py

@@ -638,15 +638,10 @@ class TechnicalEngine:
                 signal = SignalStrength.SELL
                 conf = min(70, 50 + roc_abs * 10)
                 desc = f"ROC {roc_val:.2f}% — Falling momentum"
-        elif roc_prev <= 0 < roc_val:
-            signal = SignalStrength.BUY
-            conf = 65
-            desc = f"ROC crossed above zero (+{roc_val:.2f}%)"
-        elif roc_prev >= 0 > roc_val:
-            signal = SignalStrength.SELL
-            conf = 65
-            desc = f"ROC crossed below zero ({roc_val:.2f}%)"
         else:
+            # Zero-crosses land in the branches above by definition (a
+            # cross up is positive and rising, a cross down negative and
+            # falling) — only flat/fading momentum reaches here.
             signal = SignalStrength.NEUTRAL
             conf = 30
             desc = f"ROC {roc_val:+.2f}% — No clear momentum"

main.py

@@ -221,6 +221,9 @@ def report_to_dict(report: SignalReport) -> dict:
             "risk_reward": ts.risk_reward,
             "risk_pips": round(ts.risk_pips, 1),
             "reward_pips": round(ts.reward_pips, 1),
+            "strategy": ts.strategy,
+            "trail_arm_r": ts.trail_arm_r,
+            "trail_dist_r": ts.trail_dist_r,
         }
     if report.primary_analysis:
         d["indicators"] = {
@@ -230,7 +233,7 @@ def report_to_dict(report: SignalReport) -> dict:
                 "description": ind.description,
             }
             for name, ind in report.primary_analysis.indicators.items()
-            if not name.startswith("_")
+            if not name.startswith("_") and ind.weight != 0
         }
         d["key_levels"] = {
             k: round(v, 5) for k, v in report.primary_analysis.key_levels.items()

news_engine.py

@@ -15,6 +15,7 @@ import time
 import xml.etree.ElementTree as ET
 from dataclasses import dataclass, field
 from datetime import datetime, timezone
+from email.utils import parsedate_to_datetime
 from typing import Dict, List, Optional, Tuple
 
 import requests
@@ -281,14 +282,20 @@ class NewsEngine:
         """
         base, quote = self._parse_pair(pair_symbol)
 
-        # Check upcoming high-impact events
+        # Check upcoming high-impact events. Grade from the SOONEST
+        # matching event, not the first listed — an unsorted feed could
+        # otherwise report "caution" for a far event while a nearer one
+        # warrants "danger".
+        soonest = None
         for event in context.events_upcoming_1h:
             if event.is_high_impact and event.currency in (base, quote):
-                mins = event.minutes_until
+                if soonest is None or event.minutes_until < soonest.minutes_until:
+                    soonest = event
+        if soonest is not None:
+            mins = soonest.minutes_until
             if mins <= self.config.pre_event_buffer_minutes:
-                    return "danger", f"{event.currency} {event.event} in {mins:.0f}min"
+                return "danger", f"{soonest.currency} {soonest.event} in {mins:.0f}min"
-                else:
+            return "caution", f"{soonest.currency} {soonest.event} in {mins:.0f}min"
-                    return "caution", f"{event.currency} {event.event} in {mins:.0f}min"
 
         # Check for recent high-impact releases. A surprise print (actual
         # away from forecast) keeps the market unstable longer than an
@@ -650,11 +657,20 @@ class NewsEngine:
         if not date_str:
             return None
 
+        # RFC 822/2822 forms, with or without seconds, numeric offsets AND
+        # named zones (GMT, EST/EDT, PST…). The old strptime "%Z" path
+        # accepted ANY abbreviation but attached no offset, so e.g. an EST
+        # pubDate was silently stamped UTC — a 5-hour error.
+        try:
+            dt = parsedate_to_datetime(date_str.strip())
+            if dt is not None:
+                if dt.tzinfo is None:   # unknown zone name — assume UTC
+                    return dt.replace(tzinfo=timezone.utc)
+                return dt.astimezone(timezone.utc)
+        except (TypeError, ValueError):
+            pass
+
         formats = [
-            "%a, %d %b %Y %H:%M:%S %z",   # RFC 822 with numeric tz offset
-            "%a, %d %b %Y %H:%M:%S %Z",   # RFC 822 with tz abbreviation (GMT/UTC)
-            "%a, %d %b %Y %H:%M %Z",      # RFC 822 without seconds (MyFXBook)
-            "%a, %d %b %Y %H:%M %z",      # RFC 822 without seconds, numeric offset
             "%Y-%m-%dT%H:%M:%S%z",         # ISO 8601 with offset
             "%Y-%m-%d %H:%M:%S",           # ISO 8601 naive (assume UTC)
         ]

signal_aggregator.py

@@ -50,6 +50,12 @@ class TradeSetup:
     pip_value: float
     # Entry archetype ("GENERIC", "XAU_PB") — selects trail params.
     strategy: str = "GENERIC"
+    # Trailing exit params stamped from the PRODUCING config, so per-TF
+    # factory overrides (e.g. M15 arm 0.70, M5 dist 0.85) reach every
+    # consumer. None = legacy/unstamped setup — formatters fall back to
+    # platform defaults; 0.0 means trailing is genuinely disabled.
+    trail_arm_r: Optional[float] = None
+    trail_dist_r: Optional[float] = None
 
 
 @dataclass
@@ -228,7 +234,12 @@ class SignalAggregator:
                     report.action_recommendation = "NO TRADE — Outside tradeable session"
                     report.trade_setup = None
             except Exception:
-                pass
+                # Fail-open by design (an unparseable timestamp shouldn't
+                # kill the scan) but never silently: this filter guards
+                # real money hours.
+                logger.exception(
+                    f"{symbol}: live session filter errored — filter SKIPPED this scan"
+                )
 
         return report
 
@@ -538,6 +549,7 @@ class SignalAggregator:
         min_tp = (self.config.signals.round_trip_cost
                   * self.config.signals.min_tp_cost_multiple)
         if 1.5 * sl_dist < min_tp:
+            _reject("spread_gate")
             return
 
         report.trade_setup = TradeSetup(
@@ -553,7 +565,10 @@ class SignalAggregator:
             reward_pips=1.5 * sl_dist / pip_value,
             pip_value=pip_value,
             strategy="XAU_PB",
+            trail_arm_r=self.config.signals.pb_trail_arm_r,
+            trail_dist_r=self.config.signals.pb_trail_dist_r,
         )
+        _reject("accepted")   # funnel tick, mirrors the CP path
         if report.signal_quality not in ("A+", "A", "B"):
             report.signal_quality = "B"
         report.final_signal = (SignalStrength.BUY if is_buy
@@ -629,6 +644,10 @@ class SignalAggregator:
                     ).hour
                 except Exception:
                     close_h = None
+                    logger.warning(
+                        f"{report.symbol}: bar-close hour underivable — "
+                        f"SELL hour gate skipped"
+                    )
                 if close_h in cfg.sell_blocked_hours_utc:
                     _reject("sell_hour")
                     return
@@ -683,6 +702,8 @@ class SignalAggregator:
             reward_pips=height / pip_value,
             pip_value=pip_value,
             strategy="XAU_CP",
+            trail_arm_r=self.config.signals.trail_arm_r,
+            trail_dist_r=self.config.signals.trail_dist_r,
         )
         if report.signal_quality not in ("A+", "A", "B"):
             report.signal_quality = "B"
@@ -814,7 +835,10 @@ class SignalAggregator:
                         f"DAY FILTER: {day_names.get(bar_wd, '?')} not in allowed days"
                     )
             except Exception:
-                pass
+                # Fail-open by design, but never silently (see live wrapper).
+                logger.exception(
+                    f"{symbol}: session/day filter errored — filter SKIPPED"
+                )
 
         # ── 6c. Transitioning regime quality boost ──────────────────
         if report.market_regime == "transitioning" and report.signal_quality in ("B", "C", "D"):
@@ -922,6 +946,10 @@ class SignalAggregator:
                     ).hour
                 except Exception:
                     close_h = None
+                    logger.warning(
+                        f"{report.symbol}: bar-close hour underivable — "
+                        f"SELL hour gate skipped"
+                    )
                 if close_h in self.config.xau_strategy.sell_blocked_hours_utc:
                     report.warnings.append(
                         f"SELL FILTER: {close_h:02}:00 UTC entry blocked for shorts"
@@ -1048,6 +1076,8 @@ class SignalAggregator:
             risk_pips=round(sl_distance / pip_value, 1),
             reward_pips=round(tp_distance / pip_value, 1),
             pip_value=pip_value,
+            trail_arm_r=self.config.signals.trail_arm_r,
+            trail_dist_r=self.config.signals.trail_dist_r,
         )
 
     def _check_mtf_agreement(self, report: SignalReport) -> bool:
@@ -1165,6 +1195,11 @@ class SignalAggregator:
 
             atr = primary.key_levels.get("atr", 0)
             if atr > 0 and score != 0:
+                # Distance is measured from the FARTHEST of the 3 tracked
+                # swings (sorted()[0] = deepest low / highest high), so the
+                # penalty fires earlier than a nearest-swing rule would.
+                # Calibrated as-is in every validated book — changing the
+                # anchor is a signal-logic change requiring re-validation.
                 if score > 0:
                     swing_lows = sorted(
                         v for k, v in primary.key_levels.items() if k.startswith("swing_low_")
@@ -1445,6 +1480,8 @@ class SignalAggregator:
             risk_pips=risk_pips,
             reward_pips=reward_pips,
             pip_value=pip_value,
+            trail_arm_r=self.config.signals.trail_arm_r,
+            trail_dist_r=self.config.signals.trail_dist_r,
         )
 
     def _collect_reasons(self, report: SignalReport):

telegram_bot.py

@@ -398,10 +398,15 @@ def format_signal_message(report: SignalReport, spread_detail: bool = True) -> s
         lines.append(f"⚖️  RR:     1:{ts.risk_reward:.1f}")
 
         # Trailing-stop management plan (matches backtested exit model:
-        # TP1 is advisory, the trail is the exit). Pullback setups run
+        # TP1 is advisory, the trail is the exit). Params are stamped on
-        # tighter stops and arm later.
+        # the setup by the producing config so per-TF factory overrides
-        sig_cfg = PlatformConfig().signals
+        # (M15 arm 0.70, M5 dist 0.85) are what subscribers see; only a
+        # legacy unstamped setup falls back to platform defaults.
         is_pb = getattr(ts, "strategy", "GENERIC").endswith("_PB")
+        arm_r = getattr(ts, "trail_arm_r", None)
+        dist_r = getattr(ts, "trail_dist_r", None)
+        if arm_r is None or dist_r is None:
+            sig_cfg = PlatformConfig().signals
             arm_r = sig_cfg.pb_trail_arm_r if is_pb else sig_cfg.trail_arm_r
             dist_r = sig_cfg.pb_trail_dist_r if is_pb else sig_cfg.trail_dist_r
         if arm_r > 0:
@@ -574,6 +579,7 @@ _alert_min_quality = _APP.alert_min_quality
 # signal doesn't re-send an identical message every candle.
 _last_alert_fp: Dict[str, tuple] = {}
 _last_alert_ts: Dict[str, float] = {}
+_last_alert_setup: Dict[str, tuple] = {}   # key → (entry, risk) of last ALERTED setup
 _ALERT_REARM_SECONDS = 3600   # identical signal re-alerts after 1h
 
 
@@ -582,40 +588,54 @@ def _passes_quality(quality: str) -> bool:
 
 
 def _alert_fingerprint(report: SignalReport) -> tuple:
-    """Alert identity for dedup. Includes a risk-relative entry bucket:
+    """Alert identity for dedup: signal, quality, setup direction.
-    a persisting signal whose entry drifts a few ticks between candles
+
-    keeps the same fingerprint, while a setup that moved by half its risk
+    Entry movement is checked separately in _should_alert against the
-    distance or more (new pullback, new swing anchor) re-alerts instead
+    last ALERTED entry/risk (a stable reference). Deriving a price bucket
-    of being swallowed for the full re-arm hour."""
+    from the CURRENT SL distance here made the fingerprint churn on
+    normal ATR/anchor drift — near-duplicate alerts every candle."""
     setup = report.trade_setup
-    entry_bucket = None
-    if setup is not None:
-        sl_dist = abs(setup.entry_price - setup.stop_loss)
-        entry_bucket = round(setup.entry_price / max(sl_dist * 0.5, 1e-9))
     return (
         report.final_signal.name,
         report.signal_quality,
         setup.direction if setup else None,
-        entry_bucket,
     )
 
 
 def _should_alert(key: str, report: SignalReport) -> bool:
     """True when this signal is new/changed for the key, or re-armed.
 
+    A persisting identical signal dedups for the re-arm hour, but a setup
+    whose entry moved >= half the last-alerted risk distance (new
+    pullback/swing anchor) re-alerts immediately.
+
     Pure check — call _commit_alert only after the message actually went
     out, so a failed Telegram send doesn't consume the alert for an hour.
     """
     fp = _alert_fingerprint(report)
-    return not (
+    if _last_alert_fp.get(key) != fp:
-        _last_alert_fp.get(key) == fp
+        return True
-        and time.time() - _last_alert_ts.get(key, 0) < _ALERT_REARM_SECONDS
+    if time.time() - _last_alert_ts.get(key, 0) >= _ALERT_REARM_SECONDS:
-    )
+        return True
+    setup = report.trade_setup
+    last = _last_alert_setup.get(key)
+    if setup is not None and last is not None:
+        last_entry, last_risk = last
+        if abs(setup.entry_price - last_entry) >= 0.5 * max(last_risk, 1e-9):
+            return True
+    return False
 
 
 def _commit_alert(key: str, report: SignalReport) -> None:
     _last_alert_fp[key] = _alert_fingerprint(report)
     _last_alert_ts[key] = time.time()
+    setup = report.trade_setup
+    if setup is not None:
+        _last_alert_setup[key] = (
+            setup.entry_price, abs(setup.entry_price - setup.stop_loss)
+        )
+    else:
+        _last_alert_setup.pop(key, None)
 
 
 # ── Candle-boundary scheduler ────────────────────────────────────────
@@ -673,7 +693,8 @@ async def _scan_timeframe(
 
     actionable = [
         r for r in reports
-        if _passes_quality(r.signal_quality)
+        if r.final_signal != SignalStrength.NEUTRAL   # same gate as breach path
+        and _passes_quality(r.signal_quality)
         and _should_alert(f"{r.symbol}_{tf.value}", r)
     ]
     if actionable:

test_all.py

@@ -814,6 +814,18 @@ def _():
     assert dt.tzinfo is not None
 
 
+@test("NewsEngine._parse_rss_date named zones carry their real offset")
+def _():
+    # EST is UTC−5: 12:00 EST == 17:00 UTC. The old "%Z" path accepted the
+    # abbreviation but stamped UTC — a silent 5-hour shift.
+    dt = NewsEngine._parse_rss_date("Mon, 20 Jul 2026 12:00:00 EST")
+    assert dt is not None and dt.tzinfo is not None
+    assert dt.astimezone(timezone.utc).hour == 17
+    # GMT stays a no-op; without-seconds (MyFXBook) form still parses.
+    dt2 = NewsEngine._parse_rss_date("Mon, 20 Jul 2026 12:00 GMT")
+    assert dt2 is not None and dt2.astimezone(timezone.utc).hour == 12
+
+
 @test("NewsEngine._parse_rss_date ISO format")
 def _():
     dt = NewsEngine._parse_rss_date("2025-05-12T14:30:00+00:00")
@@ -899,6 +911,27 @@ def _():
     assert risk == "caution"
 
 
+@test("NewsEngine.get_event_risk_level grades from the soonest event")
+def _():
+    # Far event listed FIRST: an unsorted feed must still yield danger
+    # for the nearer event, not caution for the first match.
+    ne = NewsEngine(NewsConfig())
+    context = FundamentalContext()
+    context.events_upcoming_1h = [
+        EconomicEvent(
+            time=datetime.now(timezone.utc) + timedelta(minutes=55),
+            currency="USD", impact="high", event="GDP",
+        ),
+        EconomicEvent(
+            time=datetime.now(timezone.utc) + timedelta(minutes=10),
+            currency="USD", impact="high", event="NFP",
+        ),
+    ]
+    risk, desc = ne.get_event_risk_level("XAU/USD", context)
+    assert risk == "danger"
+    assert "NFP" in desc
+
+
 @test("NewsEngine.get_event_risk_level clear when no events")
 def _():
     ne = NewsEngine(NewsConfig())
@@ -1113,6 +1146,24 @@ def _():
     assert ts.direction == "BUY"
     assert ts.risk_reward == 2.0
     assert ts.entry_price > ts.stop_loss
+    assert ts.trail_arm_r is None and ts.trail_dist_r is None  # unstamped default
+
+
+@test("trade setups carry trail params from the producing config")
+def _():
+    config = PlatformConfig()
+    config.signals.trail_arm_r = 0.70
+    config.signals.trail_dist_r = 0.85
+    agg = SignalAggregator(config)
+    report = SignalReport(symbol="XAU/USD", display_name="XAU/USD", timestamp="now")
+    report.final_signal = SignalStrength.BUY
+    report.final_score = 40.0
+    report.signal_quality = "A+"
+    report.current_price = 3300.0
+    ts = agg._generate_xau_atr_trade_setup(report, 10.0)
+    assert ts is not None, "A+ setup with ATR 10 should generate"
+    assert ts.trail_arm_r == 0.70, "setup must carry the producing config's arm"
+    assert ts.trail_dist_r == 0.85, "setup must carry the producing config's dist"
 
 
 @test("SignalAggregator._check_mtf_agreement with 2 bullish")
@@ -1569,6 +1620,30 @@ def _():
     assert "Indicators" in msg
 
 
+@test("format_signal_message renders stamped per-TF trail plan")
+def _():
+    def _setup(**extra):
+        return TradeSetup(
+            pair="XAU/USD", display_name="XAU/USD", direction="BUY",
+            entry_price=3300.0, stop_loss=3290.0, take_profit_1=3315.0,
+            take_profit_2=3325.0, risk_reward=1.5, risk_pips=100,
+            reward_pips=150, pip_value=0.10, **extra,
+        )
+    report = SignalReport(symbol="XAU/USD", display_name="XAU/USD", timestamp="now")
+    report.final_signal = SignalStrength.BUY
+    report.signal_quality = "A"
+    # Stamped setup (M15-style overrides) must render the stamped values…
+    report.trade_setup = _setup(trail_arm_r=0.70, trail_dist_r=0.85)
+    msg = format_signal_message(report)
+    assert "+0.7R" in msg, "must render the stamped arm, not the 0.6 default"
+    assert "$8.50" in msg, "must render the stamped dist × risk, not the default"
+    # …while a legacy unstamped setup falls back to platform defaults.
+    report.trade_setup = _setup()
+    msg = format_signal_message(report)
+    assert "+0.6R" in msg
+    assert "$7.50" in msg
+
+
 @test("_format_scan_summary produces valid text")
 def _():
     config = PlatformConfig()
@@ -2116,6 +2191,10 @@ def _():
     src = open(_gb.__file__, encoding="utf-8").read()
     assert "_apply_regime_filter_live(report, slice_dfs)" in src
     assert "backtest.regime_filter" in src
+    # …and loads the regime-filter TF like live analyze_pair does, or the
+    # method silently no-ops for M5/M15 primaries (confirm maps lack H4).
+    assert "regime_filter_timeframe not in timeframes" in src, \
+        "replay must fetch/slice the regime-filter TF when the gate is on"
 
 
 @test("disabled_indicators=['macd'] removes the pillar but keeps the histogram")
@@ -2181,6 +2260,40 @@ def _():
         break
 
 
+@test("Fresh fetch missing the just-closed bar retries once, then warns")
+def _():
+    def bars_ending(last):
+        idx = pd.date_range(end=last, periods=300, freq="15min", tz="UTC")
+        v = np.linspace(3000.0, 3100.0, 300)
+        return pd.DataFrame({"Open": v, "High": v + 1, "Low": v - 1,
+                             "Close": v, "Volume": 1000.0}, index=idx)
+
+    for _attempt in range(2):
+        provider = ForexDataProvider(PlatformConfig())
+        started = pd.Timestamp.now(tz="UTC").floor("15min")
+        last_closed = started - pd.Timedelta(minutes=15)
+        lagging = bars_ending(last_closed - pd.Timedelta(minutes=15))
+        complete = bars_ending(last_closed)
+        with patch("data_provider._is_forex_weekend", return_value=False), \
+             patch.object(ForexDataProvider, "FRESH_FETCH_RETRY_SECONDS", 0.0):
+            # (a) lagging first response → one retry returns the closed bar.
+            with patch.object(provider, "_fetch_twelve_data",
+                              side_effect=[lagging.copy(), complete.copy()]) as mf:
+                served = provider.fetch_ohlcv("XAU/USD", Timeframe.M15)
+            # (b) complete first response → no retry at all.
+            provider._cache.clear()
+            provider._cache_timestamps.clear()
+            with patch.object(provider, "_fetch_twelve_data",
+                              return_value=complete.copy()) as mf2:
+                provider.fetch_ohlcv("XAU/USD", Timeframe.M15)
+        if pd.Timestamp.now(tz="UTC").floor("15min") != started:
+            continue  # boundary crossed mid-test — retry once
+        assert mf.call_count == 2, "lagging response must trigger exactly one retry"
+        assert served.index[-1] == last_closed, "retry result must be served"
+        assert mf2.call_count == 1, "boundary-complete response must not retry"
+        break
+
+
 @test("Alert dedup: drift dedups, materially-moved setup re-alerts, no early commit")
 def _():
     import telegram_bot as tb
@@ -2201,17 +2314,23 @@ def _():
     key = "TEST_ALERT_FP"
     tb._last_alert_fp.pop(key, None)
     tb._last_alert_ts.pop(key, None)
-    r1 = rpt(3400.0, 3392.0)          # risk $8 → $4 buckets
+    tb._last_alert_setup.pop(key, None)
+    r1 = rpt(3400.0, 3392.0)          # risk $8 → re-alert threshold $4
     assert tb._should_alert(key, r1)
     assert tb._should_alert(key, r1), "_should_alert must not commit state itself"
     tb._commit_alert(key, r1)
     assert not tb._should_alert(key, r1), "identical signal must dedup"
     assert not tb._should_alert(key, rpt(3400.5, 3392.5)), \
-        "sub-bucket drift must dedup"
+        "sub-threshold drift must dedup"
+    assert not tb._should_alert(key, rpt(3401.0, 3392.6)), \
+        "SL-distance drift (ATR/anchor wobble) must not re-fingerprint"
     assert tb._should_alert(key, rpt(3412.0, 3404.0)), \
         "setup moved 1.5x risk — must re-alert"
+    assert tb._should_alert(key, rpt(3395.5, 3387.5)), \
+        "setup moved half risk AGAINST the entry — must re-alert too"
     tb._last_alert_fp.pop(key, None)
     tb._last_alert_ts.pop(key, None)
+    tb._last_alert_setup.pop(key, None)
 
 
 @test("Backtester --confirm-tfs defaults to None (factory's live map)")
@@ -2250,6 +2369,31 @@ def _():
     assert t2.entry_time == idx[2] and t2.entry_price == 110.0 and t2.sl == 102.0
 
 
+@test("Max drawdown counts the decline from flat starting equity")
+def _():
+    from gpu_backtest import Trade, compute_metrics
+
+    def mk(r):
+        t = Trade(pair="XAU/USD",
+                  entry_time=pd.Timestamp("2026-01-05 10:00", tz="UTC"),
+                  entry_price=3300.0, sl=3292.0, tp1=3312.0, tp2=3320.0,
+                  direction="BUY", pip_value=0.10, risk_pips=80.0,
+                  reward_pips=120.0, rr_planned=1.5, quality="B",
+                  final_score=30.0, final_confidence=60.0,
+                  market_regime="trending_up")
+        t.exit_reason = "SL" if r < 0 else "TP1"
+        t.r_multiple = r
+        return t
+
+    # Opening losing streak: equity 0 → −3R with no prior peak. The old
+    # high-water started at the first trade's cum (−1) and reported 2R.
+    m = compute_metrics([mk(-1.0), mk(-1.0), mk(-1.0)])
+    assert m.max_drawdown_r == 3.0, f"expected 3.0, got {m.max_drawdown_r}"
+    # Unchanged when a peak precedes the streak.
+    m2 = compute_metrics([mk(2.0), mk(-1.0), mk(-1.0), mk(-1.0)])
+    assert m2.max_drawdown_r == 3.0, f"expected 3.0, got {m2.max_drawdown_r}"
+
+
 # ══════════════════════════════════════════════════════════════════════
 #  MODULE 12: Chart-pattern entry (XAU_CP)
 # ══════════════════════════════════════════════════════════════════════
@@ -2997,7 +3141,10 @@ def _():
         "w", suffix=".env", delete=False, encoding="utf-8")
     fh.write("# comment\n\nPLAIN=abc\nexport EXPORTED=def\n"
              'QUOTED="with spaces"\nSINGLE=\'sq\'\nEXISTING=from-dotenv\n'
-             "NOT A KV LINE\n")
+             "NOT A KV LINE\n"
+             "COMMENTED=abc123 # prod key\n"
+             'QUOTED_HASH="keep #this"\n'
+             "HASH_IN_VALUE=abc#123\n")
     fh.close()
     with patch.dict(os.environ, {"EXISTING": "from-shell"}, clear=True):
         load_dotenv(fh.name)
@@ -3007,6 +3154,12 @@ def _():
         assert os.environ["SINGLE"] == "sq"
         assert os.environ["EXISTING"] == "from-shell", \
             "a real shell export must beat .env"
+        assert os.environ["COMMENTED"] == "abc123", \
+            "inline comment must be stripped, not absorbed into the value"
+        assert os.environ["QUOTED_HASH"] == "keep #this", \
+            "quoted values keep their hashes"
+        assert os.environ["HASH_IN_VALUE"] == "abc#123", \
+            "a hash with no preceding whitespace is part of the value"
     load_dotenv("/nonexistent/.env")   # missing file is a silent no-op