NeoTicker Blog

Unlike the previous tutorials that we focus on how to write a trading system, we are going to explore a special kind of trading systems – the ones that utilize randomization.

The Importance Of Studying Random Systems

Trading systems with random entries (or random exits) are very useful tools for the analysis of the underlying instrument that you wanted to trade. These systems are also good benchmarks to compare against when you work on your own trading systems.

Another application of trading systems based on random entries is the study of specially designed exit strategies. The strength and weaknesses of an exit strategy can be exposed easily by combining that with controlled random entries.

Controlled Randomness

When we work with random systems, the best way to do so is using controlled random streams.

A controlled random stream is initialized with a user defined random seed, that controls exactly what will be returned when you request for the next random value. In short, the bottom line is that you can recreate exactly the same sequence of random values every time you initialize the random stream with this same random seed.

In NeoTicker, it has a total of 100 random streams, the reason for that many separate streams is that you can work with multiple indicators and systems at the same time without interferring each other. That is especially important when you are generating research results using Grid Optimizer or other tools.

Our Example – Random Entry System

I have included a very simple random entry system for this article.

1. This system simply buys or sells at market.
2. A random decision is made on every bar.
3. The signal for a new trade happens 20% of the time.
4. Single entry per direction – thus a new signal in the same direction will not result in adding to the open position.
5. Exit on close every day.

Here is a chart of the system,

System Performance

By using the Grid Optimizer running through the first 1000 seeds, one of the seeds produced the performance below.

The equity curve does not lie, the system performed pretty impressively in terms of profit amount. But we know that it does not make any sense, because the entries are all randomly generated. So how can we distinguish a system that is purely random, from a system that actually has some merits?

The Warning Signs

Take a look at the trade summary of the system performance below.

Focus on the following things,

1. The Sharpe Ratio is weak – as a day trading system, the combined sharpe ratio is very close to 1, implying a system that has the strength of simply collecting interest. Remember that trading the index future is a much higher risk trading activity, so we are looking for a better return potential than what this system provides.

2. The Profit Factor combined with the Winning Percent is pointing towards a system that is working on a 50/50 basis. In fact, that is the truth because our system is random.

3. The Average P/L Per Trade is less than the value of 3 ticks of the underlying data. S&P Emini future has a tick size of 0.25, which translate to a value of $12.50 per tick. If the market orders has a slippage of 1 tick per side, then a system must on average getting more than 2 ticks to breakeven.

Complete Indicator

Random Entry System

Summary

Although random entry/exit systems are not useful for real trading, they can help us understand what quality to look for from our own trading systems.

Discuss this article