My next motherboard

Quote from linuxtrader:
Its does take more time - but coding and testing typically takes much more time than specifying a design.
More...
Not in general.

At some point you need to decide where your time is best spent: Once you have a couple good system design techniques mastered and you know the details of efficient implementations, it is best to let others do the work and oversee the effort. Then you can concentrate on the trading models and trade management. If you try to do everything you typically get nothing done .....
More...
You can do everything yourself IF you work smart and design a time-efficient testing architecture.

I’m in the same situation as Nitro. I do everything myself thanks to highly automated and efficient testing and trading. In minutes I can write up a script to test 10M unique system combinations, many of them profitable and weakly intercorrelated, and hopefully weakly correlated to my existing systems. Each script will run for day or two per CPU, though it is possible to get valid preliminary results in 30 minutes. Final analysis and real-time deployment takes minutes of my time using interactive GUI tools to explore the results and select systems for real time deployment. While this is all easy to do *now*, it was an overwhelming task to create this testing architecture. I didn’t believe certain aspects were possible until I was halfway through and had discovered some non-obvious tricks to boost efficiency. I’ve been working on it since February '04. I have an older backtesting/real-time architecture from January '03 and other systems and ideas from mid 2000. Fortunately this latest architecture is near perfection in terms of code-correctness, time-efficiency and scalability. No it won't test every trading hypothesis. It does have flaws. However, now I get to spend more time thinking about testable hypotheses and understanding existing results while the computers test and trade on their own with very little intervention or maintanance. My time is now better spent, requring very little coding to advance my systems.

I can't even comprehend how much money it would take to pay someone to do all of this, let alone find someone who could do it properly and profitably, who wasn't already employed by others.
 
Quote from prophet:

.....
You can do everything yourself IF you work smart and design a time-efficient testing architecture.
More...

Well.... if you can do everything yourself then the odds are that the system you are building or have built is of great value to you but of little value to others.........

When we build a system it is capitalized: it had better be of value to another company or organization. I wont fund the building of potential salvage.

Our criteria is that if we spend the time and money to build something it had better be unique enough that it can not be easily replicated in the market and does not already exist in the market: otherwise there is no reason to build it - you buy it, or lease it.

So, I guess we differ on what is an efficient use of time: doing the work of 5 other specialists seems to me to be a collossal waste of talent. If you have a viable venture and the technology to back it you get capital and oversee it using your knowledge. The output can be much more valuable than what a single individual can produce - no matter how talented they think they are ....
 
Quote from linuxtrader:
Well.... if you can do everything yourself then the odds are that the system you are building or have built is of great value to you but of little value to others.........
More...
ROTFL! Where’s your logic on this one? How is automation and rapid prototyping not valuable? You are actually right in a tangential way because my systems aren’t for sale.

When we build a system it is capitalized: it had better be of value to another company or organization. I wont fund the building of potential salvage.
More...
Are we talking about trading systems? If yes, I don’t see why anyone would plan to sell/give their system code to another organization unless they knew the systems were not profitable enough traded privately. Why give up your ideas for someone else to enhance, optimize or hybridize? I say maximize the value for your own company or investors first, unless you are in the business of selling code.

Our criteria is that if we spend the time and money to build something it had better be unique enough that it can not be easily replicated in the market and does not already exist in the market: otherwise there is no reason to build it - you buy it, or lease it.
More...
Agreed.

So, I guess we differ on what is an efficient use of time: doing the work of 5 other specialists seems to me to be a collossal waste of talent. If you have a viable venture and the technology to back it you get capital and oversee it using your knowledge. The output can be much more valuable than what a single individual can produce - no matter how talented they think they are ....
More...
Perhaps. But like I said before, once the testing/trading architecture is built, you thereafter need specialists who understand math, programming and markets well enough to advance the models or systems. Such specialists are hard to find and expensive. Sometimes they take you in all the wrong directions due to lack of experience.

If you mean specialists to help build the architecture, link in datafeeds, write algorithms, etc… then I agree with you. I could always use outside help.
 
Quote from cmaxb:

If you're into C++, you might take a look at Trolltech's toolkit (http://www.trolltech.com). Runs on Linux, Mac OS X, and Windows. Free for personal use.
More...
Thanks alot of this suggestion. Although I may not end up using it, I got the QT3 book compiled and ran many of the example programs and I really like this environment. The key will be performance.

I will be making a decision soon. Right now, I am just trying to get what I have to run under windows, but hopefully will soon be OS independent.

BTW, one thing I don't understand about all these GUI platforms is that they are not thread safe. That mean you have to sort of kludge it when calling the GUI from it's non native thread :mad:

nitro
 
Quote from prophet:

...
If you mean specialists to help build the architecture, link in datafeeds, write algorithms, etc… then I agree with you. I could always use outside help.
More...

Designing the types of systems you mention .. over and over and over again ... is not difficult, not secret, and not worth a premium in todays market. All you need is to get out and see what other firms are doing and have done - and who they employ for the design of their systems and the implementation work.

Good luck with your systems....
 
Quote from nitro:

...
BTW, one thing I don't understand about all these GUI platforms is that they are not thread safe. That mean you have to sort of kludge it when calling the GUI from it's non native thread :mad:

nitro
More...

Have you looked at GTK+ ....
 
Quote from linuxtrader:
Designing the types of systems you mention .. over and over and over again ... is not difficult, not secret, and not worth a premium in todays market.
More...
Here is that same contradiction in your argument. If designing profitable systems is “not difficult, not secret, and not worth a premium” then anyone with a bit of money can hire that cheap talent and get rich right? Are we not talking about profitable, low risk, decent size systems? That is what I’m talking about. It is not easy by anyone’s measure except your own apparently.

Experienced system designers do not design “the type of systems you mention over and over again”. My personal choice is to produce modular, expressive architectures or tools so I can spend 98% of my time testing and advancing the models with very little coding. Or am I misinterpreting what you wrote?

Good luck with your systems....
More...
Same to you.
 
You must log in or register to reply here.
Back
Top