but you belief in work separation, don't you? That more hands build bigger and better systems than single hands?
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Custom Trading Platform
- Thread starter naifwonder
- Start date
but you belief in work separation, don't you? That more hands build bigger and better systems than single hands?
sure, for that I outsource coding work to several parties and look to hire a full0-time programmer soon. I concentrate on what I do best. Competitive Advantage.
Quote from ghostrider77:
but you belief in work separation, don't you? That more hands build bigger and better systems than single hands?
@naifwonder
so, in terms of reporting, what are your requirements?
for example, do you want interest charges applied to forex positions?
do you want to have a report split by currency of your traded assets?
what do you look for?
cheers
so, in terms of reporting, what are your requirements?
for example, do you want interest charges applied to forex positions?
do you want to have a report split by currency of your traded assets?
what do you look for?
cheers
What is your point?
a) the OP never mentioned FX positions, in fact he mentioned stocks.
b) how is the question whether interest charges are applied to an asset class or how reporting looks like related AT ALL to the questions the OP asked about the system architecture?
c) Why does ET attract dummies like shit does flies?
a) the OP never mentioned FX positions, in fact he mentioned stocks.
b) how is the question whether interest charges are applied to an asset class or how reporting looks like related AT ALL to the questions the OP asked about the system architecture?
c) Why does ET attract dummies like shit does flies?
Quote from ghostrider77:
@naifwonder
so, in terms of reporting, what are your requirements?
for example, do you want interest charges applied to forex positions?
do you want to have a report split by currency of your traded assets?
what do you look for?
cheers
Quote from amazingIndustry:
What is your point?
a) the OP never mentioned FX positions, in fact he mentioned stocks.
b) how is the question whether interest charges are applied to an asset class or how reporting looks like related AT ALL to the questions the OP asked about the system architecture?
c) Why does ET attract dummies like shit does flies?
Quote from naifwonder:
- What are some good frameworks, if any, that one can use to analyze trading system performance based on a log of trades? I'm language agnostic when it comes to this since I am more interested in evaluating the programmatic logic then using the code itself.
@amazingIndustry ... did you read the post at all? Have you ever traded stocks in multiple currencies? How does your trading account look like?
Cheers mate
you are asking details about tiny implementation details that really have NOTHING to do with how the overall design architecture looks like. Whether you later on display stocks separated by base currency is utterly unrelated to how to design a proper OMS or how to store and retrieve statistical strategy results. The one who did not read the post properly was you not others.
The OP's emphasis all along was on FRAMEWORK, you quoted him yourself. So stop BSing around!!!
The OP's emphasis all along was on FRAMEWORK, you quoted him yourself. So stop BSing around!!!
Quote from ghostrider77:
@amazingIndustry ... did you read the post at all? Have you ever traded stocks in multiple currencies? How does your trading account look like?
Cheers mate
Quote from naifwonder:
I've been working on testing out some ideas that involve analyzing a great deal of data across multiple assets. The amount of data to be analyzed in combination with the nature of the systems I wanted to build made my project unfeasible for typical retail trading platforms. As a result, I built my own platform and framework for building/testing trading systems. At the moment, it can process 10 years of intraday data for several hundred stocks within a few seconds. As great as this is working, I am trying evaluate scenarios in which such a platform might fail. Do any of you guys have experience with undertaking such a project? Some of my questions are:
- What are some areas of a custom platform that you find to be particularly error prone?
- What are some things to be extra weary of while building the layer to interact with brokers?
- What are some good frameworks, if any, that one can use to analyze trading system performance based on a log of trades? I'm language agnostic when it comes to this since I am more interested in evaluating the programmatic logic then using the code itself.
- What stability and maintenance issues have you run into while undertaking such a project?
The big thing is to have a solid architecture and all the rest is details.
If you have a sound overall architecture, you can always re-implement any parts that are problematic.
I agree with you that having one's own platform can be a huge advantage, mainly because all of the existing ones, even those selling or leasing for big dollars, seem grossly inadequate to me in areas like cross-asset correlations etc., and even in simple areas like being able to handle all of the spreads and other order types that I want to model, test, and trade.
Quote from ghostrider77:
I wonder, why did you write your own platform for this?
Thanks
Conventional trading platforms are designed for analyzing smaller data sets, and almost always require you to load all the data into memory that you want to test against. For example, if you want to build a system to testing trading 1 minute bars from October 1, 2001 to October 1, 2011 for 20,000 different combinations of stocks, futures, and options, you will find that your computer will simply crash and explode. Or, if it is a powerful enough computer, take extremely long (days or weeks) to finish the calculations.
The platform I built can analyze such large data sets without breaking a sweat in a matter of minutes, but still has the conventional coding perks of most trading platforms such as being able to call code such as get_bar("stock/future name", "time frame", "start time"); that many "raw-coding" approached to analysis lack. I can also do things like test_system({"stock1", "stock2", "stock3"}, "start time", "end time");. Where it says stock1, stock2, and so forth, I can specify several hundred stocks. I can also run statistical calculations on sets of data without loading up a chart for each set, or being bound by the performance bottlenecks of software not optimized for analyzing quote data.
Conventional retail platforms are pretty good for what they are intended to do. But for analysis like this, they won't cut it. I've already built a system using this platform that would have been border-line impossible on a traditional retail platform, but want the platform to undergo more strenuous testing before progressing further.
Hey guys, I can in someway understand the jump on Ghostrider, but he is asking perfectly valid questions. I'm the CTO for my company, and I've learned through numerous discussions during the software development cycle that different people approach the feedback process in very different ways. I appreciate everyone's involvement in the discussion, but let's keep things focused on exchanging ideas on platforms, and not ideas regarding each other.
As for your question regarding interest on forex positions, I do have a module that can handle calculating account data based on positions. However, I am not trading forex right now or in the forseeable future, so this is not something I have focused on building further. I am considering going diversifying into assets in other currencies, but analyzing the currency relationships between positions adds a layer of computational complexity that I do not want to tackle until I have first further developed the more basic parts of the project.
Also, you had asked before why I did not use readily-available open-source solutions out there. My platform does use several open-source technologies for different parts. In terms of not using a complete open-source platform, in part it was because I really didn't like the open-source solutions. amazingIndustry mentioned the quality of the code; Open-source software does not inherently go through the type of stringent error testing that enterprise-level software requires. This certainly isn't always the case, but I did not find any open-source projects that had the right combination of testing and features. Also, many of them were abandoned projects. In addition, building the platform itself was a very educational process for me, and one that has contributed greatly to my knowledge base. That in itself made building the platform very rewarding. Lastly, because of the features of the platform, I did want the option to one day potentially lease out this software if it works well. Many open-source solutions have licensing requirements that would prohibit me from doing so. In an ideal world, I'd just make profits from the systems I build on the platform itself. But, if you can't find gold, have a damn good shovel you can sell to cover the costs of trying.
Quote from ghostrider77:
@naifwonder
so, in terms of reporting, what are your requirements?
for example, do you want interest charges applied to forex positions?
do you want to have a report split by currency of your traded assets?
what do you look for?
cheers
As for your question regarding interest on forex positions, I do have a module that can handle calculating account data based on positions. However, I am not trading forex right now or in the forseeable future, so this is not something I have focused on building further. I am considering going diversifying into assets in other currencies, but analyzing the currency relationships between positions adds a layer of computational complexity that I do not want to tackle until I have first further developed the more basic parts of the project.
Also, you had asked before why I did not use readily-available open-source solutions out there. My platform does use several open-source technologies for different parts. In terms of not using a complete open-source platform, in part it was because I really didn't like the open-source solutions. amazingIndustry mentioned the quality of the code; Open-source software does not inherently go through the type of stringent error testing that enterprise-level software requires. This certainly isn't always the case, but I did not find any open-source projects that had the right combination of testing and features. Also, many of them were abandoned projects. In addition, building the platform itself was a very educational process for me, and one that has contributed greatly to my knowledge base. That in itself made building the platform very rewarding. Lastly, because of the features of the platform, I did want the option to one day potentially lease out this software if it works well. Many open-source solutions have licensing requirements that would prohibit me from doing so. In an ideal world, I'd just make profits from the systems I build on the platform itself. But, if you can't find gold, have a damn good shovel you can sell to cover the costs of trying.
This is a mistake. Build it in from the start. Adding it later can be a nightmare if you haven't designed your architecture for it from the beginning.Quote from naifwonder:
...analyzing the currency relationships between positions adds a layer of computational complexity that I do not want to tackle until I have first further developed the more basic parts of the project.
Every security is a pair: SPY/USD for example. Ricardo on his deathbed acknowledged this fact to Malthus, and it has been common knowlege ever since. You need a numeraire for every security, you limit yourself if you just assume it your account base CCY.