Developing a Trading System Step by Step

[BENG]

Quote from Murray Ruggiero:
One question I have is since this thread is 18 pages I was wondering if I should start a new thread and just copy a few important entries of mine so the thread is not so long to follow for new people or continue on this thread ?

If the moderator can change the subject of this thread to something like Q&A with Murray Ruggiero, then I think it's better to start a new thread. However, I do love the Q&A here, and eventually, the new thread will get lots of Q&A that becomes off topic. Maybe we can use this new thread to address that problem.

Thanks for keeping this thread in a basket of markets approach, I think it's more important than building multiple systems for one market.

 

[BENG]

Quote from Murray Ruggiero:
In 1.3.6 we wrote a walk forward macro, this does walk forward analysis which is what I think you mean by rolling windows. This is included for free in the current version of the product. In 2.0 we are integrating this technology into the product.

This is what I mean...

Assume my strategy has 1000 different ways to enter the market, and each would have it's own stop method based on the entery price.

I run every single simulation to determine which one will give me the best sharpe ratio. Let's say from Jan 02, 95 to Dec 31, 98.

From the results, I take the best 5 (with the highest sharpe ratio in this case) entry method, and run the simulation for Jan 01, 99 to May 31, 99.

Then I run the 1000 simulations again between June 1, 95 to May 31, 99, pick the best 5 entry method, and run simulations for the next 6 months. Then I compute the results for that 6 months again, and so on until 2005.

Can TradersStudio do all 1000 simulations automatically, be able to pick the best cases based on sharpe ratio/number of trades/profit factor, then generate the annual and monthly results in one report?

Now you see what I mean by running simulations can be very complicated. Though you provide the global macro, but the process above can be very difficult to write, and I will need help.

My suggestion would be, get someone to write the templates by request, so people can download it until all possible scenarios have been covered. Or provide an easier way for people to program dynamic simulations like I mentioned. That would be my suggestion to TradersStudio.

Thanks.

 

[Murray Ruggiero]

Good idea, I will talk to Baron tomorrow.

 

[Xephen]

Quote from Murray Ruggiero:

One question I have is since this thread is 18 pages I was wondering if I should start a new thread and just copy a few important entries of mine so the thread is not so long to follow for new people or continue on this thread ?

Murray,

I am indifferent personally. But I think people who are interested in the topic will actually appreciate that there is a lot of material in the thread. If I am interested in a basket approach, then I will be glad to see that there is a lot of discussion on the thread about it, and I will think it is a really valuable resource.

With that said, I realize a lot of this isn't precisely to do with the system development path you are taking.

X

 

[7sport]

Quote from Xephen:

...But I think people who are interested in the topic will actually appreciate that there is a lot of material in the thread. If I am interested in a basket approach, then I will be glad to see that there is a lot of discussion on the thread about it, and I will think it is a really valuable resource.

With that said, I realize a lot of this isn't precisely to do with the system development path you are taking.

X

I couldn't agree more. I've found this thread very informative both on and off topic. Perhaps when you've completed the lesson at hand, it might make sense to copy over some of the key posts into another, more specific thread and adjust the two thread titles accordingly. But to start a new thread now would only result in more of what's happening here.

So again, i think its best to keep this thread intact until the end and then start a new one with the completed lesson where questions will more than likely remain specific to the topic at hand or else can be redirected to this thread.

 

[Murray Ruggiero]

Thanks, for the input.

 

[Murray Ruggiero]

I as said I am going to discuss one more core method and then develop our system. Here it goes. We all know about classic channel breakout. That is the core of the Turtle system.

This system has not done well over the past few years. The Turtle system uses a 20 day and 55 day breakout. The 20 day is the base breakout which gets filtered and then they use the 55 day breakout as the fail safe. The 20 bar length has been used since channel breakout was invented and markets have changed over time. In addition this length does not work well in many markets like the meats and metals.

During the mid 1990's I invented a methodology call Adaptive Channel breakout. This method adapts the length of the breakout based on the current dominant cycle for that market. The theory is that if the market has a 30 day cycle for example it will move 15 days in one direction and 15 days in the other. We can use that concept to say that the market will set a 15 bar high and a 15 bar low. This is the size of the cycle mode channel for a 30 day cycle. If the market sets a 30 day high or 30 day low, we have broken cycle mode and entered trend mode

This method has worked since I published in Futures Magazine in 1996!. We will continue our discussion of this method in our next installment.

 

[swingman]

Quote from Murray Ruggiero:

The Turtle system uses a 20 day and 55 day breakout. The 20 day is the base breakout which gets filtered and then they use the 55 day breakout as the fail safe.

During the mid 1990's I invented a methodology call Adaptive Channel breakout. The theory is that if the market has a 30 day cycle for example it will move 15 days in one direction and 15 days in the other. We can use that concept to say that the market will set a 15 bar high and a 15 bar low. If the market sets a 30 day high or 30 day low, we have broken cycle mode and entered trend mode

So, these are moving averages: short base breakout of 15 and longer fail safe of 30 to confirm a trend, to use your example?

BTW, I infer you think moving averages in general are the most effective T/A tools and tactics out there--better even then ORB and volatility based techniques?

Thanks.

 

[Murray Ruggiero]

This methodology is not moving averages. We are taking classic channel breakout , buy at the highest high of the past N days ,sell at the lowest low of the past N days.

We calculate N=Current dominant cycle.

This is our other core method, so we can compare and contrast during our analysis.

 

[Capablanca]

So a 3 SMA crossover will be compared to an adaptive donchian channel/turtle breakout to determine which should be the prelimenary setup signal. Okay.

Will the pullback requirement and subsequent entry still be kept or has that been dropped already?