Celes Renata
4ffde8cc06
- Database migration 018 with 13 tables for trading engine state - Trading engine service (services/trading/) with 12 pure computation modules: position sizer, stop-loss manager, reserve pool, circuit breaker, risk tier controller, correlation matrix, tax lots, trading window, gradual entry, notifications, micro-trading, backtester - Core TradingEngine with pre-trade evaluation pipeline and integration wiring - FastAPI HTTP service with 14 endpoints (health, config, decisions, metrics, backtest) - Performance tracker with Sharpe ratio, drawdown, profit factor computation - 194 Python tests (165 property-based + 29 integration) - Frontend: 13 TanStack Query hooks, 7 dashboard panels, tabbed Trading Engine page - Helm chart entry, network policy, nginx proxy, ingress for trading-engine - Shared infrastructure: enums, Redis keys, TradingConfig in AppConfig
257 lines
9.4 KiB
Python
257 lines
9.4 KiB
Python
"""Stop-loss and take-profit management for the autonomous trading engine.
|
|
|
|
Computes initial stop/take-profit levels from ATR and risk tier parameters,
|
|
re-evaluates levels when volatility or market conditions change, detects
|
|
price crossings that should trigger exits, and tightens stops under
|
|
high-heat or high-severity-event conditions.
|
|
|
|
All public methods are synchronous (pure computation, no DB access).
|
|
Persistence is handled by the caller (engine.py).
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
from datetime import datetime
|
|
|
|
from services.trading.models import (
|
|
OpenPosition,
|
|
RiskTierConfig,
|
|
StopLevels,
|
|
StopTrigger,
|
|
)
|
|
|
|
|
|
class StopLossManager:
|
|
"""Compute and maintain dynamic stop-loss / take-profit levels."""
|
|
|
|
# ------------------------------------------------------------------
|
|
# Public API
|
|
# ------------------------------------------------------------------
|
|
|
|
def compute_initial_levels(
|
|
self,
|
|
entry_price: float,
|
|
atr: float,
|
|
risk_tier: RiskTierConfig,
|
|
is_micro_trade: bool = False,
|
|
) -> StopLevels:
|
|
"""Compute initial stop-loss and take-profit for a new position.
|
|
|
|
For standard trades the risk tier's ATR multiplier and reward/risk
|
|
ratio are used. Micro-trades use a tighter 1.0x ATR multiplier
|
|
and 1.5x stop distance for the take-profit target.
|
|
"""
|
|
if is_micro_trade:
|
|
atr_multiplier = 1.0
|
|
reward_risk_ratio = 1.5
|
|
else:
|
|
atr_multiplier = risk_tier.stop_loss_atr_multiplier
|
|
reward_risk_ratio = risk_tier.reward_risk_ratio
|
|
|
|
stop_distance = atr * atr_multiplier
|
|
stop_loss_price = entry_price - stop_distance
|
|
|
|
if is_micro_trade:
|
|
take_profit_price = entry_price + (stop_distance * reward_risk_ratio)
|
|
else:
|
|
take_profit_price = entry_price + (stop_distance * reward_risk_ratio)
|
|
|
|
return StopLevels(
|
|
stop_loss_price=stop_loss_price,
|
|
take_profit_price=take_profit_price,
|
|
trailing_stop_active=False,
|
|
atr_value=atr,
|
|
atr_multiplier=atr_multiplier,
|
|
reward_risk_ratio=reward_risk_ratio,
|
|
last_updated=datetime.utcnow(),
|
|
)
|
|
|
|
def re_evaluate_levels(
|
|
self,
|
|
position: OpenPosition,
|
|
current_price: float,
|
|
atr: float,
|
|
risk_tier: RiskTierConfig,
|
|
last_levels: StopLevels,
|
|
high_severity_event: bool = False,
|
|
earnings_within_3_days: bool = False,
|
|
portfolio_heat_pct: float = 0.0,
|
|
max_portfolio_heat: float = 0.20,
|
|
) -> StopLevels | None:
|
|
"""Re-evaluate stop/take-profit levels for an open position.
|
|
|
|
Returns updated ``StopLevels`` when a material change is needed,
|
|
or ``None`` when the existing levels are still appropriate.
|
|
|
|
Material change triggers:
|
|
* ATR shift > 10 %
|
|
* Trailing stop activation (price moved > 50 % of TP distance)
|
|
* Earnings proximity tightening (0.7x ATR multiplier)
|
|
* High-severity macro event tightening (0.5x normal multiplier)
|
|
* Proactive heat tightening (heat > 80 % of max)
|
|
"""
|
|
entry_price = position.entry_price
|
|
|
|
# --- Determine effective ATR multiplier -----------------------
|
|
base_multiplier = risk_tier.stop_loss_atr_multiplier
|
|
|
|
if high_severity_event:
|
|
effective_multiplier = base_multiplier * 0.5
|
|
elif earnings_within_3_days:
|
|
effective_multiplier = base_multiplier * 0.7
|
|
else:
|
|
effective_multiplier = base_multiplier
|
|
|
|
# --- Trailing stop check --------------------------------------
|
|
trailing_stop_active = last_levels.trailing_stop_active
|
|
tp_distance = last_levels.take_profit_price - entry_price
|
|
favorable_move = current_price - entry_price
|
|
|
|
if tp_distance > 0 and favorable_move > 0.5 * tp_distance:
|
|
trailing_stop_active = True
|
|
|
|
# --- Compute candidate stop-loss ------------------------------
|
|
candidate_stop = entry_price - (atr * effective_multiplier)
|
|
|
|
# If trailing stop is active, floor the stop at entry (breakeven)
|
|
if trailing_stop_active:
|
|
candidate_stop = max(candidate_stop, entry_price)
|
|
|
|
# --- Proactive heat tightening --------------------------------
|
|
# When portfolio heat exceeds 80% of max, tighten further
|
|
if max_portfolio_heat > 0 and portfolio_heat_pct > 0.8 * max_portfolio_heat:
|
|
heat_tightening_factor = 0.7
|
|
tightened_stop = entry_price - (
|
|
atr * effective_multiplier * heat_tightening_factor
|
|
)
|
|
if trailing_stop_active:
|
|
tightened_stop = max(tightened_stop, entry_price)
|
|
candidate_stop = max(candidate_stop, tightened_stop)
|
|
|
|
# --- Decide whether the change is material --------------------
|
|
atr_change_pct = (
|
|
abs(atr - last_levels.atr_value) / last_levels.atr_value
|
|
if last_levels.atr_value > 0
|
|
else 0.0
|
|
)
|
|
|
|
trailing_changed = trailing_stop_active != last_levels.trailing_stop_active
|
|
multiplier_changed = effective_multiplier != last_levels.atr_multiplier
|
|
|
|
if not trailing_changed and not multiplier_changed and atr_change_pct < 0.10:
|
|
return None # no material change
|
|
|
|
# --- Compute new take-profit ----------------------------------
|
|
stop_distance = atr * effective_multiplier
|
|
reward_risk_ratio = risk_tier.reward_risk_ratio
|
|
candidate_tp = entry_price + (stop_distance * reward_risk_ratio)
|
|
|
|
return StopLevels(
|
|
stop_loss_price=candidate_stop,
|
|
take_profit_price=candidate_tp,
|
|
trailing_stop_active=trailing_stop_active,
|
|
atr_value=atr,
|
|
atr_multiplier=effective_multiplier,
|
|
reward_risk_ratio=reward_risk_ratio,
|
|
last_updated=datetime.utcnow(),
|
|
)
|
|
|
|
def check_price_crossings(
|
|
self,
|
|
positions: list[OpenPosition],
|
|
prices: dict[str, float],
|
|
stop_levels: dict[str, StopLevels],
|
|
) -> list[StopTrigger]:
|
|
"""Return triggers for positions whose price has crossed a level.
|
|
|
|
A ``StopTrigger`` is emitted when:
|
|
* current price <= stop_loss_price → ``"stop_loss"``
|
|
* current price >= take_profit_price → ``"take_profit"``
|
|
"""
|
|
triggers: list[StopTrigger] = []
|
|
|
|
for pos in positions:
|
|
current_price = prices.get(pos.ticker)
|
|
if current_price is None:
|
|
continue
|
|
|
|
levels = stop_levels.get(pos.ticker)
|
|
if levels is None:
|
|
continue
|
|
|
|
if current_price <= levels.stop_loss_price:
|
|
triggers.append(
|
|
StopTrigger(
|
|
ticker=pos.ticker,
|
|
trigger_type="stop_loss",
|
|
current_price=current_price,
|
|
trigger_price=levels.stop_loss_price,
|
|
)
|
|
)
|
|
elif current_price >= levels.take_profit_price:
|
|
triggers.append(
|
|
StopTrigger(
|
|
ticker=pos.ticker,
|
|
trigger_type="take_profit",
|
|
current_price=current_price,
|
|
trigger_price=levels.take_profit_price,
|
|
)
|
|
)
|
|
|
|
return triggers
|
|
|
|
def tighten_for_heat(
|
|
self,
|
|
positions: list[OpenPosition],
|
|
stop_levels: dict[str, StopLevels],
|
|
portfolio_heat: float,
|
|
max_heat: float,
|
|
active_pool: float,
|
|
) -> dict[str, StopLevels]:
|
|
"""Tighten stops on lowest-confidence positions when heat is high.
|
|
|
|
When ``portfolio_heat > 0.8 * max_heat``, the lowest-confidence
|
|
positions get their stops tightened first (moved closer to current
|
|
price) to reduce overall portfolio heat.
|
|
|
|
Returns a *new* dict containing only the tickers whose levels
|
|
were actually changed.
|
|
"""
|
|
if max_heat <= 0 or portfolio_heat <= 0.8 * max_heat:
|
|
return {}
|
|
|
|
# Sort positions by confidence ascending (lowest first)
|
|
sorted_positions = sorted(positions, key=lambda p: p.signal_confidence)
|
|
|
|
updated: dict[str, StopLevels] = {}
|
|
|
|
for pos in sorted_positions:
|
|
levels = stop_levels.get(pos.ticker)
|
|
if levels is None:
|
|
continue
|
|
|
|
# Tighten by reducing the stop distance by 30 %
|
|
heat_factor = 0.7
|
|
new_stop_distance = levels.atr_value * levels.atr_multiplier * heat_factor
|
|
new_stop = pos.entry_price - new_stop_distance
|
|
|
|
# Never move stop further away from current price
|
|
new_stop = max(new_stop, levels.stop_loss_price)
|
|
|
|
# If trailing stop is active, floor at entry
|
|
if levels.trailing_stop_active:
|
|
new_stop = max(new_stop, pos.entry_price)
|
|
|
|
if new_stop != levels.stop_loss_price:
|
|
updated[pos.ticker] = StopLevels(
|
|
stop_loss_price=new_stop,
|
|
take_profit_price=levels.take_profit_price,
|
|
trailing_stop_active=levels.trailing_stop_active,
|
|
atr_value=levels.atr_value,
|
|
atr_multiplier=levels.atr_multiplier,
|
|
reward_risk_ratio=levels.reward_risk_ratio,
|
|
last_updated=datetime.utcnow(),
|
|
)
|
|
|
|
return updated
|