Deciding on a Backtesting and Trading Platform

[JackR]

Quote from Synonym:
Have TradersStudio given any idea when they aim to have the RT version ready?
Syn

In Feb, 2007 they said it would be released later in 2007.
In Aug 2007 they said there would be an 8 to 10 month slip.
In March of this year they said they were confident of the end of this year (2008), maybe the third quarter.
Last month they said they were changing the design to support multi-core processing.

I've never seen them make a schedule and at this time there doesn't seem to be a schedule so they can't miss it.

Nothing wrong with the current product, just don't expect real-time any time soon.

Jack

 

[fundjunkie]

Quote from JackR:

In Feb, 2007 they said it would be released later in 2007.
In Aug 2007 they said there would be an 8 to 10 month slip.
In March of this year they said they were confident of the end of this year (2008), maybe the third quarter.
Last month they said they were changing the design to support multi-core processing.

I've never seen them make a schedule and at this time there doesn't seem to be a schedule so they can't miss it.

Nothing wrong with the current product, just don't expect real-time any time soon.

Jack

Yes, this doesn't reflect very well on the management and operation control of the company. Fortunately for them, the application is solid at what it was designed for, so delivery of a real-time version isn't a huge issue for me. But it probably is for the macho-men out there who believe the only "real" trading is done intra-day with tick data.

I gave up long ago wondering when the real-time version would arrive. But I do need it. My understand is that when it finally emerges EOD strategy automation (precursor to intra-day automation) will become possible enabling me to get out of the decision making around entries and exits completely. The sooner that day arrives, the better


Thx
D

 

[Murray Ruggiero]

Quote from fundjunkie:

Yes, this doesn't reflect very well on the management and operation control of the company. Fortunately for them, the application is solid at what it was designed for, so delivery of a real-time version isn't a huge issue for me. But it probably is for the macho-men out there who believe the only "real" trading is done intra-day with tick data.

I gave up long ago wondering when the real-time version would arrive. But I do need it. My understand is that when it finally emerges EOD strategy automation (precursor to intra-day automation) will become possible enabling me to get out of the decision making around entries and exits completely. The sooner that day arrives, the better


Thx
D

We are working on real time, The truth is that we had a early beta ready and it's performance did not scale well as we added symbols. Instead of trying to patch up these issues I decided to go back and redesign things. I rather not release a real time product which I know can't handle more than someone trading only a few items.

In addition yes we have a management problem, me, I keep adding features to the end of day product. We have added many of them to the current beta. Some have been requested by customers , both indivdual and institutional and other I needed for my work putting together my hedge fund project. My fault is I keep adding things as I see they would be useful.

We have increase the size of our programming team so we can move forward more quickly , I realize our users have been waiting for both this new version and real time. This increase in staffing happened about a month ago, so they are now trained and we are moving forward quickly.

 

[Synonym]

Hi droskill
Thanks for your advice.

Quote from droskill:

No idea when they're going to get it out.

Seems it's in the pipeline and due to appear at some point...

API = Application Programming Interface - it's a means of accessing/controlling a given application. For instance, Quotetracker can use the Interactive Brokers API to get data from it.

TC, which I also use, doesn't do any of these things and never will. I think they're putting their money on their Blocks software (www.blocks.com). It's interesting. Definitely designed for people without programming experience. The backtesting is very simplistic - no portfolio backtesting, but it's ok for simple tests. I believe it works with the TC real-time data, but that feed, I believe, is limited to 1m intervals if that matters.

Thanks for clarifying and for the link. I do actually have Blocks (use it for sectors). Simplistic like you say. I don't actually like the way you move around in it; and the despite the clarity of the charting, i find working with it much more clunky than TC, particularly the drawing tools. I know that you use TS and Ami, a combo that has potential for me. I didn't know you had TC. If you'd don't mind me asking, I'd be interested to know what you use it for...scanning? PM me if you'd prefer.

Your chief obstacle is that every tool mentioned requires some knowledge of programming. And I don't think there is a way out of it.

I'm happy to have a basic understanding, enough to cobble together code for my ideas. But as i want to keep strategies simple in esscence, i thought i might be able to avoid the complexities of real programming...for the time being anyway.

Stratasearch is a nice tool but I find their basic approach rather questionable - scanning thousands of strategies to find one that works. It's nice, but it's not real time.

I didn't know that. No i don't find that appealling either!

Seertrading is interesting - but their support is typical of a European company - none. I've posted messages to their board and gotten no answers.

Poor. Not encouraging.

Tradecision is an interesting tool, but the backtesting, the last time I looked at it, it did not do portfolio-level testing.

It does now, and walkforward testing. But i'm not sure just how good its testing capabilities really are.

You'll need two programs at the moment.

If you're going for cheap, QT is impossible to beat. As you know, I'm a AB fan. One more to consider that I've been playing around with and like: NinjaTrader. For trade automation, I've found no good solution that doesn't require programming - NinjaTrader is very good in this area. AB can autotrade, but it takes a lot of programming work.

Right, thanks. Yes, im moving to the conclusion that two seems to be the way.

Like you, i'm a hawk on cost but not to the detriment of any useful qualities. Getting the best value is just good business sense.

So you're saying AB will be good for my charting and possibly scanning? Am i right in guessing that QT is comparable for those things? If so, might I be best with QT for the charting, as i wouldn't be needing the extra capability of AB (strat devel and testing), due to the fact that this work could be done on my other sw...e.g. TradersStudio?

I've heard mixed things about Ninja's support. Plus, a small thing, but one i think is fundemantal (at the moment), it doesn't support the logarythmic scaling of its charts. If i could code, maybe i'd find a way to make it.

I've found no one data feed that solves the problems. Tradersstudio uses CSI for a specific reason - the number of decimals associated with each equity and separation of dividends and splits so you can construct what TS calls "TS adjusted" stocks. This allows you to test into the past correctly. They're the only ones that offer this.

I would also mention the importance of a community. AB really has a great community (as does Trading Blox) - Tradersstudio has yet to develop that kind of support, but I have my hopes. NinjaTrader seems to have a nice community as well.

So can i use CSI data for AB too? If not, what do you use for it?

I've heard a fair bit on Trading Blox vs TradersStudio. People have great things to say about both. Ignoring cost and focussing on value, do you know of a reason to choose Blox over TS? For example, would top-end Blox mimise my need to learn coding and so save time? If so, it could be money well spent.

Thanks for your help.
Best
Syn

 

[Synonym]

Quote from JackR:

In Feb, 2007 they said it would be released later in 2007.
In Aug 2007 they said there would be an 8 to 10 month slip.
In March of this year they said they were confident of the end of this year (2008), maybe the third quarter.
Last month they said they were changing the design to support multi-core processing.

I've never seen them make a schedule and at this time there doesn't seem to be a schedule so they can't miss it.

Nothing wrong with the current product, just don't expect real-time any time soon.

Jack

Hi Jack

thanks for the post.
So, it will come, but it's not something to wait on.

Syn

 

[heng]

It has been very interesting for me to read the reviews by Gyles and the numerous comments by others. I have decided to buy some kind of trading system testing software and , as there are hardly any demos available, independent testing is best to rely on.
This past weekend I visited a Trading show in Germany and two of the speakers happened to be Curtis Faith and Thomas Stridsman.
Curtis Faith is one of the original Turtles and the developper of Trading Blox. So I asked him about his Trading Blox software. I was astonished to learn that he has sold off the software and has no more financial interest in it. He explained to me that TB is not a trading system, the developpment goal having been a first of class backtesting system which he thinks it still is. He also mentioned the less elaborate graphing capabilities. He also mentioned the programming language being close to VB. In all, it was at first disappointing as he seemed to warn me more as to what it is not instead of being enthusiastic about his baby. But thinking of it after his lecture, this looks rather like being his personality, so it is ok.
Thomas Stridsman lectured on trading system development and that's what he is world famous for. I asked him which software he uses to that purpose. The answer was Tradersstudio. He did not give any details or reasons and I could not drill down on it. But he mentioned other good systems like TradeStation and Trading Blox. Regarding the later, he added that it is probably the best system for backtesting and definitively the best for people who do not want to program.

 

[fundjunkie]

Quote from Murray Ruggiero:

We are working on real time, The truth is that we had a early beta ready and it's performance did not scale well as we added symbols. Instead of trying to patch up these issues I decided to go back and redesign things. I rather not release a real time product which I know can't handle more than someone trading only a few items.

Multicharts has this problem but chose not to address it. I think you're right to look at redesign. Without turning this into a workshop on TS planned RT design can I ask how you propose to address the twin requirements of a tick handling application:

- Ability to handle and store symbols in bulk (sequential write heavy operations)
- Ability to respond quickly for specific sybols actually being traded

A database will always slow the second part down. So I hope you're looking at doing somethig like cacheing tickfiles on disk (for trading) and perhaps later storing in a DB for storage and simulations etc. I struggle to see how performance can scale if you try to do everything from a relational DB (too much overhead). I expect every company out there struggles with this question.


Thx
D

 

[Synonym]

Quote from heng:

It has been very interesting for me to read the reviews by Gyles and the numerous comments by others. I have decided to buy some kind of trading system testing software and , as there are hardly any demos available, independent testing is best to rely on.
-----
Thomas Stridsman lectured on trading system development and that's what he is world famous for. I asked him which software he uses to that purpose. The answer was Tradersstudio. He did not give any details or reasons and I could not drill down on it. But he mentioned other good systems like TradeStation and Trading Blox. Regarding the later, he added that it is probably the best system for backtesting and definitively the best for people who do not want to program.

Hi Heng
Interesting comments. Thanks for sharing.
Syn

 

[droskill]

Quote from Synonym:

Hi droskill
Thanks for clarifying and for the link. I do actually have Blocks (use it for sectors). Simplistic like you say. I don't actually like the way you move around in it; and the despite the clarity of the charting, i find working with it much more clunky than TC, particularly the drawing tools. I know that you use TS and Ami, a combo that has potential for me. I didn't know you had TC. If you'd don't mind me asking, I'd be interested to know what you use it for...scanning? PM me if you'd prefer.

I use TC for data for AB - which is really the result of pure laziness. I'm looking at switching my AB over to CSI data so that I can have consistency, but I've got a lot of time spent coding some indicators in AB that depend on TC data.

It does now, and walkforward testing. But i'm not sure just how good its testing capabilities really are.

Looks more interesting having spent a bit more time on their website. My issue with Tradecision would be the price - and the lack, currently, of any apparent demo on their site. I remember having a demo version to look at, but they don't seem to offer that.

So you're saying AB will be good for my charting and possibly scanning? Am i right in guessing that QT is comparable for those things? If so, might I be best with QT for the charting, as i wouldn't be needing the extra capability of AB (strat devel and testing), due to the fact that this work could be done on my other sw...e.g. TradersStudio?

Well, there's no cost to trying QT - so you should try it out. I find AB's scanning quite good, but I do not use it intraday. Your mileage might vary if that's a strong need for you.

I've heard mixed things about Ninja's support. Plus, a small thing, but one i think is fundemantal (at the moment), it doesn't support the logarythmic scaling of its charts. If i could code, maybe i'd find a way to make it.

Hmmmm....not sure if it support log charts or not. I'll take a look.

I've heard a fair bit on Trading Blox vs TradersStudio. People have great things to say about both. Ignoring cost and focussing on value, do you know of a reason to choose Blox over TS? For example, would top-end Blox mimise my need to learn coding and so save time? If so, it could be money well spent.

I would say that the following are in Trading Blox's favor:

- Very good community.
- Modular approach to system development - so you develop money management, entry, exit, etc all as separate blocks. This makes it easier to try out different combos. Tradersstudio has a somewhat similar feature in Trading Plans, but for myself I find it rather clumsy.
- Can be very plug-and-play - add an entry, add an exit, add a money management technique and away you go. It might work better for you given your newness to coding.

Against TB:
- Cost: I've downloaded a demo of Trading Blox twice and I always have the same reaction: "that's it???" Meaning I just don't feel I get a lot of value from it. But that's just me - a lot of people think it is THE BEST platform, and in many ways I can see their point.
- Lack of progress: I feel like TB has been rather stuck in what they do, and I haven't seen a lot of advances in what the program can do.

But again, the people that own TB are religious about how they feel about it. I would suggest downloading a trial and then judging for yourself.

More info on this:
http://www.tradingblox.com/forum/viewtopic.php?t=5853

Good luck!

 

[Synonym]

Many thanks for the reply droskill.

Quote from droskill:
Looks more interesting having spent a bit more time on their website. My issue with Tradecision would be the price - and the lack, currently, of any apparent demo on their site. I remember having a demo version to look at, but they don't seem to offer that.

Yes those two things stand out. I'm not sure what it is, i just don't find their site very reassuring. Seems to have a lot of gloss, not so sure about what's under it. Plus, you cannot access their forums without a log in.

Well, there's no cost to trying QT - so you should try it out. I find AB's scanning quite good, but I do not use it intraday. Your mileage might vary if that's a strong need for you.

Very true. It's more a future consideration of future scalability really.

I would say that the following are in Trading Blox's favor:

- Very good community.
- Modular approach to system development - so you develop money management, entry, exit, etc all as separate blocks. This makes it easier to try out different combos. Tradersstudio has a somewhat similar feature in Trading Plans, but for myself I find it rather clumsy.
- Can be very plug-and-play - add an entry, add an exit, add a money management technique and away you go. It might work better for you given your newness to coding.

Against TB:
- Cost: I've downloaded a demo of Trading Blox twice and I always have the same reaction: "that's it???" Meaning I just don't feel I get a lot of value from it. But that's just me - a lot of people think it is THE BEST platform, and in many ways I can see their point.
- Lack of progress: I feel like TB has been rather stuck in what they do, and I haven't seen a lot of advances in what the program can do.

But again, the people that own TB are religious about how they feel about it. I would suggest downloading a trial and then judging for yourself.

More info on this:
http://www.tradingblox.com/forum/viewtopic.php?t=5853

Good luck!

Thanks for the advice and the link - i'd not until now checked out TB's forums. Enlightening. The community certainly seems to be an intelligent, experienced and supportive bunch. I find having anything that can speed up the coding side of things is really appealing.

I've now had a look at Stratsearch properly. Not something to replace TB or TStudio in my opinion. If i had the time, i'd be interested to take a trial and see how it's actually designed and what it can do. I have a lot of concerns i'd have to allay before i trusted it! But i think my times much better spent elsewhere at the moment.

Still to look into Seertrading. But if support is questionable then if just not worth it in my opinion.

Not yet made my decision but getting nearer i think.

Thanks again. Best of luck in your trading too!