Guest Post by DV 34
Here’s a really useful 3 for 1 post by our guest blogger DV34 providing
- A look at the long term chart for the S&P 500
- A really clear explanation of the basic harmonic chart patterns and
- Some high level thoughts on backtesting trading strategies
DV is asking whether readers would be interested in a series of articles covering his approaches to backtesting and his experiences with it. I must say I think this could be really useful resource for traders who want to test their ideas so if you are interested please encourage DV to make the effort by leaving a comment.
As a quick overview I thought I would look at the S&P500,
(dovetailing on Michaels post which was released when I was half way through writing this…!)
Below is a chart of the S&P 500, on a very long term quarterly chart going back to 1996
What I thought was interesting in this was the long term wedge formation and how the market has respected the long term trend lines even during major sell offs during the dot com bust and GFC crash.
The upper trend line is only a slight amount above or basically right at where we are now.
The other interesting point was the cycles of the rallies vs. declines over the last 13 years, where it followed a rough 5 year rise vs. a 2.5 year decline, again, right now we have rallied for approx. 4.5 years from Jan 2009.
On this chart the standard slow stochastic is clearly overbought at a resistance level and the MACD shows declining peaks from each new high suggesting we are in a large range bound market on the very long term cycle, you can see this clearly by simply looking at the 10, 20ema’s that have crossed over themselves numerous times in this range.
On the positive side, central banks continue to flood the market with liquidity (very similar to what Alan Greenspan did in 2001 – following in footsteps!) which has been finding its way into stocks and pushing the market higher without any major reason to slow, so on this chart right now there is no ‘real’ reason to sell just yet without a catalyst.
S&P500 Monthly Chart
On the monthly there is a possibility of an evening star at resistance if this month closes bearish, this is a big if… at the moment the indicators are still converging and agreeing with the uptrend so far…
Given what has happened previously in 2000 and 2007, any initial failure would likely be retested and possibly move above the prior high before any major correction.
If this month does close bearish then it could spark a significant selloff back to the origin of the wedge longer term…
Any initial sell off should see some support around the 20ema/ bottom of the wedge level and possibly bounce back to retest the highs which could offer an even better short trade either below or above the high – but too early to say yet… and could be easily invalidated.
We would really need a fundamental catalyst and something break in the system to see any real follow through on the downside.
Types of Harmonic Patterns
Following on from a question posted on my previous blog by Andy P (thanks Andy!), I thought I would give an overview of harmonic patterns and the different types so it gives some structure around them.
Harmonic patterns are primarily based on swing high/ low analysis to trade corrections, and involve geometric relationships between swings. They are often used to either get into a trend on a correction/ pullback or possibly fade the trend at extremes or even trade within ranges when price action gets choppy.
The great thing I like about the more advanced patterns is that they often break the cardinal trend following rules of HH and HL etc – so trigger in when others are getting out
I created this sheet below today that does not really show the details of each individual pattern, but gives an outline as to what they look like overall and what the key Fibonacci levels are… note the dashed red lines marking the X-A zones.
The ratios between swings vary between patterns so this is more of a high level overview, although if you Google the name you will find plenty of examples showing the correct ratios.
Technically, there is no “Type” of harmonic pattern and I have simply named them Types 1-4 to give some indicative classification system… and make them easier to understand
Type 1: B,C,D all stay within X-A
Type 2: B,C within X-A, D extends past X
Type 3: C extends past A, D within X-A
Type 4: Both C, D extend past X-A
The patterns that are simplest to understand are the classic patterns with simple AB=CD lightning bolt structures, and if you look at my previous blogs I have already shown 6 out of 7 of the classic patterns in real time on various instruments
I only use about half of these patterns personally and know traders that use only one…!
A few things I have observed over time is the importance of the B-C leg, this helps eliminate the patterns/ options quickly.
In the case of the Type 1 vs Type 2 trades, I have used a rule of thumb that if the A-B leg retraces more than 50% of X-A and the C leg is a shallow pullback, the more likely the pattern will extend past X to a Type 2 pattern or possibly further…
For the Type 1, Type 3 patterns the key ratios are 61.8%, 70.7or76.4%(sometimes), 78.6% and 88.6% of X-A. (For Gartleys using 38.2 and 50% is ok in some circumstances)
For Type 2 and Type 4 patterns the key ratios are 113%, 127.2%, 141.4% and 161.8% of X-A
From my experience the challenge has always been around where to place stops especially in Type 2 and 4 patterns without giving away too much profit potential.
So clearly it helps if there is some other structure or S/R level that can allow you to define your risk better – combining other technical’s as well.
Series of Articles on Manual Backtesting
I have recently spent some time back testing some trading ideas I had and considered writing a series of articles about how I do it in the hope that it helps beginners understand some methods and key principals.
I personally believe in the old adage of ‘trust… but verify’, and coming from an engineering type background I can’t help but wanting to dig into details. This is where back testing comes into play
I know there are a myriad of ways that people use for testing from looking back at charts manually with a simple pad and pencil/ excel spreadsheets through to automated EA (expert advisor) testing and optimisation using computer software.
The method I personally use is excel based, so should be relatively easy to follow and anybody with average computer skills should have no problem.
The system I tested was a pattern based discretionary system on a 1hr chart in FX over 12 years of data
The articles will be focussed a lot around Van Tharp’s thinking regarding trading systems which are a must-read – especially his earlier work, and I will also look at the types of testing (forward/ back/ in sample/ out of sample/ monte carlo simulation etc), data quality and why it’s crucial, win%, reward to risk ratios, expectancy, expectunity, position sizing, draw downs, distribution of trades, probability of string of losses and time of day etc…
I will dig a lot more into this, but would like to get some feedback first to see if this may be useful or not from the reader’s perspective – preferably before I start…
Hope this helps and look forward to hearing your feedback