## Simple Trading Strategy Development Data

Aug 19, 2007: 12:05 PM CST

Previously in the post ATR Strategies to Test, I mentioned that you can create a trading system with random entries that produces a greater than 80% win rate. Despite this extraordinarily high winning trade percentage, the system had a zero or negative expectancy â€“ in other words, it was a failing strategy. However, with the addition of a few simple filters, you can create a positive expectancy with seemingly random entries. Read on to find out how.

Reader JKW provided the simple mathematical formula for random entries and stated that there is no way to create a random entry system with a positive expectancy and he is right. The formula he gave is the following:

“Assuming a random walk, the probability of going up by b before it goes down by a is a/(a+b). The expected winnings would be ab/(a+b). The expected losses would also be ab/(a+b). The total expected profit is ab/(a+b)-ab/(a+b)=0, regardless of what the values for a and b are. So whatever numbers are chosen, the expected profit before commissions and slippage comes out to 0.”

I mentioned that the profit and stop targets were determined by a multiple of the Average True Range after a random entry was chosen for trade execution. The high win rate occurs when you play for an extremely small ATR multiple target while entering an extremely high ATR multiple stop.

In other words, a system will return greater than 80% win if you play for .25 or .50 ATR for the price target and 8 ATRs or greater for the stop. I demonstrated previously how, even though the winning % is extremely high, the expectancy is negative or zero because the 20% or less times the wide stop is hit, significant profits are completely erased.

What does simple strategy testing show that can create a profitable system by adding simple filters?

Trend Filter

A Trend Filter can be added to filter out trades that are randomly entered directly against the prevailing trend, if one is present. This method is a bit arbitrary, and depends on the various definition of “Trend”, but mechanical filters can be used such as defining the positioning of key moving averages, percentage of time price is above a key moving average, and number of closes above a key moving average. One could also use a filter for the Average Directional Index, or ADX, such that random entries would be taken if the value is above 30 (or some lower threshold).

Using a trend filter has mixed results in backtesting, and the results often depend on the type of filter chosen and the underlying market conditions throughout the period the test occurred.

Momentum and Volatility Filters

Using the principle “Momentum Precedes Price,” we can create strategies that utilize this principle that capitalizes on the principle that short term increases in momentum (supply/demand imbalance) tend to create situations of short-term price continuation in the direction of the momentum impulse.

One could use a filter similar to those used in volatility breakout trading strategies, including number of new price highs recently made, number of closes above a volatility band such as the Bollinger Bands, or % distance away from a key moving average.

The theory behind this filter is that price with strong momentum will have a greater chance of continuing in that direction in the short term, meaning that a small ATR target will be more likely to be achieved than a large ATR stop.

Again, tests are mixed depending on the type of filter and parameters of the indicators used to define “Momentum”

The Best Method: A Time Stop

The most effective method for shifting a high win % trading system into the positive is the use of an additional stop strategy: The Time Stop.

Simple logic states that price must travel over a larger period of time to move 8 ATRs than 1 ATR (or similar multiple), testing would show that winning trades (Small targets) would be achieved within far fewer bars than larger targets. If the ATRs are based on daily bars, a winning trade would often take less than a week to be achieved, while losing trades would take weeks to perhaps a month to be achieved.

Apply a filter into the random entry strategy that states: If price has not reached the upper smaller target, nor the lower larger stop target by one week (or some number of bars), then exit the trade at the market.

What this would do would eliminate the significant losses that make the system unprofitable.

However, inherently built into this strategy would be the fact that the winning percentage would decrease along with the size of the losing trades and drawdowns. Again, the length of the time stop would determine overall results, but testing shows an increase in the profitability of the system to generate a slightly positive expectancy over time.

I will be posting a summary of these results, as well as a quick discussion of how the above methods can lay a benchmark to be overcome with the development of any system, and ensure that the results of your system are better than that of a random entry system.

### 8 Responses to “Simple Trading Strategy Development Data”

1. reno Says:

Hello,

Your blog are really well organized and full of information. Thank you for sharing… (smile)

My name is Cornel Tanady, and have been doing research on options trading.

Kindly visit my blog at :

Thank you very much.

Sincerely,

3. Corey Rosenbloom Says:

reno,

Thank you for the comment! There’s a lot to say about this and I don’t want to overwhelm readers in a singular post. I’ll be returning to this thread once or twice per week depending on how active the market is and how much I need to write on other relevant issues.

Stay tuned!

4. Mike W. Says:

What software will you be using to back test possible strategies?

5. Corey Rosenbloom Says:

Mike,

I utilize the backtesting and programming features in TradeStation for my results. I try to keep the analysis parameters as simple as possible and I only look at one variable at a time. I’m having to run tests individually over multiple randomly selected stocks and most major futures contracts. I’m finding it’s a bit more difficult than I expected, and I’m learning a lot in the process.

If you have any suggestions, please let me know.

Corey

6. Corey Rosenbloom Says:

Mike,

I utilize the backtesting and programming features in TradeStation for my results. I try to keep the analysis parameters as simple as possible and I only look at one variable at a time. I'm having to run tests individually over multiple randomly selected stocks and most major futures contracts. I'm finding it's a bit more difficult than I expected, and I'm learning a lot in the process.

If you have any suggestions, please let me know.

Corey

7. piyush Says:

hi. regarding trade station, where can i get the same. Does it come for free or how much does it cost. And is it simple to use or again for any such testing knowing how to program is a must.

8. piyush Says:

hi. regarding trade station, where can i get the same. Does it come for free or how much does it cost. And is it simple to use or again for any such testing knowing how to program is a must.