Allocation of the robots

[Pascal]

Thank you Dave, your comment is correct.
For the old (backtest) 2007 trades, I will check with Pascal or I'm sure he will jump in the thread.
About the allocations, they will be further detailed in the Algo & Multi pivot forum over time.
If you trade only IWM/RWM, then the optimal proportion of GDX is much higher than if you trade the double or triple-leveraged IWM related ETFs. However, note that the backtest period included the most volatile market in a decade and future results may differ much if we enter a period of stable volatility.
Pascal will post soon an enlighting robot drawdown study for GDX compared to IWM that demonstrates that non-leveraged and double-leveraged IWM trading should be combined with GDX trading for optimal portfolio risk-adjusted performance.

Dave's explanations are correct.
Here is the link to the DD study that I finsihed this morning:

http://www.effectivevolume.eu/conten...n_analysis.pdf


Pascal

[mingpan.lam]

Dave's explanations are correct.
Here is the link to the DD study that I finsihed this morning:

http://www.effectivevolume.eu/conten...n_analysis.pdf


Pascal

Hi Pascal,

Thanks for the reply.
During 2008, both IWM and GDX have max draw down more than 25%. But IWM/GDX model has about 10% draw only, during that time. Why that will happen, I will expect it will be still 25% drawdown, I understand they have low correlation factor, but just based on the diagram, they are both making a 25% drawndown at 10/15/2008 but it is only 10% when they combined.

Also, any suggestion about free dividends adjusted data I can find. I just want to verify the trade myself and if I follow the robot, I can follow it correctly.

Cheers,

Ellis

[Pascal]

Hi Pascal,

Thanks for the reply.
During 2008, both IWM and GDX have max draw down more than 25%. But IWM/GDX model has about 10% draw only, during that time. Why that will happen, I will expect it will be still 25% drawdown, I understand they have low correlation factor, but just based on the diagram, they are both making a 25% drawndown at 10/15/2008 but it is only 10% when they combined.

Also, any suggestion about free dividends adjusted data I can find. I just want to verify the trade myself and if I follow the robot, I can follow it correctly.

Cheers,

Ellis

It is because the diagram is not precise enough. Both IWM/GDX did not have their worst drawdowns at the same dates, but two weeks apart.

On Oct 8, 2008, the GDX DD was -28% and IWM was -3%
On Oct 23, 2008, the GDX DD was 0% and the IWM DD was -29%

During that period, IWM/GDX often had opposite direction signals.

The DD of the portfolio is of course not the average DD of each component.

Yahoo supplies the dividends corrected data.


Pascal

[Pascal]

Exact!

pascal

[mingpan.lam]

Hi Pascal,

Based on this doc, "IWM_GDX_Portfolio_Analysis".
If you are 100% long on GDX and then after the close of the day, the robots give the following signal.

IWM long
GDX strong long

Should we close half of GDX and buy IWM with limit order?
Or, should we simply allocate half/half to GDX/IWM and follow the robots separately?

What is your suggestion?

Cheers,

Ellis

[Pascal]

Hi Pascal,

Based on this doc, "IWM_GDX_Portfolio_Analysis".
If you are 100% long on GDX and then after the close of the day, the robots give the following signal.

IWM long
GDX strong long

Should we close half of GDX and buy IWM with limit order?
Or, should we simply allocate half/half to GDX/IWM and follow the robots separately?

What is your suggestion?

Cheers,

Ellis

In the back-test, long/strong long are treated the same.
If you are 100% invested in one robot and then the other robot issues a signal, the strategy is to sell 50% and move it to the robot with the new signal.


Pascal

[mingpan.lam]

Dave's explanations are correct.
Here is the link to the DD study that I finsihed this morning:

http://www.effectivevolume.eu/conten...n_analysis.pdf


Pascal

Hi Pascal,

Can I assume the following? For example,

Shorted 7/25/2007 8/1/2007 77.42 74.37 3.94% 1.12%

Means, at 7/25/2007, put a short limited order at 77.42 before the market open and it should be filled, but do you have a protective stop loss order along with the first short limited order?
at 8/1/2007, closed at 74.37, it should be closed by a trailing stop, can I assume this stop is available before the market open?

Cheers,

Ellis

[Pascal]

Hi Pascal,

Can I assume the following? For example,

Shorted 7/25/2007 8/1/2007 77.42 74.37 3.94% 1.12%

Means, at 7/25/2007, put a short limited order at 77.42 before the market open and it should be filled, but do you have a protective stop loss order along with the first short limited order?
at 8/1/2007, closed at 74.37, it should be closed by a trailing stop, can I assume this stop is available before the market open?

Cheers,

Ellis

There is no stop loss on a market direction model simulation. Only on the Multi-time pivots.


Pascal

[asomani]

Pascal, in the document you posted today, did you use intraday or end-of-day drawdowns in the analysis?

Thank you.