Combo-MF, a simple and robust trading strategy ?!?

[Pascal]

Using the time free because of the current neutral state the 20DMF is in, I did some back testing to find an easy trading strategy.
I have written a paper about it and publish it here.
Because my native language is Dutch there is a possibility that my writing in English is not perfect.
I have only been trading for 1½ year and consider myself still very novice.
There are certainly more brighter and more experienced people in this forum.
The idea behind this paper is to obtain feedback.
The results are quite good to my personal humble opinion.
Maybe too good to be true…
Any feedback is more than welcome

PdP-Brugge

Attachment 13180

This is really great work PDP.
I understand that it took you much time.
I encourage trading this with GDX as you suggested, because it will give you much confidence to trade a system that you have back-tested.

Regarding the "poor" results for IWM in 2011: 15.85%, against SPY: 22.43%.
This is not normal, as both trades are entered as the same dates.
I suspect that there is maybe a calculaton or data error somewhere for that year.
You might check the data trade by trade: you pull IWM/SPY data from the Yahoo historical site and then compare the entry/exit prices of your system for each trade. Since there are only 10 trades per year, this should be very quick.


Pascal

[pdp-brugge]

I did some text correcting and some layout changes and inclose the updated document here

Attachment 13183

[pdp-brugge]

Pascal,

I do not find the error in the quotes that you state.
I suppose that the correlation between SPY - IWM - QQQQ is not always optimal and for the trades that the Combo-MF has generated, the IWM came in with the worst timing

I enclose my excel sheet for these period:

Attachment 13184

[Pascal]

Pascal,

I do not find the error in the quotes that you state.
I suppose that the correlation between SPY - IWM - QQQQ is not always optimal and for the trades that the Combo-MF has generated, the IWM came in with the worst timing

I enclose my excel sheet for these period:

Attachment 13184

Ok. I trust this is correct then.
Another note: I see that you are using the opening price in your calculations.
The 20DMF is an EOD signal. Hence, a short signal on 9/21/2011 for example, is only 20DMF activated at the EOD of 9/21.
You might check that too.


Pascal

[pdp-brugge]

I have checked that too.
The 20DMF was on 9/20/2011 in cash and on 9/21/2011 in short mode.
The Vstpro was on 9/20/2011 in SHORT and stayed so on 9/21/2011.
That is the reason that the Combo-MF went one day before the 20DMF in SHORT mode and hence had a better timing.

PdP

[Pascal]

I have checked that too.
The 20DMF was on 9/20/2011 in cash and on 9/21/2011 in short mode.
The Vstpro was on 9/20/2011 in SHORT and stayed so on 9/21/2011.
That is the reason that the Combo-MF went one day before the 20DMF in SHORT mode and hence had a better timing.

PdP

This is great, because your returns will be even better when you'll use the 20DMF RT signals.


Pascal

[Harry]

PdP,

Thank you for sharing! I hope to finish my own backtesting, though I doubt it'll ever achive returns as great as these.

Harry

[ernsttanaka]

PdP

Impressive work -- Looking at your research I have one remark. It seems you backtested the years 2007 till 2011. This period has been exceptional. Indexes and all stocks had a enormous correlation. Stockpicking skill were not really rewarded in comparison with good directional calls on the indexes.

This will not always be the case. I personally expect (pretty soon) a period were the indexes will have modest change but some individual stock will do extremely well.

Thank you for sharing your work.

Ernst

[pdp-brugge]

Update for March 6, 2012

Status MFs:
The VStpro model is LONG since October 6, 2011
The 20DMF is NEUTRAL since January 4, 2012
Following my judgement the Combo-MF is still LONG
The PM MF has turned SHORT on March 2, 2012

My current trades:
I am personally long TQQQ since January 5 and I am long DUST since March 5

Possible changes:
If the 20DMF is EOD below 0 and the Avg Tev of the four inversed ETFs is higher than 4%, then the 20DMF will go SHORT
The VSTpro will remain LONG as long as the moving average of the S&P500 is not going down for 2 days in a row (this is part of the code behind the VSTpro MF)
So there is a possibility that the Combo-MF might go to CASH at the end of this day

Disclaimer: this information is not an investment advice on what to trade next