Hello / Extending Java code?

Discussions about custom-built testing platforms written in C, C++, or Java.
Post Reply
steven
Contributing Member
Contributing Member
Posts: 9
Joined: Mon Jan 26, 2009 1:00 pm
Location: Dubai, UAE

Hello / Extending Java code?

Post by steven » Mon Jan 26, 2009 11:58 pm

Hi Everyone,

I recently found this forum and have just signed up. I have only been through a few threads but I have already found pertinent information and experience on the subjects I have looked up. So thanks to the contributors for sharing and I hope I can engage some of you going fwd.

I work as a commodity trader and have done so for some years so I have good experience taking risk but none using formal & numerical systems. I have seen them in action through customer flow though. Around 6 months ago, I decide to try to test the trend-following trading approach to see if I might be able to profit from it. As a starting point, I follow Ed Seykota's Trading System project and decide to write my own java framework to implement this. I spend a few hours a week on it and consider this is a good approach for me to grasp the concepts involved in simulation/back-testing.

I am now on the last step of this project and trying to find a good way to proceed from here. The code framework is quite solid but I can see that it needs to be extended further in order to run some realistic back-testing/optimization. Mainly, I need to add capability for multiple underlyings, position addition/reductions and multiple daily signals. I can then test some of the systems I have in mind.

However, I am still unsure on whether to persist with my own framework or whether to switch to a commercial package now that I have an idea of the problem domain involved in running simulations. The end-result I want is a well defined, historically profitable and well behaved trading system. If I pursue the former route, I need to work more on design and implementation of code, lots of testing to iron out the bugs but much less investment in the design of the actual systems to test since I am already familiar with the functionality of the framework. The latter requires I make a decent investment to determine which back-testing packages have the right functionality and then learning their language. I have not really looked at any packages yet. I hope that someone here may have faced the same crossroads or have something useful to contribute. If so, I welcome your input.

Bye for now,

Steve

sluggo
Roundtable Knight
Roundtable Knight
Posts: 2986
Joined: Fri Jun 11, 2004 2:50 pm

Post by sluggo » Tue Jan 27, 2009 9:42 am

Welcome, Steve. Like many others, I followed the same path that you're on. Eventually I decided that it was a good trade, to pay money and buy time. I paid money to a vendor that sold trading system backtesting software, and gained back hundreds of hours of my own time that I would have spent, writing testing programs. Now I could spend those hours creating trading systems and testing them.

Quickly I learned that no backtesting software implements 100.00 percent of the features that I wanted. So it became a choice: live with something Pretty Good, or start over and build something Perfect? I chose the former. Quite a few of my friends chose the latter, spouting off the famous quote The reasonable man adapts himself to the world. The unreasonable man adapts the world to himself. Therefore all progress depends upon the unreasonable man. Most of them eventually got frustrated and quit; sometimes, without ever placing an actual trade.

You will find there are lots of system backtester programs on the market. The latest issue of Stocks and Commodities magazine, for example, presents a certain trading system coded for eleven different backtesters. Over the years I've bought and used a half dozen of them, and I've concluded that the best-of-breed program is Trading Blox. That's what I use now, both for backtesting and for actual trading. Their sales site has got some free videos ("tutorials") that show the program in action, so you can get a pretty good impression of whether it does what you want done.

In some ways, it's "better" to start out by purchasing cheap backtesting software. This will quickly help you discover what you WISH software would do, but bargain programs don't. It'll give you a better appreciation of what top-tier programs do, as a matter of course, without fanfare. You'll realize exactly what your money is buying, when you write the check for a top-tier program.

RedRock
Roundtable Knight
Roundtable Knight
Posts: 941
Joined: Fri Jan 30, 2004 3:54 pm
Location: Chicago

Post by RedRock » Tue Jan 27, 2009 4:51 pm

TBBs very roots extend back to serious professional traders seeking a solution for their own needs. The software, "started out on the correct footing" and has move forward on this track, since.

Roscoe
Roundtable Knight
Roundtable Knight
Posts: 250
Joined: Sat Jan 24, 2004 2:06 am
Location: Houston TX

Post by Roscoe » Wed Jan 28, 2009 4:08 am

Hello Steven and welcome.

I faced the same decision some years back and I elected to spend a few months to research all the available applications. Only one application fully met my extensive list of requirements and so I purchased it (PowerST) and I have been very pleased with the result. I could not have developed anything even a fraction as capable and it has helped me enormously.

RedRock
Roundtable Knight
Roundtable Knight
Posts: 941
Joined: Fri Jan 30, 2004 3:54 pm
Location: Chicago

Post by RedRock » Wed Jan 28, 2009 4:32 am

Roscoe wrote:Hello Steven and welcome.

I faced the same decision some years back and I elected to spend a few months to research all the available applications. Only one application fully met my extensive list of requirements and so I purchased it (PowerST) and I have been very pleased with the result. I could not have developed anything even a fraction as capable and it has helped me enormously.
Cost:

The initial purchase cost for PowerST is $25,000.

The Calculation Engine Source Code is a separate product. Please call for pricing.

steven
Contributing Member
Contributing Member
Posts: 9
Joined: Mon Jan 26, 2009 1:00 pm
Location: Dubai, UAE

Post by steven » Thu Jan 29, 2009 7:41 am

Thank you all for your input.

I think 25k is more than I am willing to invest at this stage of investigation which probably means PowerST is not a viable choice for me.

I have had a quick look through the TradingBlox help file and on the face of it, it seems to do much of what I am after - at least whilst I am still a novice :) . I plan to try to test it out this w/e. I doubt I can get fully into the TradingBlox part of the program during the trial period so I am curious if how other users have found that part of it. For those components that don't come as part of the initial package, does the development facility provide an accessible approach towards coding them? Is there a big investment required to learn how?

Thanks again, Steve

Roscoe
Roundtable Knight
Roundtable Knight
Posts: 250
Joined: Sat Jan 24, 2004 2:06 am
Location: Houston TX

Post by Roscoe » Thu Jan 29, 2009 10:27 pm

Hi RedRock!

$25K is for the full application. The Calculation Engine Source Code option is there to cater for CTA's who need to own the complete source code as well for security/continuity reasons.

I am not kidding when I say that this application is brilliant and well worth the price if you are really serious about testing and trading.

Post Reply