|
|
import numpy as np |
|
|
import pandas as pd |
|
|
import logging |
|
|
|
|
|
logger = logging.getLogger(__name__) |
|
|
|
|
|
class ExecutionEngine: |
|
|
""" |
|
|
Models realistic slippage based on notional vs. ADV. |
|
|
""" |
|
|
|
|
|
def __init__(self, slippage_coefficient: float): |
|
|
""" |
|
|
:param slippage_coefficient: e.g., 0.0001 (1 bp per 0.1% ADV) |
|
|
""" |
|
|
self.slip_coeff = slippage_coefficient |
|
|
|
|
|
def compute_slippage( |
|
|
self, |
|
|
notional: pd.Series, |
|
|
volume: pd.Series, |
|
|
price: pd.Series |
|
|
) -> pd.Series: |
|
|
""" |
|
|
Computes slippage cost = slip_coeff * (notional / (ADV * price)). |
|
|
:param notional: Series of absolute dollar notional traded. |
|
|
:param volume: Series of share volume (ADV proxy). |
|
|
:param price: Series of price to convert volume to ADV notional. |
|
|
:return: Series of slippage costs (as fraction of capital). |
|
|
""" |
|
|
|
|
|
adv_dollar = volume * price |
|
|
adv_dollar = adv_dollar.replace(0, np.nan).fillna(method="ffill").fillna(1e9) |
|
|
slip = self.slip_coeff * (notional / adv_dollar) |
|
|
return slip |
|
|
|
