Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

On Data-Driven Drawdown Control with Restart Mechanism in Trading

Chung-Han Hsieh

TL;DR

This paper extends the existing drawdown modulation control policy to include a novel restart mechanism for trading and finds that with the restart mechanism, the policy may achieve a superior trading performance to that without the restart, even with a nonzero transaction costs setting.

Abstract

This paper extends the existing drawdown modulation control policy to include a novel restart mechanism for trading. It is known that the drawdown modulation policy guarantees the maximum percentage drawdown no larger than a prespecified drawdown limit for all time with probability one. However, when the prespecified limit is approaching in practice, such a modulation policy becomes a stop-loss order, which may miss the profitable follow-up opportunities if any. Motivated by this, we add a data-driven restart mechanism into the drawdown modulation trading system to auto-tune the performance. We find that with the restart mechanism, our policy may achieve a superior trading performance to that without the restart, even with a nonzero transaction costs setting. To support our findings, some empirical studies using equity ETF and cryptocurrency with historical price data are provided.

On Data-Driven Drawdown Control with Restart Mechanism in Trading

TL;DR

This paper extends the existing drawdown modulation control policy to include a novel restart mechanism for trading and finds that with the restart mechanism, the policy may achieve a superior trading performance to that without the restart, even with a nonzero transaction costs setting.

Abstract

This paper extends the existing drawdown modulation control policy to include a novel restart mechanism for trading. It is known that the drawdown modulation policy guarantees the maximum percentage drawdown no larger than a prespecified drawdown limit for all time with probability one. However, when the prespecified limit is approaching in practice, such a modulation policy becomes a stop-loss order, which may miss the profitable follow-up opportunities if any. Motivated by this, we add a data-driven restart mechanism into the drawdown modulation trading system to auto-tune the performance. We find that with the restart mechanism, our policy may achieve a superior trading performance to that without the restart, even with a nonzero transaction costs setting. To support our findings, some empirical studies using equity ETF and cryptocurrency with historical price data are provided.
Paper Structure (16 sections, 2 theorems, 25 equations, 6 figures, 2 tables)

This paper contains 16 sections, 2 theorems, 25 equations, 6 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Downside Risks
  3. Control of Drawdown
  4. Contributions of the Paper
  5. Preliminaries
  6. Asset Trading Formulation
  7. Account Value Dynamics
  8. Drawdown Modulation with Restart
  9. Drawdown Modulation Policy
  10. Optimal Drawdown Control Problem
  11. Modulation with Restart
  12. Illustrative Examples
  13. Trading with ETF and Riskless Asset
  14. Trading with Cryptocurrency and Riskless Asset
  15. Transaction Costs Considerations.
  16. ...and 1 more sections

Key Result

Lemma 3.1

Let $d_{\max} \in (0,1)$ be given. An trading policy $u(\cdot)$ guarantees prespecified drawdown limit satisfying $d(k) \leq d_{\max}$ for all $k$ with probability one if and only if for all $k$, the condition is satisfied along all sample paths where

Figures (6)

  • Figure 1: Stock Prices of VT
  • Figure 2: Drawdown Modulation with/without Restart (Green Dots Indicate a Restart)
  • Figure 3: Prices of BTC-USD (In-Sample and Out-of-Sample)
  • Figure 4: Seeking the Optimum: $J(\gamma)$ Versus $\gamma \in \Gamma$
  • Figure 5: Trading Performance of BTC-USD with $d_{\max}:=0.2$ and $\varepsilon:= d_{\max}/10$.
  • ...and 1 more figures

Theorems & Definitions (8)

  • Remark 2.1
  • Definition 2.1: Maximum Percentage Drawdown
  • Remark 2.2
  • Lemma 3.1: Drawdown Modulation
  • Remark 3.1
  • Corollary 3.1: Maximum Drawdown Protection
  • Remark 3.2: Non-Convexity
  • Remark 3.3