The Risk of Trading: Mastering the Most Important Element in Financial Speculation

OBTAINING A VALID SAMPLE SIZE

A general principle of sample sizing is that the larger the number of records in your database to measure, the greater level of precision you will obtain. This sample does not have to be limited to your trade history logs. A sample of chart patterns, candlestick bars, price reactions to stochastic extremes are all examples of items that can be back-tested in your quest for success. All things considered, access to a large database of transactions in all market conditions does not always provide a realistic or optimal result. For one, reviewing such a large number of records, even with the use of analytics software, can be time consuming. Your time as a trader is precious and can be used performing other tasks related to your business. It can also be costly if you have a coach, auditor, or risk professional perform these for you. Rarely should your entire trade database be used or should you back-test a pattern over several decades. To the contrary, a small sample such as the one in our coin flip test may provide a wide range of results if we only flipped it 10–15 times.

For a hedge fund or proprietary desks, it is suggested that extended stratified sampling methods be considered. The reason is twofold. Since the assets at risk are commonly higher and from multiple investment sources, such entities should have the advanced tools to support the capital at risk. Furthermore, these infrastructures often require more scrutiny and reporting to multiple stakeholder ...