Standard Error (SE) Bands Entry Testing

I like reading Dr. Van Tharp’s weekly newsletters. In a previous issue, there was an article that caught my eye.  Ali Moin-Afshari wrote a 4-part article on Band trading which was particularly interesting to me as I’ve done a lot of work with the MACD BB indicator which is basically Bollinger Bands around the distance between two Moving Averages. In Part 3 of this series, Ali discusses the Standard Error (SE) Bands and how they are superior due to being built upon Regression theory in Statistics. He also discusses a possible entry strategy  that I thought I would test and see how well it worked.

Standard Error bands are constructed by first plotting a Regression Line which can act like a more reactive Moving Average. The Standard Error is then calculated, multiplied by the band multiplier and then plotted above and below the Regression Line.

Strategy Setup #1

  • Regression Period = 21
  • SE Multiplier = 2

Entry Conditions: If the current bar closes outside of one of  bands, enter a trade in the direction of the breakout.

Exit Conditions: Standard entry testing exit. Exit after 5, 10, 15 and 20 bars and look only at Win % to see how often the entry is correct.  Ideally, we are looking for a win % greater than 55%. Also observe how many trade opportunities are presented.

Here’s the win % results:

ES (S&P500 E-mini contract)

Daily Bars

Jan 2010 to Sept 24 2013

Both Short and Long Trades

Instrument Num Bars to Exit Total # of Trades Percent Profitable
ES ##-## 5 44 30%
ES ##-## 10 38 34%
ES ##-## 15 32 50%
ES ##-## 20 28 39%

The results were definitely not that great on the ES so I decided to do an optimization on it to see if changing the parameter values helped any. Of course, I found a few better results, but less than 10% of the optimizations were above 55%, so this doesn’t seem to be a good entry technique for the ES. This isn’t too surprising because the ES is more of a mean reversion instrument while this particular entry technique is more of a trend following strategy. However, I also wanted to see if there was a difference between long and short trades. Here’s the results:

Long Only Trades

Instrument Parameters Total # of Trades Percent Profitable
ES ##-## 5 17 47%
ES ##-## 10 17 47%
ES ##-## 15 17 71%
ES ##-## 20 16 56%

Short Only Trades

Instrument Num Bars To Exit Total # of Trades Percent Profitable
ES ##-## 5 32 25%
ES ##-## 10 29 38%
ES ##-## 15 27 33%
ES ##-## 20 22 32%

As you can see, Long trades definitely fared better than Short trades. It’s possible it’s due to the market type being mostly Bullish over 2010 to now. Further study over a period that spams multiple market types would help.

Next I wanted to try it on a basket of currencies as currencies tend to trend more than the ES. Here are the results:

Currencies

Daily Bars

Jan 2010 to Sept 24, 2013

Instrument Num Bars To Exit Total # of Trades Percent Profitable
$AUDUSD 5 44 66%
$AUDUSD 10 42 57%
$AUDUSD 15 35 54%
$AUDUSD 20 29 48%
$EURCHF 5 43 49%
$EURCHF 15 31 42%
$EURCHF 20 26 65%
$EURCHF 10 35 43%
$EURGBP 5 46 43%
$EURGBP 10 38 50%
$EURGBP 15 32 56%
$EURGBP 20 28 64%
$EURJPY 5 44 52%
$EURJPY 10 36 36%
$EURJPY 15 29 59%
$EURJPY 20 28 68%
$EURUSD 5 48 50%
$EURUSD 10 42 55%
$EURUSD 15 35 49%
$EURUSD 20 32 44%
$GBPUSD 5 49 61%
$GBPUSD 10 39 59%
$GBPUSD 15 35 54%
$GBPUSD 20 27 59%
$USDCAD 5 53 51%
$USDCAD 10 44 45%
$USDCAD 15 37 54%
$USDCAD 20 30 57%
$USDCHF 5 54 54%
$USDCHF 10 44 61%
$USDCHF 15 37 51%
$USDCHF 20 32 53%
$USDJPY 5 51 31%
$USDJPY 10 40 43%
$USDJPY 15 37 43%
$USDJPY 20 32 47%

These results definitely look better. Certain currencies show promise; $GBPUSD is almost above 55% on all 4 parameters, $EURGBP improves as the number of bars increases, and $USDCHF also shows promise. Additionally, the Long trades performed slightly better than Short trades, making the edge even greater.

I also ran the strategy on a basket of ETFs and found that about 25% of runs were above a 55% win rate.  The instruments  with the greatest potential were: BSV, ERO, FXE, FXF, PBE, PGJ, SLV, and TIP.

The next entry I wanted to test was the entry strategy discussed in Ali’s article. It’s pretty much the same except we wait for 2 consecutive closes outside the band before we enter.

Strategy Setup #2

  • Regression Period = 21
  • SE Multiplier = 2

Entry Conditions: If we have 2 consecutive bars close outside of one of  bands, enter a trade in the direction of the breakout.

Exit Conditions: Standard entry testing exit. Exit after 5, 10, 15 and 20 bars and look only at Win % to see how often the entry is correct.  Ideally, we are looking for a win % greater than 55%. Also observe how many trade opportunities are presented.

Here’s the win % results:

ES (S&P500 E-mini contract)

Daily Bars

Jan 2010 to Sept 30 2013

Instrument Num Bars to Exit Total # of Trades Percent Profitable Long Trades Short Trades
ES ##-## 5 15 47% 67% 33%
ES ##-## 10 15 53% 50% 56%
ES ##-## 15 13 54% 83% 33%
ES ##-## 20 12 50% 80% 44%

Here we see that the results are better than with just a single close outside the bands. There are fewer trades, as we would expect with tighter entry criteria.

Currencies

Daily Bars

Jan 2010 to Sept 30, 2013

Instrument Num Bars To Exit Total # of Trades Percent Profitable Long Trades Short Trades
$AUDUSD 5 24 50% 50% 50%
$AUDUSD 10 23 52% 50% 53%
$AUDUSD 15 22 50% 63% 47%
$AUDUSD 20 19 42% 43% 47%
$EURCHF 5 18 50% 36% 71%
$EURCHF 10 16 50% 45% 57%
$EURCHF 15 16 63% 45% 86%
$EURCHF 20 15 67% 55% 83%
$EURGBP 5 23 39% 27% 50%
$EURGBP 10 22 50% 50% 50%
$EURGBP 15 20 50% 50% 50%
$EURGBP 20 18 61% 70% 55%
$EURJPY 5 18 44% 50% 38%
$EURJPY 10 17 59% 60% 63%
$EURJPY 15 17 71% 60% 88%
$EURJPY 20 17 71% 60% 88%
$EURUSD 5 24 58% 50% 63%
$EURUSD 10 22 59% 50% 63%
$EURUSD 15 20 55% 63% 50%
$EURUSD 20 19 68% 63% 69%
$GBPUSD 5 17 76% 86% 70%
$GBPUSD 10 16 75% 71% 78%
$GBPUSD 15 14 64% 71% 67%
$GBPUSD 20 14 79% 57% 78%
$USDCAD 5 24 54% 43% 70%
$USDCAD 10 22 45% 50% 44%
$USDCAD 15 18 44% 58% 44%
$USDCAD 20 17 41% 42% 44%
$USDCHF 5 22 50% 36% 75%
$USDCHF 10 22 55% 43% 75%
$USDCHF 15 19 53% 36% 75%
$USDCHF 20 19 58% 43% 75%
$USDJPY 5 29 38% 35% 38%
$USDJPY 10 28 46% 41% 54%
$USDJPY 15 26 35% 25% 58%
$USDJPY 20 21 52% 40% 64%

What’s interesting is that some currencies did better and some did worse when comparing them with the 1 bar close outside the bands, while  others did worse. What is also interesting is that for some of the currencies, short trades worked better than the long trades ($USDCHF, EURCHF). Again that could be due to the market type over the period, but additional study is required to determine if there is a correlation.

I also tested the strategy against the same basket of ETFs. Instruments that did well include ERO, EU, FXB, EWW, GBB, IXJ, KOL, PBE, RSX, TAO, and UDN.

So it appear that for some instruments, this entry strategy certainly would be exploring in detail more. Additional study is needed to determine if there is a correlation between market type and the discrepancy between long and short trades. However, with a good exit strategy, this entry has the potential of offering an edge for your trading.

Advertisements

2 responses to “Standard Error (SE) Bands Entry Testing

  1. I see a lot of interesting articles on your blog. You have to spend a lot of time writing, i know how
    to save you a lot of work, there is a tool that creates unique,
    google friendly articles in couple of minutes, just search in google – laranita’s free
    content source

  2. I read a lot of interesting articles here. Probably you spend a
    lot of time writing, i know how to save you a lot of time,
    there is an online tool that creates unique, google friendly posts in seconds, just type in google – laranitas free content
    source

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s