Lets Roll Our Own Software!@

[darthtrader3.0beta 10]

There are obviously flaws in the business models of comercial trading software. Namely for the people who post on this site its that you can't stay in business without catering to the newbs who it would be alien to not have MACD or RSI as a base...The popularity of this board though shows there is an untapped market for software that has no concern for that kind of uselessness.

There are alot of smart people here who I assume probly know other smart people who don't post here. I can think of 3 mad "scientists" I'm friends with who don't post here but a software project would get them to.

James has already locked in probly the best marketting name for such software that at least I can imagine..."Traders Laboratory".

I assume some other people here have seen the TickZoom guys open source C# software ideas on elite...the code has been released but seems to have died on the vine. He should have come here instead of elite

It would probly be quite an interesting "experiment" if we invite the TickZoom guy to here to work on "Traders Laboratory"...he is clearly a monster software engineer but doesn't appear to be a monster as far as a project lead.

The project lead should be our collective "swarm intelligence" here.

[Soultrader 12]

Quite interesting darthtrader. I would be impressed if you can pull this off but want to make sure there is a disclaimer for such products using Traders Labortory as a name since there is no way for me to support it. Nor do I want any problems arising from order related problems that can cause traders to lose money. Its quite a task you are trying to move forward with but let me know if you need any help.

[darthtrader3.0beta 10]

Quite interesting darthtrader. I would be impressed if you can pull this off but want to make sure there is a disclaimer for such products using Traders Labortory as a name since there is no way for me to support it.

 

No, No, that is missing the "context" of what I am proposing...The "context" of what I'm going for has to do with that I'm completely obsessed with ants. No metaphor, the little black creatures. Everynight I read some of Holldobler and Wilson masterpiece on ants..What you quickly get from this is a feel for how much better "swarm intelligence" is than a single human brain.

http://en.wikipedia.org/wiki/Swarm_intelligence

 

If you realize it or not, you have built a swarm intelligence trading network here that crushes anything else done previously. In that context to say you can't support such software is obviously wrong. Its like saying that this place would be better if you could reply to every thread, but that is obviously false..the "swarm" here knows far far more than both you or I taken singlely..

To me that right there is the flaw/opportunity with trading software.

Me "pulling this off" in that context makes no sense...you would obviously need to be the Queen Bee though. I'm only interested in being a worker bee.

The Queen bee though doesn't plan anything like a human project manager..it just gets laid all day..

I just think that a decentralized, open source trading software can not be beat because of the inputs, especially if the inputs come from this board.

[Soultrader 12]

Thanks for the input darthtrader. Youre right about this.... most softwares are limited in different ways due the the lack of resources and support each firm can provide. A free open source version definitely has the potential. I currently do not know enough about open source software laws but will look into this. I am only concerned about protecting TL in case something goes wrong like a trader losing $$$$$ due to software issues and who is the one to blame? I guess this can be overcome by a simple digital agreement for software usage?

[BlowFish 10]

I know (or 'eknow') a couple of people that have rolled there own. One of the more impressive is (racks brain......) Dcraig over at that other forum. There are various toolboxes and open source bits that can be cobbled together (don't underestimate the cobblng however).

 

Btw if you are interested in swarms check out BioComps Dakota I almost bought it to play with....really couldn't justify it though.

[Eiger 10]

I'd really like software that does exactly what i want and can handle some of the calcs needed and not hang up. I am looking at this as a potential foundation or 'shell' from which to build a better mousetrap.

 

http://www.modulusfe.com/stockchartx/

 

It's not building from the ground up, but seems to offer a useable framework to build on. My major problem is that I know zip about programming, so I would need a counterpart to help. Any thoughts on this or does anyone have any experience with this software?

[AgeKay 10]

While this would obviously be great, I see two reasons why this is highly unlikely:

 

1. There is no incentive for the programmers unless they are getting paid. If they do create something really useful then they would probably just keep it to themselves.

 

2. Most programmers do not understand trading and most traders do not understand programming. This explains why all the trading platforms out there are so crappy. NinjaTrader is a an example of a trading platform written by programmers who do not understand trading (check out their "advanced" matching algorithm on the simulator, ridiculous). X_Trader is an example of a trading platform designed by a trader (Harris Brumsfield) who does not know what is actually possible with computers (limited charts and very limited in performance analysis). I have not used CQG, but I am sure you will get eye cancer from using it over a long period of time ;-) (their designers must be blind).

 

I just checked TickZoom. They seem to use the open source ZedGraph component for charting. I have used this for my own software before too since it is extremely flexible and very well designed, but found it to be buggy and having poor performance when scrolling on a lot of data and or using it in real-time. So I am currently rewriting all of my charts in WPF 2D Graphics which are hardware accelerated out of the box since it uses DirectX (the API that games use for their drawing) underneath. You just can't beat the performance unless you writing directly the DirectX or the graphics card. It's really amazing.

 

StockChartX looks good though. I would definitely use their WPF version though because of the hardware acceleration.

 

BlowFish, can you please provide a link to "Dcraig over at that other forum". I'd like to check it out. I've googled for it, but didn't find anything.

[darthtrader3.0beta 10]

While this would obviously be great, I see two reasons why this is highly unlikely:

 

Well I don't disagree with you at all. From thinking about it more, its better to forget the TickZoom guy. From what I've read he seems like a great software engineer, his charting stuff is a secondary to producing a monsterly fast tick backtesting engine. My open source suggestion was just because thats what tickzoom aims to be, but my experience is open source is really only good at mostly producing half finished software and sometimes you get an outlier.

 

Forgetting opensource, I think a more interesting idea is how feasible would it be to make a traderslab "in house" software? That way there is no delusions of marketting to a losest common demoninator like commercial software but it still has the "feel" of an open source project as far as the "swarm intelligence" here.

 

The basic guts of that software I would view as

you have a datefeed

that datefeed writes to a database

then you pull a chart from that database

 

My biggest problem with current software is it seems to be datafeed -> chart -> then maybe think about storing data because most people don't care about storing data. Even though its obviously much faster to pull from a database than to go out across the internet to retrive data.

[darthtrader3.0beta 10]

"Dcraig over at that other forum". I'd like to check it out. I've googled for it, but didn't find anything.

 

Dcraig on elite...If you search for him and delta stuff you should find his charts. They are pretty badass.

As far as .net charting, you ever check out the commercial chart packages?

Seems like that would be the path to take, I'm sure this is where the rubber meets the road for guys like ninja as far as headaches.

 

An interesting excersise here would be descibe your ultimate trading software, forgetting about any constraints...

 

mine is:

datafeed ->

into an HDF5 database, capturing everything i'm paying for as far as data goes ->

a flexible charting package

at the other end a backtester that can access the database

Somewhere in there is execution but that has to be after the fact.

 

The obvious issue in that equation for something like this getting off the ground is the datafeed. That datafeed is what puts us on different pages.

For that to me you have to go with DTN starting out capturing SPY tick data...only real dirt cheap non shuttled data that is feasible for the price as far as adding to minimal overhead that would not mess with anyones trading if this project takes 5 years.

[darthtrader3.0beta 10]

I am only concerned about protecting TL in case something goes wrong like a trader losing $$$$$ due to software issues and who is the one to blame?

 

I have no idea. It seems to me like how Jim Cramer can give an hour of investment advice, then run the blurb...You don't see every trading board though append the blurb to each post though. Obviously that kind of thing as far as software glitches is covered in the license agreement..it doesn't make much sense to be worried about that without appending "this does not constitute investment advice" to every post on this board.

[darthtrader3.0beta 10]

Btw if you are interested in swarms check out BioComps Dakota I almost bought it to play with....really couldn't justify it though.

 

Meh, its not like this area of AI is not well studied. It just is an area that does not lend itself well to commercial sofware. I would never consider a software that pretended to intigrate trading and AI...

Just look into SVM

http://en.wikipedia.org/wiki/Support_vector_machine

Why is there no commercial software that "yells" about SVM?

Obviously, because its much easier to go yell about things you dont understand, like neural networks, whatever.."

while you will keep the real "AI" to yourself., most importantly "if you have it all"?" then market it as something advanced.

That is the trading software "market" to me in a nutshell....

[AgeKay 10]

As far as .net charting, you ever check out the commercial chart packages?

 

I've looked at Dundas Charts. Look very good (visually). They have been around forever and seem to offer every chart under the sky. But I have not personally tried their charts. Using an external charting component, you always have to worry about flexibility and performance. For example, I was completely blown away by the flexibility of ZedGraph, but like many other charts, they are not very good for real time financial data.

 

An interesting excersise here would be descibe your ultimate trading software, forgetting about any constraints...

 

My ultimate trading software is basically something that helps me trade my personal strategy optimally. So it will look different for every strategy. Sure, it will use the same components underneath, but it would essentially be completely different for every strategy. I think it would be very hard to develop something that would suit everyone. This only works well for software platforms based on technical indicators (like NinjaTrader), which I am not a big fan of.

 

 

 

For that to me you have to go with DTN starting out capturing SPY tick data...

 

I looked at DTN. You need to pay like $600 just to get access to their API (they say it's for "support"). You cannot even have a look at their API documentation before you buy it. To get access to the TT feed (through the X_Trader API), you need X_Trader Pro, which costs depending on your broker around $1250-$1500 per month. Zen-Fire would probably be cheaper but I don't know whether it is even available outside of NinjaTrader.

[lynx23 10]

Meh, its not like this area of AI is not well studied. It just is an area that does not lend itself well to commercial sofware. I would never consider a software that pretended to intigrate trading and AI...

Just look into SVM

http://en.wikipedia.org/wiki/Support_vector_machine

 

I checked out that SVM link you provided, it seemed very interesting. I've messed around with neural networks in the past, with no real success with regards to algorithmic trading.

 

It seems to me that the hard problem is figuring out how to represent price data as "points in n-dimensional space", in a way that makes sense for the algorithm. I had the same problem with neural networks, figuring out how to translate prices into something a NN could use as meaningful input.

 

I'm a programmer by trade, but I have very little experience with AI so maybe I'm just dense... :doh:

 

As far as making our own trading software goes, it's a really big job. It the classic case where the potential users think it's easy because they (we) are very familiar with the problem domain, and can sum it up quickly when asked. But in reality, when you consider the myriad details it would be an almost impossible undertaking to make something with a small team in a reasonable period of time that does decent charting, indicators, analysis, data management and automated order entry.

 

There's a reason why all the commercial packages seem to fall short, and it's not because they all don't know what they are doing. Rather, I'd blame short release cycles and pressure from competition to keep up with the latest bullet point features. Not to mention differing ideas of what is most important in a software package.

 

Although I do sort of suspect that the financial industry tends to have sub-par programming for some reason. Perhaps because the programmers are hired and managed by finance MBAs and not technical folks. No offense to any programmers or MBAs out there, it's a just general impression.

 

For me the most important thing is that I want to be able to code my strategies in C# or some other "real" language, rather than some kind of brain-dead "easy" scripting language.

[AgeKay 10]

I just found out about DynamicDataDisplay - an open source charting library written in WPF.

 

This is a library of WPF controls for dynamic data visualization. It features efficient binding mechanisms and real-time interactivity capable of charting millions of data points. Current release allows flexible drawing of line/marker plots.

 

DynamicDataDisplay controls can be used in your application in the same way as other WPF controls. Any data can be used as a source of coordinates on your chart. It can be an array, a DataTable or a function. Every change in the data source causes update of the chart.

 

I have not tried it personally yet. Has anyone tried this?

 

P.S.: I am completely dumbfounded by the performance of 2D Graphics in WPF (using DrawingVisuals). I have more than 20k drawings on my chart and when scrolling the chart quickly the drawings are rendered faster than my eyes can conceive. Absolutely amazing. And best of all, my CPU just sleeps because the GPU is doing all the work. And I am not even virtualizing the drawings yet.

[darthtrader3.0beta 10]

As far as making our own trading software goes, it's a really big job. It the classic case where the potential users think it's easy because they (we) are very familiar with the problem domain, and can sum it up quickly when asked. But in reality, when you consider the myriad details it would be an almost impossible undertaking to make something with a small team in a reasonable period of time that does decent charting, indicators, analysis, data management and automated order entry.

 

Very true...I guess the more and more I've thought about this, the biggest reason this is impossible is the same reason retail software will always fall short for yourself. Since everyone does something different trading wise, most the resources are going to be wasted on things that have nothing to do with how YOU trade. Even if you go a non commercial route where you could strip away having to cater to the lowest common denominator, there would still be massive wastes from every person's individual perspective and no one would be happy.

[darthtrader3.0beta 10]

P.S.: I am completely dumbfounded by the performance of 2D Graphics in WPF (using DrawingVisuals). I have more than 20k drawings on my chart and when scrolling the chart quickly the drawings are rendered faster than my eyes can conceive. Absolutely amazing. And best of all, my CPU just sleeps because the GPU is doing all the work. And I am not even virtualizing the drawings yet.

 

That does look pretty sick. This kind of thing does seem to be the future. I seen some videos of something like this for matlab in 3d doing stuff gpu vs cpu and it was off the map. I guess it makes sense when you consider the complexity of what is going on in a modern computer game vs how simple what we are trying to do.

[darthtrader3.0beta 10]

Just wanted to add I found this amazing blog from the guy who made MATLAB2IB.

 

http://tradingwithmatlab.blogspot.com/

 

If you check out the code for creating market profile from tick data in matlab its pretty stunning how simple it is.

He also has code for replicating the VIX...considering matlab can do that, it would imagine it blows away any retail software as far as breadth goes.

Seems like if you really want to roll your own software its pretty sensless to not go with matlab/IB..the real grunt work is already done.

$300 bucks a year.. They also have software for connecting matlab to TT and are working on a generic hookup to FIX.

[tickzoom 10]

While this would obviously be great, I see two reasons why this is highly unlikely:

 

1. There is no incentive for the programmers unless they are getting paid. If they do create something really useful then they would probably just keep it to themselves.

True. Valid point. TZ is bent on being open source and also generating revenue to solve this issue. Those seem to conflict but a compromise was found that seems to make most users happy. Now, everything is open source in TZ and just switched to LGPL except for the engine that calculates all the bars and fires all the events is closed source. That has all the technology to process ticks at ZOOM speed.

 

2. Most programmers do not understand trading and most traders do not understand programming. This explains why all the trading platforms out there are so crappy.

 

Agreed!! As the author of TickZOOM, i can't agree more. I tried all those you mentioned myself. However I'm one of the rare people who is both a successful trader and many year software developer. So I have built TZ ground up to fix the major flaws that I see in other systems as a trader and developer.

 

However, it still lacks some bells and whistles in the GUI. So it's all driven programmatically at present. But that gap will close. Plus, it has been frustruting on other systems that certain features were impossible to control programmatically. On TickZOOM the requirement is that is must be programmatically accessible first, GUI comes second.

 

Also the TradeLink open source project is joining forces to close the gap and that will add support for many more brokers and data feeds.

 

NinjaTrader is a an example of a trading platform written by programmers who do not understand trading (check out their "advanced" matching algorithm on the simulator, ridiculous). X_Trader is an example of a trading platform designed by a trader (Harris Brumsfield) who does not know what is actually possible with computers (limited charts and very limited in performance analysis). I have not used CQG, but I am sure you will get eye cancer from using it over a long period of time ;-) (their designers must be blind).

 

 

I just checked TickZoom. They seem to use the open source ZedGraph component for charting. I have used this for my own software before too since it is extremely flexible and very well designed, but found it to be buggy and having poor performance when scrolling on a lot of data and or using it in real-time.

Good point. BTW, the reason I deliver the source code of ZegGraph with TickZOOM is that I have fixed the bugs in ZegGraph itself so it works awesome now with TZ--runs at zoom speed.

 

If you watch the video demo that processes 10million ticks of 5 years of data in 40 seconds, that includes the time to create the chart with ZedGraph. So there's no issues there whatsoever.

 

Also, I built a special feature for real time charting. It centers the last bar on the screen and then let's the bars build across to the right without any scrolling until they get close to the edge. At that point it very smoothly scrolls the current bar back to the center.

 

It also readjusts if the market rallies or sells off such that the price is heading off the screen. It readjusts automatically and smoothly completely hands free.

 

I did that because I liked having channels on the screen which extend 15 or 20 bars forward and wanted to see price filling up the channels and wanted my hands free for trading rather than screen positioning. It's rather sweet.

 

But on the other hand, the only discretionary control is market by or sell through the GUI. Again the GUI is limited for the moment.

 

So I am currently rewriting all of my charts in WPF 2D Graphics which are hardware accelerated out of the box since it uses DirectX (the API that games use for their drawing) underneath. You just can't beat the performance unless you writing directly the DirectX or the graphics card. It's really amazing.

 

That IS very cool. I must completely admit. Wow.

 

StockChartX looks good though. I would definitely use their WPF version though because of the hardware acceleration.

 

BlowFish, can you please provide a link to "Dcraig over at that other forum". I'd like to check it out. I've googled for it, but didn't find anything.

[tickzoom 10]

Well I don't disagree with you at all. From thinking about it more, its better to forget the TickZoom guy. From what I've read he seems like a great software engineer, his charting stuff is a secondary to producing a monsterly fast tick backtesting engine. My open source suggestion was just because thats what tickzoom aims to be, but my experience is open source is really only good at mostly producing half finished software and sometimes you get an outlier.

Thanks. I'm glad to find people who agree that open source isn't always the Panacea that some think it is with many people altruistically contributing code.

 

TickZoom wants to be open source however so that users can see the code and tweak it or otherwise.

 

But the project must generate revenue to keep innovation and provide support moving which is your point, I think.

 

To that end, the solution was to keep only part of the system which provide the ZOOM speed streaming technology for generating bars and firing events as closed source and requiring a low cost license.

 

Forgetting opensource, I think a more interesting idea is how feasible would it be to make a traderslab "in house" software? That way there is no delusions of marketting to a losest common demoninator like commercial software but it still has the "feel" of an open source project as far as the "swarm intelligence" here.

 

The basic guts of that software I would view as

you have a datefeed

that datefeed writes to a database

then you pull a chart from that database

 

Interesting. That's how TickZOOM works but it also does what you say below simultaneously--best of both worlds.

 

Like this: TickZOOM has a Windows Service which captures ticks from your data provider or broker and writes them to a file for history.

 

That way, you can start or stop your GUI and do back testing, without missing any ticks. You can even run that data service at a data center for 24/7 reliability.

 

When you want to trade real time, the GUI talks to the data service to get history its missing and the real time feed.

 

My biggest problem with current software is it seems to be datafeed -> chart -> then maybe think about storing data because most people don't care about storing data. Even though its obviously much faster to pull from a database than to go out across the internet to retrive data.

 

Good point.

[tickzoom 10]

Very true...I guess the more and more I've thought about this, the biggest reason this is impossible is the same reason retail software will always fall short for yourself. Since everyone does something different trading wise, most the resources are going to be wasted on things that have nothing to do with how YOU trade. Even if you go a non commercial route where you could strip away having to cater to the lowest common denominator, there would still be massive wastes from every person's individual perspective and no one would be happy.

 

I went through the same analysis myself before building TickZOOM and was forced to make my own software. The tick filter for bad ticks in one commercial system was terrible and not user replaceable. The performance at tick testing with another was bad ,etc.

 

What made it so difficult was they were closed source.

 

I don't think TickZOOM is the end all solution to this dilemma.

 

However, it solves a specific problem which you can extremely rarely find programmers qualified to solve. That is raw speed in processing tick data to create bars that are updated per tick.

 

TickZOOM zooms through 10million ticks in 40 seconds while updating tick, minute, and hour bars for every tick while updating daily, weekly, monthly, yearly every minute.

 

With streaming technology, it has zero memory constraints.

 

It offers range, volume, change, ticks, point & figure bars, and more completely configurable to use many different time frames simultaneously.

 

It has rolling bar technology and configurable resets for bars.

 

Plus it hungrily seek other ideas for bar types because it would love to support every bar type imaginable.

[tickzoom 10]

My ultimate trading software is basically something that helps me trade my personal strategy optimally. So it will look different for every strategy. Sure, it will use the same components underneath, but it would essentially be completely different for every strategy. I think it would be very hard to develop something that would suit everyone. This only works well for software platforms based on technical indicators (like NinjaTrader), which I am not a big fan of.

 

Good point. It's not only very hard. It's impossible. For example, some people, amazingly, don't like Google search for example.

 

So, software authors must always select their target market or market(s) and focus features on what they request and need.

 

TickZOOM is attracting an interesting lot so far.

 

3 crowds: do you fit in any of these?

 

1. Traders who are also programmers are using TickZOOM an loving it since it's the only platform out there right now which has nearly entirely open source code and also a zoom speed testing and trading engine.

 

2. Discretionary traders who don't know how to program want TickZOOM for its historical testing speed. But at the moment they're stymied by the lack of documentation and have posted tickets to that effect. So we hope to satisfy them eventually and working on that currently.

 

3. Mid and small size institutions or prop shops who can't afford the 6 figure priced commercial system but have programmers on staff to add anything they need to TickZOOM.

 

Interestingly, as programmers all these people recognize how hard (and sometimes painful) it is to build high speed software with multi-threading, quad core support and other high performance, near real time features.

So they're willing to pay for that part of the system.

 

Out of those groups, the programmers already have most of what they need with TickZOOM except more speed options like GPU support and multi symbol support (comming soon). After that, they will be mostly satisfied.

 

So the focus will become building a more elaborate training and GUI for the non-programmers. That is far easier (any body can do it almost) than building the engine.

 

It seems the focus will be more towards teaching the non-programmers who to do things programmatically because it gives far greater options.

 

It's a farce to use a GUI interface to write a strategy--as we all know.

 

Wayne

[AgeKay 10]

Good point. BTW, the reason I deliver the source code of ZegGraph with TickZOOM is that I have fixed the bugs in ZegGraph itself so it works awesome now with TZ--runs at zoom speed.

 

If you watch the video demo that processes 10million ticks of 5 years of data in 40 seconds, that includes the time to create the chart with ZedGraph. So there's no issues there whatsoever.

 

I had fixed a few bugs myself in ZedGraph too, but the performance was still not acceptable for me. How is your performance if you zoom into your 10 million ticks and then scroll quickly to either side? I would be surprised if you get smooth scrolling. If so, did you do any performance optimizations inside ZedGraph too?

 

 

So I am currently rewriting all of my charts in WPF 2D Graphics which are hardware accelerated out of the box since it uses DirectX (the API that games use for their drawing) underneath. You just can't beat the performance unless you writing directly the DirectX or the graphics card. It's really amazing.

 

I am willing to share a few gotchas/insides I have found about writing the most performant 2D Graphics in WPF if someone wants to go that route too. It's mostly based on measuring and comparing the performance of my own code and stepping through the .NET Framework Source Code. My current implementation is impossibly fast even though I am not even virtualizing yet.

 

It offers range, volume, change, ticks, point & figure bars, and more completely configurable to use many different time frames simultaneously.

 

Oh, that's interesting. I somehow missed that when I checked out your website. Do you have screenshots of these charts?

 

Plus it hungrily seek other ideas for bar types because it would love to support every bar type imaginable.

 

I've come up with a couple chart types you will not have seen anywhere but they are not based on bars and I am not willing to share them either ;-).

 

1. Traders who are also programmers are using TickZOOM an loving it since it's the only platform out there right now which has nearly entirely open source code and also a zoom speed testing and trading engine.

 

What do you mean when you say "zoom speed testing"?

 

3. Mid and small size institutions or prop shops who can't afford the 6 figure priced commercial system but have programmers on staff to add anything they need to TickZOOM.

 

Does not make sense to me. Your programmer, especially if you have more than 1, will cost you more than (low) 6 figures per year, so why not just buy the commercial system?

[BlowFish 10]

Hey TickZoom, I'm sure your contribution here is welcome but as your core engine ins proprietary and costs a monthly lease I'm pretty sure that promoting it here (unless you are a sponsor) is a no no. (I know you aren't explicitly promoting it but it seems to be mentioned every other paragraph )

 

More importantly doesn't building a 'roll your own' on a locked propitiatory engine kind of defeat the object some what? Again don't want to diminish your contribution but can't help feeling that you might have a secondary agenda that may or may not be appropriate here. It might be worth having a chat with James (Soultrader)

[darthtrader3.0beta 10]

Again don't want to diminish your contribution but can't help feeling that you might have a secondary agenda that may or may not be appropriate here.

 

I think we need to put the breaks on "secondary agendas" here. It clearly drove away that poker guy.

I'm very glad to see tickzoom here...Keep in mind I didn't invite him here like I posted earlier in this thread, I guess someone else pointed him here. If you followed the elitetrader threads at all what you quickly see is he just really likes to talk If you didn't follow those threads it will look like aggressive marketting.

I would like to ask James to consider that this is early alpha stage software that is clearly trying to do something different. I can't see how you can really be a sponsor until you have a first release...In the mean time he seems extremely open to ideas, so bomb away.

[tickzoom 10]

Hey TickZoom, I'm sure your contribution here is welcome but as your core engine ins proprietary and costs a monthly lease I'm pretty sure that promoting it here (unless you are a sponsor) is a no no. (I know you aren't explicitly promoting it but it seems to be mentioned every other paragraph )

 

More importantly doesn't building a 'roll your own' on a locked propitiatory engine kind of defeat the object some what? Again don't want to diminish your contribution but can't help feeling that you might have a secondary agenda that may or may not be appropriate here. It might be worth having a chat with James (Soultrader)

 

My point for this thread was just that everyone has indicated the downsides of buying a commercial app, the downsides of using open source software and also the downsides of rolling your own software.

 

Where does that leave you?

 

Allow me to answer without mentioning any specific solution so it's more clear.

 

The answer is an "a la carte" combination of these options. Point is that "roll your own" for any common sense discussion actually means to "cherry pick" components to build your system. To totally roll your own and write every line of code yourself as an individual is nearly unthinkable.

 

For example, if someone rolls their own software, does it make sense to write their own graphics engine with so many commercial and open source ones out there?

 

Does it make sense to write a high speed tick processing engine?

 

Does it make sense to rewrite quantlib or talib, etc. from scratch?

 

Does it make sense to write your own code to communicate with your broker if TradeLink (open source and free) does that already with a easy api?

 

So you see, a la carte, is the way to go if you have the skills to program your own system.

 

As regards your question of "defeating the purpose" of rolling your own, consider this. In the "a la carte" mentality to rolling your own, it seems a smart trader must always analyze the buy versus build versus open source rational for any component.

 

The problem with open source was described above, but don't those same problems exist for rolling your own software only far worse?

 

When you roll your own, you have only yourself to turn to for any assistance if things don't work. And your code won't progress or add features or get bugs fixed unless you do it yourself.

 

So a la carte has enormous advantages since other people are working on, supporting, fixing, improving, the different parts of your system. And you can upgrade any individual part at will.

 

Sorry if I over emphasized a specific components. Hopefully this is more clear as the ideal direction to go if you have the technical ability to "roll your own".

 

Wayne