paulpas Role: Trading System Simulator — implements comprehensive paper trading simulations that replicate live trading conditions including slippage, fees, partial fills, and market impact for accurate performance estimation.
Philosophy: Live Trading Replication — paper trading must simulate real-world execution friction to provide accurate expectations of live trading performance; unrealistic simulations lead to overoptimistic performance projections.
Key Principles
Execution Friction Modeling: Include all real-world costs (commissions, spreads, slippage, fees) that affect live trading returns.
Order Book Dynamics: Simulate partial fills and fill probability based on order book depth and size.
Market Impact: Account for price movement caused by own orders, especially for larger positions.
Fill Latency: Include realistic latency between order placement and fill execution.
Scenario-Based Testing: Test under multiple market conditions (trending, volatile, rangebound).
Implementation Guidelines
Structure
- Core logic:
skills/paper-trading/simulator.py - Order book models:
skills/paper-trading/order_book.py - Tests:
skills/tests/test_paper_trading_simulation.py
Patterns to Follow
- Use stateful simulator classes to track account state
- Implement order book models with realistic liquidity patterns
- Separate order placement from fill execution
- Track latency between order submission and execution
- Use vectorized operations for efficient simulation
Code Examples
Realistic Paper Trading Simulator
from dataclasses import dataclass, field
from typing import List, Dict, Optional, Tuple
from enum import Enum
import numpy as np
import pandas as pd
from datetime import datetime, timedelta
from collections import defaultdict
class OrderType(Enum):
"""Order types."""
MARKET = "market" # Execute immediately at market price
LIMIT = "limit" # Execute at specified price or better
STOP = "stop" # Trigger market order when price hits stop level
STOP_LIMIT = "stop_limit" # Trigger limit order when price hits stop level
class OrderSide(Enum):
"""Order direction."""
BUY = "buy"
SELL = "sell"
class FillStatus(Enum):
"""Order fill status."""
PENDING = "pending"
PARTIAL = "partial"
FILLED = "filled"
CANCELLED = "cancelled"
REJECTED = "rejected"
@dataclass
class Order:
"""Trading order."""
order_id: int
order_type: OrderType
order_side: OrderSide
quantity: float
price: Optional[float] = None
stop_price: Optional[float] = None
submitted_at: Optional[pd.Timestamp] = None
filled_at: Optional[pd.Timestamp] = None
filled_price: Optional[float] = None
filled_quantity: float = 0.0
status: FillStatus = FillStatus.PENDING
commission: float = 0.0
slippage: float = 0.0
@dataclass
class Position:
"""Current position state."""
symbol: str
quantity: float
entry_price: float
entry_timestamp: pd.Timestamp
unrealized_pnl: float = 0.0
realized_pnl: float = 0.0
@dataclass
class Trade:
"""Executed trade."""
trade_id: int
order_id: int
symbol: str
side: OrderSide
quantity: float
price: float
timestamp: pd.Timestamp
commission: float
slippage: float
@dataclass
class AccountState:
"""Account state at a point in time."""
timestamp: pd.Timestamp
cash: float
equity: float
margin_used: float
margin_available: float
positions: Dict[str, Position]
open_orders: List[Order]
trades: List[Trade]
daily_pnl: float
cumulative_pnl: float
max_drawdown: float
class OrderBook:
"""
Simulated order book with realistic liquidity patterns.
"""
def __init__(self,
spread: float = 0.001, # 10 bps default
depth: int = 10,
volatility: float = 0.02,
impact_coefficient: float = 0.01):
self.spread = spread
self.depth = depth
self.volatility = volatility
self.impact_coefficient = impact_coefficient
self.bids: List[Tuple[float, float]] = [] # (price, quantity)
self.asks: List[Tuple[float, float]] = [] # (price, quantity)
def update(self, mid_price: float, timestamp: pd.Timestamp):
"""Update order book with current market conditions."""
# Generate realistic order book
self.bids = []
self.asks = []
for i in range(1, self.depth + 1):
# Bids below mid price
bid_price = mid_price * (1 - self.spread / 2 - i * 0.0005)
bid_quantity = 100 * (1 + 0.1 * (self.depth - i)) * np.random.uniform(0.5, 1.5)
self.bids.append((bid_price, bid_quantity))
for i in range(1, self.depth + 1):
# Asks above mid price
ask_price = mid_price * (1 + self.spread / 2 + i * 0.0005)
ask_quantity = 100 * (1 + 0.1 * (self.depth - i)) * np.random.uniform(0.5, 1.5)
self.asks.append((ask_price, ask_quantity))
def get_spread(self, mid_price: float) -> Tuple[float, float]:
"""Get current bid-ask spread."""
if not self.bids or not self.asks:
return mid_price * (1 - self.spread / 2), mid_price * (1 + self.spread / 2)
return self.bids[0][0], self.asks[0][0]
def get_liquidity(self, side: OrderSide) -> Dict[int, float]:
"""Get cumulative liquidity at each price level."""
levels = {}
if side == OrderSide.BUY:
price, quantity = self.asks[0]
cumulative = 0
for i, (p, q) in enumerate(self.asks):
cumulative += q
levels[i + 1] = cumulative
else:
price, quantity = self.bids[0]
cumulative = 0
for i, (p, q) in enumerate(self.bids):
cumulative += q
levels[i + 1] = cumulative
return levels
def calculate_market_impact(self, order_side: OrderSide, quantity: float,
mid_price: float) -> float:
"""Calculate price impact of executing an order."""
# Square root impact model
total_liquidity = sum(q for _, q in self.asks if order_side == OrderSide.BUY else self.bids)
impact = self.impact_coefficient * (quantity / total_liquidity) ** 0.5
if order_side == OrderSide.BUY:
return mid_price * (1 + impact)
else:
return mid_price * (1 - impact)
class ExecutionModel:
"""
Models order execution with realistic behavior.
"""
def __init__(self,
fill_probability_model: str = "quantity_based",
latency_mean: float = 0.1, # seconds
latency_std: float = 0.05,
commission_rate: float = 0.001, # 10 bps
commission_minimum: float = 1.0,
slippage_model: str = "uniform",
slippage_range: Tuple[float, float] = (-0.001, 0.001)):
self.fill_probability_model = fill_probability_model
self.latency_mean = latency_mean
self.latency_std = latency_std
self.commission_rate = commission_rate
self.commission_minimum = commission_minimum
self.slippage_model = slippage_model
self.slippage_range = slippage_range
def get_fill_probability(self, order: Order, order_book: OrderBook,
mid_price: float, timestamp: pd.Timestamp) -> float:
"""Calculate probability of order being filled."""
if order.order_type == OrderType.MARKET:
# Market orders fill immediately
return 1.0
elif order.order_type == OrderType.LIMIT:
# Limit orders fill if price condition is met
if order.order_side == OrderSide.BUY:
# Fill if bid price >= limit price
best_bid, _ = order_book.get_spread(mid_price)
return 1.0 if best_bid >= order.price else 0.0
else:
# Fill if ask price <= limit price
_, best_ask = order_book.get_spread(mid_price)
return 1.0 if best_ask <= order.price else 0.0
elif order.order_type == OrderType.STOP:
# Stop orders trigger when price hits stop level
if order.order_side == OrderSide.BUY:
return 1.0 if mid_price >= order.stop_price else 0.0
else:
return 1.0 if mid_price <= order.stop_price else 0.0
elif order.order_type == OrderType.STOP_LIMIT:
# Stop-limit orders trigger at stop, then limit
return 0.5 # Partial probability for stop-limit
return 0.0
def calculate_slippage(self, order: Order, order_book: OrderBook,
mid_price: float) -> float:
"""Calculate expected slippage for an order."""
if self.slippage_model == "uniform":
# Uniform slippage within range
return np.random.uniform(
self.slippage_range[0],
self.slippage_range[1]
)
elif self.slippage_model == "quantity_weighted":
# Slippage scales with order size
base_slippage = np.random.uniform(
self.slippage_range[0],
self.slippage_range[1]
)
return base_slippage * (1 + order.quantity / 1000)
elif self.slippage_model == "market_impact":
# Slippage based on market impact
impact = order_book.calculate_market_impact(
order.order_side, order.quantity, mid_price
)
return (impact - mid_price) / mid_price
return 0.0
def calculate_commission(self, quantity: float, price: float) -> float:
"""Calculate commission for a trade."""
commission = quantity * price * self.commission_rate
return max(commission, self.commission_minimum)
def simulate_fill(self, order: Order, order_book: OrderBook,
mid_price: float, timestamp: pd.Timestamp) -> Tuple[float, float]:
"""
Simulate order fill.
Returns (filled_price, filled_quantity).
"""
if order.order_type == OrderType.MARKET:
# Market order fills at best available price + slippage
if order.order_side == OrderSide.BUY:
_, ask_price = order_book.get_spread(mid_price)
filled_price = ask_price * (1 + self.slippage_range[1])
else:
bid_price, _ = order_book.get_spread(mid_price)
filled_price = bid_price * (1 + self.slippage_range[0])
return filled_price, order.quantity
elif order.order_type == OrderType.LIMIT:
# Limit order fills at limit price or better
return order.price, order.quantity
elif order.order_type == OrderType.STOP:
# Stop order triggers market order
if order.order_side == OrderSide.BUY:
_, ask_price = order_book.get_spread(mid_price)
filled_price = ask_price
else:
bid_price, _ = order_book.get_spread(mid_price)
filled_price = bid_price
return filled_price, order.quantity
elif order.order_type == OrderType.STOP_LIMIT:
# Stop-limit fills at limit price if triggered
return order.price, order.quantity
return mid_price, 0.0
class PaperTradingSimulator:
"""
Comprehensive paper trading simulator with realistic execution.
"""
def __init__(self,
initial_cash: float = 100000,
commission_rate: float = 0.001,
commission_minimum: float = 1.0,
slippage_range: Tuple[float, float] = (-0.001, 0.001),
market_impact_coefficient: float = 0.01,
order_latency_mean: float = 0.1):
self.initial_cash = initial_cash
self.commission_rate = commission_rate
self.commission_minimum = commission_minimum
self.slippage_range = slippage_range
self.market_impact_coefficient = market_impact_coefficient
self.account = {
"cash": initial_cash,
"equity": initial_cash,
"positions": {},
"open_orders": [],
"trades": [],
"trade_id": 0,
"order_id": 0,
"equity_curve": [],
"max_drawdown": 0.0,
"cumulative_pnl": 0.0
}
self.order_book = OrderBook(
spread=0.001,
impact_coefficient=market_impact_coefficient
)
self.execution_model = ExecutionModel(
commission_rate=commission_rate,
commission_minimum=commission_minimum,
slippage_range=slippage_range
)
self.current_timestamp: Optional[pd.Timestamp] = None
self.trading_active = True
self.order_latency_mean = order_latency_mean
def submit_order(self, symbol: str, order_type: OrderType,
order_side: OrderSide, quantity: float,
price: Optional[float] = None,
stop_price: Optional[float] = None,
timestamp: Optional[pd.Timestamp] = None) -> Order:
"""Submit a new order."""
if not self.trading_active:
return None
if timestamp is None:
timestamp = pd.Timestamp.now()
self.current_timestamp = timestamp
self.account["order_id"] += 1
order = Order(
order_id=self.account["order_id"],
order_type=order_type,
order_side=order_side,
quantity=quantity,
price=price,
stop_price=stop_price,
submitted_at=timestamp
)
self.account["open_orders"].append(order)
return order
def cancel_order(self, order_id: int) -> bool:
"""Cancel an open order."""
for order in self.account["open_orders"]:
if order.order_id == order_id:
order.status = FillStatus.CANCELLED
return True
return False
def update_market(self, prices: pd.Series, timestamp: pd.Timestamp):
"""Update market state for a new timestamp."""
self.current_timestamp = timestamp
# Update order book
current_price = prices.iloc[-1]
self.order_book.update(current_price, timestamp)
# Process pending orders
self._process_orders(prices, timestamp)
# Update position values
self._update_positions(prices, timestamp)
# Calculate account metrics
self._calculate_account_metrics()
# Store equity curve point
self.account["equity_curve"].append({
"timestamp": timestamp,
"equity": self.account["equity"],
"cash": self.account["cash"],
"positions": sum(
pos.quantity * prices.iloc[-1]
for pos in self.account["positions"].values()
)
})
# Update max drawdown
self._update_max_drawdown()
def _process_orders(self, prices: pd.Series, timestamp: pd.Timestamp):
"""Process pending orders for current market state."""
for order in self.account["open_orders"][:]:
if order.status != FillStatus.PENDING:
continue
# Check fill probability
fill_prob = self.execution_model.get_fill_probability(
order, self.order_book, prices.iloc[-1], timestamp
)
if fill_prob > np.random.random():
# Simulate fill
filled_price, filled_quantity = self.execution_model.simulate_fill(
order, self.order_book, prices.iloc[-1], timestamp
)
# Generate fill latency
latency = np.random.normal(
self.execution_model.latency_mean,
self.execution_model.latency_std
)
latency = max(0, latency)
fill_timestamp = timestamp + timedelta(seconds=latency)
# Execute trade
self._execute_trade(
order, filled_price, filled_quantity, fill_timestamp
)
def _execute_trade(self, order: Order, filled_price: float,
filled_quantity: float, timestamp: pd.Timestamp):
"""Execute a trade and update account state."""
# Update order
order.filled_price = filled_price
order.filled_quantity = filled_quantity
order.filled_at = timestamp
order.status = FillStatus.FILLED
# Remove from open orders
self.account["open_orders"] = [
o for o in self.account["open_orders"]
if o.order_id != order.order_id
]
# Calculate commissions
commission = self.execution_model.calculate_commission(
filled_quantity, filled_price
)
# Calculate slippage
mid_price = self.order_book.bids[0][0] if order.order_side == OrderSide.BUY else self.order_book.asks[0][0]
slippage = self.execution_model.calculate_slippage(
order, self.order_book, mid_price
)
order.commission = commission
order.slippage = slippage
# Update positions
self._update_position(order, filled_price, filled_quantity)
# Update cash
if order.order_side == OrderSide.BUY:
self.account["cash"] -= filled_quantity * filled_price + commission
else:
self.account["cash"] += filled_quantity * filled_price - commission
# Create trade record
self.account["trade_id"] += 1
trade = Trade(
trade_id=self.account["trade_id"],
order_id=order.order_id,
symbol="SYMBOL", # Simplified for single symbol
side=order.order_side,
quantity=filled_quantity,
price=filled_price,
timestamp=timestamp,
commission=commission,
slippage=slippage
)
self.account["trades"].append(trade)
def _update_position(self, order: Order, filled_price: float,
filled_quantity: float):
"""Update position for executed order."""
symbol = "SYMBOL" # Simplified
if symbol not in self.account["positions"]:
self.account["positions"][symbol] = Position(
symbol=symbol,
quantity=0,
entry_price=0,
entry_timestamp=self.current_timestamp
)
position = self.account["positions"][symbol]
if order.order_side == OrderSide.BUY:
# Buying increases position
total_cost = position.quantity * position.entry_price + \
filled_quantity * filled_price
position.quantity += filled_quantity
if position.quantity > 0:
position.entry_price = total_cost / position.quantity
else:
# Selling decreases position
# Simplified FIFO for this example
cost_of_goods_sold = filled_quantity * position.entry_price
position.quantity -= filled_quantity
if position.quantity <= 0:
position.quantity = 0
position.entry_price = 0
# Update unrealized P&L
current_price = self.order_book.get_spread(
self.account["equity"] / max(1, position.quantity) if position.quantity > 0 else 1
)[0]
position.unrealized_pnl = (current_price - position.entry_price) * position.quantity
def _update_positions(self, prices: pd.Series, timestamp: pd.Timestamp):
"""Update all positions with current market prices."""
current_price = prices.iloc[-1]
for symbol, position in self.account["positions"].items():
position.unrealized_pnl = (current_price - position.entry_price) * position.quantity
def _calculate_account_metrics(self):
"""Calculate account metrics."""
# Calculate positions value
positions_value = sum(
pos.quantity * self.order_book.bids[0][0]
for pos in self.account["positions"].values()
)
# Calculate equity
self.account["equity"] = self.account["cash"] + positions_value
# Calculate margin metrics (simplified)
self.account["margin_used"] = 0.0 # Simplified for cash account
self.account["margin_available"] = self.account["cash"]
# Calculate daily P&L
if len(self.account["equity_curve"]) > 1:
previous_equity = self.account["equity_curve"][-2]["equity"]
self.account["daily_pnl"] = self.account["equity"] - previous_equity
else:
self.account["daily_pnl"] = 0.0
# Calculate cumulative P&L
self.account["cumulative_pnl"] = self.account["equity"] - self.initial_cash
def _update_max_drawdown(self):
"""Update maximum drawdown calculation."""
if len(self.account["equity_curve"]) > 1:
equity_values = [point["equity"] for point in self.account["equity_curve"]]
running_max = np.maximum.accumulate(equity_values)
drawdowns = (running_max - equity_values) / running_max
max_dd = np.max(drawdowns)
self.account["max_drawdown"] = max(self.account["max_drawdown"], max_dd)
def get_account_state(self, timestamp: Optional[pd.Timestamp] = None) -> AccountState:
"""Get current account state."""
if timestamp is None:
timestamp = self.current_timestamp
return AccountState(
timestamp=timestamp or pd.Timestamp.now(),
cash=self.account["cash"],
equity=self.account["equity"],
margin_used=self.account["margin_used"],
margin_available=self.account["margin_available"],
positions=dict(self.account["positions"]),
open_orders=list(self.account["open_orders"]),
trades=list(self.account["trades"]),
daily_pnl=self.account["daily_pnl"],
cumulative_pnl=self.account["cumulative_pnl"],
max_drawdown=self.account["max_drawdown"]
)
def generate_report(self) -> Dict:
"""Generate comprehensive trading report."""
trades = self.account["trades"]
equity_curve = self.account["equity_curve"]
if not equity_curve:
return {"error": "No trading data available"}
# Calculate performance metrics
returns = []
for i in range(1, len(equity_curve)):
if equity_curve[i-1]["equity"] > 0:
returns.append(
equity_curve[i]["equity"] / equity_curve[i-1]["equity"] - 1
)
returns_series = pd.Series(returns)
total_return = self.account["cumulative_pnl"] / self.initial_cash
if len(returns_series) > 1:
daily_std = returns_series.std()
annualized_return = (1 + total_return) ** (252 / len(equity_curve)) - 1
annualized_volatility = daily_std * np.sqrt(252)
sharpe_ratio = (annualized_return - 0.02) / annualized_volatility if annualized_volatility > 0 else 0
sortino_ratio = (annualized_return - 0.02) / (daily_std * np.sqrt(252) * (returns_series < 0).sum() / len(returns_series)) if (returns_series < 0).sum() > 0 else 0
else:
annualized_return = total_return * 252
annualized_volatility = 0
sharpe_ratio = 0
sortino_ratio = 0
# Trade statistics
wins = sum(1 for t in trades if (t.side == OrderSide.BUY and t.price > self.initial_price) or
(t.side == OrderSide.SELL and t.price < self.initial_price))
# Calculate initial price for comparison
initial_price = self.order_book.bids[0][0] if self.order_book.bids else 100
return {
"performance": {
"total_return": total_return,
"annualized_return": annualized_return,
"annualized_volatility": annualized_volatility,
"sharpe_ratio": sharpe_ratio,
"sortino_ratio": sortino_ratio,
"max_drawdown": self.account["max_drawdown"],
"profit_factor": self._calculate_profit_factor()
},
"trades": {
"total_trades": len(trades),
"win_rate": wins / len(trades) if trades else 0,
"total_commissions": sum(t.commission for t in trades),
"total_slippage": sum(t.slippage * t.quantity * t.price for t in trades)
},
"account": {
"initial_cash": self.initial_cash,
"final_equity": self.account["equity"],
"daily_pnl": self.account["daily_pnl"],
"cumulative_pnl": self.account["cumulative_pnl"]
}
}
def _calculate_profit_factor(self) -> float:
"""Calculate profit factor (gross profits / gross losses)."""
if not self.account["trades"]:
return 1.0
gross_profits = sum(
t.quantity * (t.price - self.order_book.get_spread(t.price)[0])
for t in self.account["trades"]
if t.side == OrderSide.BUY and t.price > self.order_book.get_spread(t.price)[1]
)
gross_losses = abs(sum(
t.quantity * (t.price - self.order_book.get_spread(t.price)[0])
for t in self.account["trades"]
if t.side == OrderSide.BUY and t.price < self.order_book.get_spread(t.price)[1]
))
return gross_profits / gross_losses if gross_losses > 0 else float('inf') if gross_profits > 0 else 1.0
# Example usage
if __name__ == "__main__":
# Create synthetic price data
np.random.seed(42)
n_days = 252
prices = pd.Series(
100 * np.cumprod(1 + np.random.normal(0.001, 0.015, n_days)),
index=pd.date_range('2024-01-01', periods=n_days, freq='D')
)
# Initialize simulator
simulator = PaperTradingSimulator(
initial_cash=100000,
commission_rate=0.001,
slippage_range=(-0.001, 0.001)
)
# Store initial price
simulator.initial_price = 100
# Simulate trading
for i in range(len(prices)):
timestamp = prices.index[i]
current_prices = prices.iloc[:i+1]
# Update market
simulator.update_market(current_prices, timestamp)
# Submit some orders (simplified strategy)
if i > 20:
if np.random.random() > 0.95:
# Random trading
if np.random.random() > 0.5:
simulator.submit_order(
"SYMBOL", OrderType.MARKET, OrderSide.BUY, 10
)
else:
simulator.submit_order(
"SYMBOL", OrderType.MARKET, OrderSide.SELL, 10
)
# Generate report
report = simulator.generate_report()
print("Paper Trading Simulation Report:")
print(f" Total Return: {report['performance']['total_return']:.2%}")
print(f" Annualized Return: {report['performance']['annualized_return']:.2%}")
print(f" Sharpe Ratio: {report['performance']['sharpe_ratio']:.3f}")
print(f" Max Drawdown: {report['performance']['max_drawdown']:.2%}")
print(f" Total Trades: {report['trades']['total_trades']}")
print(f" Win Rate: {report['trades']['win_rate']:.2%}")
print(f" Total Commissions: ${report['trades']['total_commissions']:.2f}")
print(f" Final Equity: ${report['account']['final_equity']:.2f}")
Market Impact Simulation
class MarketImpactSimulator:
"""
Simulates market impact on large orders.
Implements square root and linear impact models.
"""
def __init__(self,
volatility: float = 0.20,
half_life: float = 0.1, # hours
impact_coefficient: float = 0.01,
alpha: float = 0.5):
self.volatility = volatility
self.half_life = half_life
self.impact_coefficient = impact_coefficient
self.alpha = alpha
def calculate_square_root_impact(self, volume: float,
market_volume: float,
mid_price: float) -> float:
"""
Calculate impact using square root model.
Impact ∝ √(order_size / market_volume)
"""
if market_volume == 0:
return 0
ratio = volume / market_volume
impact = self.impact_coefficient * np.sqrt(ratio)
return mid_price * impact
def calculate_linear_impact(self, volume: float,
market_volume: float,
mid_price: float) -> float:
"""
Calculate impact using linear model.
Impact ∝ (order_size / market_volume)
"""
if market_volume == 0:
return 0
ratio = volume / market_volume
impact = self.impact_coefficient * ratio
return mid_price * impact
def calculate_liquidation_impact(self, positions: Dict[str, float],
volumes: Dict[str, float],
prices: Dict[str, float]) -> Dict[str, float]:
"""
Calculate impact of liquidating positions.
"""
impacts = {}
for symbol, position in positions.items():
if symbol not in volumes or symbol not in prices:
continue
volume = volumes[symbol]
price = prices[symbol]
# Calculate impact for closing position
if position > 0:
# Closing long position
impact = self.calculate_square_root_impact(
abs(position), volume, price
)
impacts[symbol] = -impact # Negative impact (price moves against us)
else:
# Closing short position
impact = self.calculate_square_root_impact(
abs(position), volume, price
)
impacts[symbol] = impact # Positive impact (price moves against us)
return impacts
# Example usage
if __name__ == "__main__":
# Create market impact simulator
impact_sim = MarketImpactSimulator(
volatility=0.20,
impact_coefficient=0.01,
alpha=0.5
)
# Calculate impact for different order sizes
market_volume = 1000000 # 1M daily volume
mid_price = 100
order_sizes = [1000, 5000, 10000, 50000, 100000]
print("Market Impact Analysis:")
print(f"Market Volume: {market_volume:,}")
print(f"Mid Price: ${mid_price}")
print("\nOrder Size | Square Root | Linear | Percent Impact")
print("-" * 50)
for size in order_sizes:
square_root_impact = impact_sim.calculate_square_root_impact(size, market_volume, mid_price)
linear_impact = impact_sim.calculate_linear_impact(size, market_volume, mid_price)
sq_pct = square_root_impact / mid_price * 100
linear_pct = linear_impact / mid_price * 100
print(f"${size:>8,} | ${square_root_impact:>10.2f} | ${linear_impact:>6.2f} | {sq_pct:>6.2%}")
Adherence Checklist
Before completing your task, verify:
- Execution Friction Modeling: All real-world costs (commissions, spreads, slippage) are included
- Order Book Dynamics: Simulated order book with realistic liquidity patterns
- Market Impact: Order execution accounts for price impact based on order size
- Fill Latency: Realistic latency between order submission and execution
- Scenario-Based Testing: Simulation can test under different market conditions
Common Mistakes to Avoid
- No Slippage: Ignoring slippage leads to overoptimistic performance estimates
- Instant Fills: Assuming orders fill immediately without latency
- Full Fills: Assuming all orders fill completely regardless of market liquidity
- No Commissions: Omitting trading commissions and fees
- Fixed Impact: Using constant market impact instead of variable impact based on order size
- Perfect Execution: Simulating perfect fill at limit price for all orders
- No Partial Fills: Not accounting for partial fills on larger orders
- Unrealistic Liquidity: Assuming infinite market depth for large orders
References
Almgren, R., & Chriss, N. (2000). Optimal Execution of Portfolio Transactions. Journal of Risk, 3(2), 5-39.
Bouchaud, J. P., & Potters, M. (2003). Theory of Financial Risk and Derivative Pricing. Springer.
Gatheral, J. (2010). No-Dynamic-Arbitrage and Market Impact. Quantitative Finance, 10(7), 749-759.
Cont, R., & de Larrard, A. (2013). Price impact of meta-orders in continuous time models. SIAM Journal on Financial Mathematics, 4(1), 1-25.
Moro, E.,zx
- Market Impact: Order execution accounts for price impact based on order size
- Fill Latency: Realistic latency between order submission and execution
- Scenario-Based Testing: Simulation can test under different market conditions
Common Mistakes to Avoid
- No Slippage: Ignoring slippage leads to overoptimistic performance estimates
- Instant Fills: Assuming orders fill immediately without latency
- Full Fills: Assuming all orders fill completely regardless of market liquidity
- No Commissions: Omitting trading commissions and fees
- Fixed Impact: Using constant market impact instead of variable impact based on order size
- Unrealistic Liquidity: Assuming infinite market depth for large orders
- Fill Latency: Realistic latency between order submission and execution
- Scenario-Based Testing: Simulation can test under different market conditions
Common Mistakes to Avoid
- No Slippage: Ignoring slippage leads to overoptimistic performance estimates
- Instant Fills: Assuming orders fill immediately without latency
- Full Fills: Assuming all orders fill completely regardless of market liquidity
- No Commissions: Omitting trading commissions and fees
- Fixed Impact: Using constant market impact instead of variable impact based on order size
- Unrealistic Liquidity: Assuming infinite market depth for large orders
- Fill Latency: Realistic latency between order submission and execution
- Scenario-Based Testing: Simulation can test under different market conditions
Common Mistakes to Avoid
- No Slippage: Ignoring slippage leads to overoptimistic performance estimates
- Instant Fills: Assuming orders fill immediately without latency
- Full Fills: Assuming all orders fill completely regardless of market liquidity
- No Commissions: Omitting trading commissions and fees
- Fixed Impact: Using constant market impact instead of variable impact based on order size
- Unrealistic Liquidity: Assuming infinite market depth for large orders
Constraints
MUST DO
- Implement commission structures that reflect real broker fee schedules including per-share, per-contract, and regulatory fees
- Model market impact for paper orders: simulate price movement caused by your order based on order book depth
- Use the same data feeds and latency characteristics as live trading to ensure paper results are realistic
- Track slippage separately from commissions and calculate both fill-level and execution-level slippage metrics
- Validate paper trading P&L against theoretical expectations at regular intervals to detect simulation bugs
MUST NOT DO
- Do not assume fills at the next available price without modeling order book dynamics or queue position
- Avoid using perfect historical data for paper trading — add realistic noise and latency characteristics
- Never use different data sources between backtesting, paper trading, and live execution without documenting the delta
- Do not ignore dividend adjustments in equity paper trading — missed dividends create false P&L discrepancies
Live References
Authoritative documentation links for this skill's domain. The model follows markdown links at load time to resolve external references and inline content.