Return to site

Do You Need To Backtest A Trading Strategy?

Do You Need To Backtest A Trading Strategy?

broken image

Determining Optimal Trading Rules Without Backtesting

Some strategies use econometric arguments to forecast financial variables such as GDP or inflation; other strategies use fundamental and accounting information to price securities; or search for arbitrage-like opportunities in the pricing of derivatives products, etc. 

For instance, suppose that banking corporations tend to sell off the-run bonds two days before U.S. Treasury auctions, in order to reserve balance sheet for the new “paper”. One could monetize on that knowledge by selling off-the-run bonds three days before auctions. But how? Each investment strategy requires an implementation tactic, often referred to as trading rules.

There are dozens of hedge fund styles, each running dozens of unique investment strategies.

While strategies can be very heterogeneous in nature, tactics are relatively homogeneous. Trading rules provide the algorithm that must be followed to enter and exit a position, for example, a position will be entered when the strategy’s signal reaches a certain value.

Moreover, conditions for exiting a position are often defined through thresholds for profit-taking and stop-losses, these entry and exit rules rely on parameters that are usually calibrated via historical simulations. This practice leads to the problem of backtest overfitting. Because these parameters target specific observations in-sample, to the point that the investment strategy becomes so attached to the past that becomes unfit for the future.

An important clarification is an interest in the exit corridor conditions that maximize performance.

In other words, the position already exists, and the question is how to exit it optimally. Thus, the dilemma often faced by execution traders. And it should not become mistaken with the determination of entry and exit thresholds for some underlying instrument.

Calibrating a trading rule using a historical simulation (also called backtest) contributes to backtest overfitting, which in turn leads to underperformance.

In this paper Professors Carr & Lopez de Prado propose a procedure for determining the optimal trading rule (OTR) without running alternative model configurations through a backtest engine.

Professors Carr & Lopez de Prado present empirical evidence of the existence of such optimal solutions for the case of prices following a discrete Ornstein-Uhlenbeck process. And show how they can become computed numerically. Although they do not derive a closed-form solution for the calculation of OTRs, they conjecture its existence on the basis of the empirical evidence presented.

Read The Full Paper: