Sydnee nightly — exit_flow audit — 2026-06-10 P0 findings: 0 P1 findings: 7 Risks: 5 - Area: Exit flow (how and when open positions are closed) - Branch: `dev` (afddfff — peak-monitor log, no strategy code changed overnight) - Files scanned: `bot.py` (~18 497 lines), `docs/strategy_decisions.md`, yesterday's exit-flow audit (`audit_2026-06-09_exit_flow.md`), `git log --oneline -30` - Bugs found (P0 / P1): 0 / 7 (1 new, 6 carried) - Risks noted: 5 (0 new, 5 carried) Cross-check: `git log --oneline -30` shows no strategy or exit-path commits since the 2026-06-09 audit. All 6 carried P1s are confirmed unfixed. 1 new P1 found: TRIM3 has the same `_trimmed_3` / 1-share flag-clearing gap as the already-carried TRIM1 and TRIM2 bugs. --- Full report (dev branch): https://github.com/kanex1/sydnee.signals/blob/dev/docs/audit_2026-06-10_exit_flow.md Reply FROM [email protected] to [email protected] to request fixes, e.g.: "code_task on sydnee-signals-dev: apply fix for the P0 about RVOL threshold in bot.py" Sydnee Agent will propose + you APPROVE (or plain 'approve') + auto-push to dev. --- Full audit below (first 12 KB) --- # Nightly Audit 2026-06-10 — EXIT FLOW ## Summary - Area: Exit flow (how and when open positions are closed) - Branch: `dev` (afddfff — peak-monitor log, no strategy code changed overnight) - Files scanned: `bot.py` (~18 497 lines), `docs/strategy_decisions.md`, yesterday's exit-flow audit (`audit_2026-06-09_exit_flow.md`), `git log --oneline -30` - Bugs found (P0 / P1): 0 / 7 (1 new, 6 carried) - Risks noted: 5 (0 new, 5 carried) Cross-check: `git log --oneline -30` shows no strategy or exit-path commits since the 2026-06-09 audit. All 6 carried P1s are confirmed unfixed. 1 new P1 found: TRIM3 has the same `_trimmed_3` / 1-share flag-clearing gap as the already-carried TRIM1 and TRIM2 bugs. --- ## Findings ### BUG [P1] (new): TRIM3 `_trimmed_3` flag never set for 1-share runners — target fires every bar, trailing stop permanently suppressed **File:** `bot.py:5297–5325` (EXIT 3 TRIM3 block) **Evidence:** ```python # bot.py:5296-5325 — TRIM3 block: if not hit and is_trimmed_2 and not is_trimmed_3 and t.target_price: _tgt_hit = (t.direction == 1 and live_price >= float(t.target_price)) or ... if _tgt_hit: trim = True # ← unconditional trim3_shares = t.shares // 2 # = 0 for 1-share runner if trim3_shares > 0: t._trimmed_3 = True # ← inside guard — never set for 1-share self._update_trade_field(t.trade_id, trimmed_3=True) # never runs new_stop3 = t.entry_price + 0.5 * _daily_atr * t.direction # never runs ... # no else branch ``` A runner that reaches the target price with exactly 1 share remaining (the common case after TRIM1 on a 3-share entry → 2 shares, then TRIM2 → 1 share) produces: - `trim3_shares = 1 // 2 = 0` - `t._trimmed_3` stays `False` → `not is_trimmed_3 = True` → TRIM3 condition fires again every subsequent bar while `live_price >= target_price` - `trim = True` unconditionally each bar - FIX 1 (trailing stop, line 5473) checks `not trim` → blocked every bar - Stop is never advanced to `entry_price + 0.5 × ATR` (that code is inside the guard) **Concrete cascade for a 3-share LONG entry:** 1. Entry 3sh → TRIM1: 1sh child out, 2sh runner, `t._trimmed = True` 2. TRIM2: 1sh child out, 1sh runner, `t._trimmed_2 = True` 3. Target reached → TRIM3: `trim3_shares = 0`, `t._trimmed_3 = False` unchanged 4. Every subsequent bar with `live_price >= target`: `trim = True` → FIX 1 blocked 5. Stop remains at the TRIM2 level (`entry + 0.25×ATR`) — never tightened to `entry + 0.5×ATR` 6. Hold log line (`elif not trim:`) never written — activity log is silent for the runner 7. Loop ends only when the original stale stop fires or a full-close exit (RSI/BXt) triggers **Impact:** All 1-share runners reaching TRIM3 target silently lose their stop-advance and trailing stop. Stop is stuck at the TRIM2 level, leaving ~0.25×ATR of additional buffer on the table. Activity log goes dark for the runner. Frequency: any 3-share or odd-lot entry where both prior trims succeed before target is reached. **Fix:** Add an `else` after `if trim3_shares > 0:` at `bot.py:5308`: ```python else: t._trimmed_3 = True self._update_trade_field(t.trade_id, trimmed_3=True) # 1-share runner: advance stop to entry+0.5×ATR (same as the normal path) new_stop3 = t.entry_price + 0.5 * _daily_atr * t.direction if _daily_atr > 0 else t.stop_price cur_stop3 = float(t.stop_price or 0) if t.direction == 1: new_stop3 = max(new_stop3, cur_stop3) elif cur_stop3 > 0: new_stop3 = min(new_stop3, cur_stop3) self._update_trade_field(t.trade_id, stop_price=new_stop3) self._log_activity(sym, f"TRIM3 1-share runner @${live_price:.2f} — lock stop → ${new_stop3:.2f}", "trade") ``` --- ### BUG [P1] (carried from 2026-06-09): TRIM1 `_rsi_force_trim` and `_trimmed` flags never set/cleared for 1-share parents **File:** `bot.py:5395–5456` (EXIT 4 TRIM1 block) **Evidence:** ```python # bot.py:5397-5412 — inside EXIT 4: if not hit and not is_trimmed and _trim_hit: if True: trim = True trim_shares = t.shares if _flat_full_close else t.shares // 2 # = 0 for 1-share, non-FLAT if trim_shares > 0: t._trimmed = True # ← inside guard — never set for 1-share t._rsi_force_trim = False # ← inside guard — never cleared for 1-share ... # no else branch ``` For a 1-share parent: `trim = True` every bar, neither flag cleared, BE stop (EXIT 5) and trailing stop (FIX 1) permanently blocked. **Fix:** Add `else: t._trimmed = True; t._rsi_force_trim = False; self._update_trade_field(...)` after `if trim_shares > 0:`. --- ### BUG [P1] (carried from 2026-06-02): `profit_per_share` uses stale bar close — `_in_profit` gate misroutes RSI/BXt exits when intrabar price is at a loss **File:** `bot.py:4855` (`profit_per_share` assignment); `bot.py:5122` (`_in_profit` gate) **Evidence:** ```python # bot.py:4854-4855: live_price = float(ss.last_price) if ss.last_price else close_price profit_per_share = (close_price - t.entry_price) * t.direction # ← stale close # bot.py:5122: _in_profit = profit_per_share > 0 ``` When intrabar price drops below entry after the last 5m close, `_in_profit = True` (stale) routes to trim ladder; BE stop then placed above current price → immediate stop-out on the runner. **Fix:** `_in_profit = (live_price - float(t.entry_price)) * t.direction > 0` at line 5122 and 5271. --- ### BUG [P1] (carried from 2026-06-02): `_check_stops` defined but never called — intrabar bar-high/low stop check is dead code **File:** `bot.py:6583` (definition); zero callers in all `.py` files **Evidence:** ```python # bot.py:6583: def _check_stops(self, trade: Trade, bar) -> None: if trade.stop_price: if trade.direction == 1 and bar.low <= trade.stop_price: # bar-low check ... ``` The 5m exit loop only checks `close_price` against `stop_price`. An intrabar wick through the stop but closing above it is never caught. `_check_stops` was intended to fix this but was never wired in. **Fix:** Wire `_check_stops` into `_evaluate_symbol` near the `_stop_hit_pending` block using `signal_df["low"].iloc[-1]` / `signal_df["high"].iloc[-1]` already in scope. --- ### BUG [P1] (carried from 2026-05-26): OBV Rule F permanently disabled after in-loss skip **File:** `bot.py:4936–4943` **Evidence:** ```python # bot.py:4941 — in-loss / non-loss-exit path: t._obv_be_applied = True # ← consumed WITHOUT moving stop ``` When `OBV_BE_STOP=true`, `RSI_TRIM_PROFIT_GUARD=true`, `OBV_LOSS_EXIT=false` and trade is at a loss: skip sets `_obv_be_applied = True` permanently without moving the stop. Rule F never fires again even after trade recovers to profit. **Fix:** Remove `t._obv_be_applied = True` from the in-loss / non-loss-exit branch; keep only in the profit and OBV_LOSS_EXIT paths. --- ### BUG [P1] (carried from 2026-05-26): TRIM2 `_rsi_force_trim2` flag never cleared for 1-share parents **File:** `bot.py:5337–5371` **Evidence:** ```python # bot.py:5341: trim2_shares = t.shares // 2 # = 0 for 1-share if trim2_shares > 0: t._trimmed_2 = True t._rsi_force_trim2 = False # ← inside guard — never clears when shares=1 ... # no else branch ``` 1-share position: `_rsi_force_trim2 = True` and `force_trim2_pending=True` in DB stay set. TRIM2 re-fires every bar, `trim = True`, suppressing hold log and trailing stop. **Fix:** Add `else: t._trimmed_2 = True; t._rsi_force_trim2 = False; self._update_trade_field(...)`. --- ### BUG [P1] (carried from 2026-05-31): `TV_BXT_TRAIL_AGAINST` default mismatch — live `1.0` vs history endpoint `1.5` **File:** `bot.py:3774` (live OBV trail); `bot.py:14352` (history replay / stop-analysis endpoint) **Evidence:** ```python # bot.py:3774 — live _obv_trail_stop_hit: agnst = float(os.environ.get("TV_BXT_TRAIL_AGAINST", "1.0")) # bot.py:14352 — /api/trade_obv_trail dashboard: agnst = float(os.environ.get("TV_BXT_TRAIL_AGAINST", "1.5")) ``` When env var is unset, live trail uses `1.0×ATR`; dashboard stop-analysis shows `1.5×ATR`. Operator sees a 50% wider trail in the dashboard than what actually ran. **Fix:** Change `bot.py:14352` default `"1.5"` → `"1.0"`. --- ### RISK: `_stop_hit_pending` overwrites earlier `hit=True` reason with no `not hit` guard **File:** `bot.py:5109` **Evidence:** ```python if _stop_hit_pending: # ← no `not hit` check hit, reason = True, f"Stop-loss ${t.stop_price:.2f} · ..." # ← overwrites OBV/HOLD_CAP ``` If HOLD_CAP or OBV_LOSS_EXIT already set `hit=True`, and the same bar's close also hits the stop, `_stop_hit_pending=True` overwrites the prior exit reason. No incorrect close; P&L attribution in DB is wrong. **Fix:** `if _stop_hit_pending and not hit:` at line 5109. --- ### RISK (carried): Tick exit path (`_tick_exit_step`) has no TRIM2 check — `WS_TICK_EXITS=true` skips the trim2 level **File:** `bot.py` (`_tick_exit_step` ~line 16750) After tick T1 fires (50% trim), the remaining 50% can be closed by TICK TARGET before bar close, compressing the 4-level trim ladder into 2 levels. Behavioral divergence from `WS_TICK_EXITS=false` runs. `WS_TICK_EXITS` is default off so live impact is low. --- ### RISK (carried): 5m-naked OBV trail stop unavailable first 75 min of pre-market **File:** `bot.py:3769` (`_tv_atr_obv`); `bot.py:4123` (`_eval_bxt_1m_naked`) OBV trail requires `n >= 15` 5m bars. Entries at/after 04:15 ET have no OBV trail until ~05:30 ET. TV-BXt's `TV_BXT_RTH_ONLY` flag mitigates for TV mode; 5m-naked entries remain unguarded. --- ### RISK (carried): FLAT_EXIT parent closed with empty `exit_reason` in DB **File:** `bot.py:1319` (`_close_trade_in_store`) ```python self.db.close_trade( trade_id=trade_id, ..., exit_reason="", # ← always empty ``` `_close_trade_in_store` hard-codes `exit_reason=""`. FLAT_EXIT parent rows are unclassifiable in SQL exit-type queries. Trim child holds the real reason; parent does not. --- ### RISK (carried): `trim3` parent-DB comment stale **File:** `bot.py:8272` ```python elif trade.trade_id.endswith("_trim3"): pass # no DB column for trim3 yet — state tracked in-memory ``` `trimmed_3` column was added in commit `a6be962` (2026-04-15). The `pass` is correct (parent updated at line 5302 when trim3 is set up), but the stale comment risks a developer adding a redundant update to this handler. --- ## OK (checked, working or recently fixed): - **No overnight code commits:** `git log --oneline -30` shows only peak-monitor hourly logs since 2026-06-09 audit. All carry decisions from yesterday hold. ✓ - **TRIM3 monotonicity guard:** `new_stop3 = max(new_stop3, cur_stop3)` for longs prevents loosening. (This runs only when `trim3_shares > 0`; the 1-share gap is the new P1 above.) ✓ - **Scale-in first-fill fix (4517516):** `st["slots_done"] = slot_n` and `st["start_ms"] = now_ms` preserve slot count and pin the 10-min window to first fill. ✓ - **TRIM1 `if True:` wrapper:** cosmetic inert wrapper; trim logic runs unconditionally when the outer `not is_trimmed and _trim_hit` guard passes. ✓ - **BXt reversal `_pending_reversal` drain:** `t._pending_reversal = False` + `self._update_trade_field(pending_reversal=False)` at line 5268 correctly consumes the flag in-memory and DB on confirmation-bar fire. ✓ - **BXt reversal correct trim routing:** `_do_reversal` path routes through `is_trimmed` / `_is_trimmed_2_now` correctly (arming force-trim flags, not direct `hit=True`) for positions with partial trim history. ✓ - **Deferred-close dedup:** `seen_close_ids` set in `_process_deferred_orders` (line 6542) prevents duplicate close orders from stacked tick callbacks. ✓ - **HOLD_CAP `not hit` guard:** `if _hold_cap > 0 and not hit and ...` at line 4970 prevents HOLD_CAP from overwriting a prior `hit=True`. ✓ - **EXIT 5 (BE) `not trim` guard:** `if not hit and not trim and _daily_atr > 0 and not is_trimmed:` at line 5459 — co