A validated statistical-arbitrage signal — one that has passed DSR, CPCV, and the seven-sins checklist — still needs to be translated into a portfolio and then into orders. This translation introduces three sources of return drag that the backtesting unit treated as a single deduction: transaction costs (spread + impact), position-sizing and portfolio construction choices, and portfolio-level risk events. This topic makes each of those sources concrete.
The flow through this topic follows the live trading pipeline:
- Transaction cost models inform the portfolio construction engine of the cost of any given trade before it is placed.
- Portfolio construction uses the alpha, risk, and cost signals together to select the target weights.
- Execution algorithms (VWAP, TWAP, implementation-shortfall minimisation) deliver those target weights to market at minimal slippage.
- Portfolio risk management — factor hedging, leverage limits, drawdown circuit-breakers, and the 2007 quant quake as the canonical stress scenario — protects the capital base during market dislocations.
By the end of this topic you should be able to:
- decompose a round-trip trading cost into commissions, slippage, and market impact components, and apply the flat / linear / piecewise-linear / quadratic cost models from Narang;
- contrast rule-based portfolio construction (equal-dollar, equal-risk, signal-weighted) with optimiser approaches and articulate when each is preferred;
- describe VWAP and TWAP benchmark schedules, explain why VWAP is problematic as a performance measure, and define implementation shortfall as arrival-price deviation;
- explain factor hedging in a long-short book, specify the four drivers of the 2007 quant quake, and design a drawdown-based leverage circuit-breaker.
In this lesson
- Contents
- The big picture
The full lesson is for members.
Read the full lessonNew to SelfMadeQuant.com? See membership options or browse the Statistical Arbitrage syllabus.