X hits on this document





3 / 6

Trading as a Business


The system is the combination of the logic and the data. Characteristics that are important include:

  • The issue to trade, including its liquidity.

  • Being sufficiently mechanical that it can be tested and statisti- cally validated.

  • The length of the bars used to generate the trading signals, such as hourly, daily, or weekly.

  • Methods of entering trades, such as at a limit price, at a stop price, or at the market at the next trading opportunity, includ- ing being able to fill actual orders at prices suggested in the simulation.

  • The length of time the typical trade is held.

  • Frequency of trading.

  • The statistical distribution of gains per trade. Analysis of these stats will show the mean, standard deviation, percent- age of trades that are winners, win to loss ratio, and fatness of the tails. Actual trade results (or hypothetical results) are used rather than theoretical distributions.

  • Proper use of leverage. A futures system that always buys one contract, without regard to the size of the account or the trading history, is unleveraged. Similarly, a stock system that always buys $10,000 worth of an issue is unleveraged. Grow- ing an account requires use of leverage and compounding.

iS The SySTem Working or Broken?

hoW do i knoW The SySTem WorkS?

Future profits depend on the system being profitable in the future. The best estimates we can get of future performance are results of trades made using data that was not used during the development of the sys- tem – that is, out-of-sample data. It would be folly to trade a system that is untested; and equally poor judgment to trade a system known to be unprofitable in backtests. Even if the system appeared to be profitable using in-sample backtest results, that is no guarantee that it will be profitable in the future. Use of in-sample results always over estimates future profitability and under estimates future drawdown. It is not uncommon for systems that are profitable using in-sample data to be unprofitable in the future. Use in-sample results to plan your trading business at your financial peril.

Copyright © 2011 by Howard Bandy All rights reserved This document is a chapter of “Modeling Trading System Performance” Published by Blue Owl Press, Inc. www.modelingtradingsystemperformance.com

Document info
Document views6
Page views6
Page last viewedThu Oct 20 21:41:26 UTC 2016