Scalable platform for running EasyLanguage strategies

Joel Reymont · Jan 22, 2006

Folks,

I would like you input on the platform that I'm designing. I'm thinking of building Uptick as a scalable backtesting and execution platform for EasyLanguage strategies. Open source by all means.

My unique proposition is to enable you to build a cluster of cheap boxes for running your strategies in a telco-grade fault-tolerant environment.

This will require me to write a translator from EasyLanguage to Erlang but this is straightforward. What is not clear to me yet is how to get the data into this environment.

I very much appreciate your input!

Thanks, Joel

horribilicus · Jan 22, 2006

Easy Language will be the best, and the worst, feature of this endeavor. Best, because it lets the thundering herd of Tradestation users "port" their systems with a minimum of difficulty. Worst, because Easy Language has numerous deficiencies

One-instrument-at-a-time simulation
No access to portfolio-level statistics such as portfolio equity, portfolio risk, sector equity, sector risk, etc.
No support for time-varying BigPointValue (instruments denominated in foreign currency whose exchange rate vs. USD fluctuates daily)
Inflexible commissions & slippage primatives (compared to WL and others)
Inability to express the concept of multiple systems trading simultaneously, sharing a common equity pool

I have grown rather fond of the TBB worldview, in which it is assumed that multiple systems with perhaps different portfolios are being traded simultaneously, and it is further assumed that thousands of parameter sets are being tested (for robustness checking, etc.) Here is an example.
http://www.elitetrader.com/vb/showthread.php?s=&postid=953233#post953233

Joel Reymont · Jan 22, 2006

Quote from horribilicus:
Easy Language will be the best, and the worst, feature of this endeavor. Best, because it lets the thundering herd of Tradestation users "port" their systems with a minimum of difficulty. Worst, because Easy Language has numerous deficiencies

This is not a problem at all. I will need to translate EL into something that my system can execute internally (Erlang). This implies that you can see what the translated version looks like and use it for further development, or you can do all your new development in the internal language while running your old unmodified EL code. It also brings forth the possibility of my extending EL to add the features that it's missing.

I have grown rather fond of the TBB worldview, in which it is assumed that multiple systems with perhaps different portfolios are being traded simultaneously, and it is further assumed that thousands of parameter sets are being tested (for robustness checking, etc.) Here is an example.
http://www.elitetrader.com/vb/showthread.php?s=&postid=953233#post953233

You should be able to do it ultimately. That is the whole point of my platform. Being able to run EL is expressly for the purpose of easing migration for the thundering herd of TradeStation users :-)

.

Nana Trader · Jan 22, 2006

Quote from horribilicus:

telco-grade fault-tolerant environment

[*]Inflexible commissions & slippage primatives (compared to WL and others)
[*]Inability to express the concept of multiple systems trading simultaneously, sharing a common equity pool[/list]I have grown rather fond of the TBB worldview, in which it is assumed that multiple systems with perhaps different portfolios are being traded simultaneously, and it is further assumed that thousands of parameter sets are being tested (for robustness checking, etc.) Here is an example.
http://www.elitetrader.com/vb/showthread.php?s=&postid=953233#post953233

What are these crap you posting? Hence someone stupid will come to defend you, majority won't takes time to figure out what you are trying to say, because they are not worth shit

Can't hightec like you trade, instead of developing products

horribilicus=Joel Reymont

Joel Reymont · Jan 22, 2006

I have to point out that I'm not horribilicus. A little investigation into our previous messages will surely prove that. With all due respect your comment is very much off-base.

Joel Reymont · Jan 22, 2006

Can't hightec like you trade, instead of developing products

You need to come up with the capital somehow ;-). Developing products also fills some psychological voids, etc. I'm trying to sort of do both, except I won't make money on the core software since it will be free.

Nana Trader · Jan 22, 2006

Quote from Joel Reymont:

I have to point out that I'm not horribilicus. A little investigation into our previous messages will surely prove that. With all due respect your comment is very much off-base.

Dude, you are very smart ass , but I can be more smart ass than you.

I added both your name and horribilicus as my buddy list, If you are not the same person, you both had to appear online at same time

Joel Reymont · Jan 22, 2006

Quote from Nana Trader:

Dude, you are very smart ass , but I can be more smart ass than you.

I added both your name and horribilicus as my buddy list, If you are not the same person, you both had to appear online at same time

Two people will appear online if they are online at the same time. Only one or none will appear if they are not. Lets stick to the topic, shall we?

What do you think of my design? Do you have any suggestions on how to get data into the system?

Nana Trader · Jan 22, 2006

Quote from Joel Reymont:

You need to come up with the capital somehow ;-). Developing products also fills some psychological voids, etc. I'm trying to sort of do both, except I won't make money on the core software since it will be free.

Free' huh? You bloody cockroach, isn't this your site
http://wagerlabs.com/
---------------------------------------------------
Who are my customers?

I see three groups of traders:

Banks, brokerages, hedge funds and outfits with their own programming staff. They either have their own software or the burden of supporting them is too large for me to handle.

Enterprising programmers. These are either die-hard developers who will insist on rolling their own or well-in-the-money .NET lovers who are already served by SmartQuant. I will try to serve the die-hard developers indirectly.

The TradeStation folk. This is a diverse group consisting of individual traders, power traders, small trading firms, hedge funds and everyone in between. A large portion of this group is dissatisfied with TradeStation for various reason and has migrated to WealthLab, NeoTicker[NT], etc.

Group #3 makes up the bulk of my market. These are my customers.

Why would my customers buy from me?

My unique selling proposition is a scalable and robust backtesting and execution platform that runs strategies written in TradeStationâs EasyLanguage.

Joel Reymont · Jan 22, 2006

@NanaTrader

This spirited discussion is not going in the right direction. I edited my blog post to clear up my intentions. To wit:

Yes, the core platform will be free. Open source, available to everyone, etc. Yes, I still treat it as a business and thus have customers who buy from me. I think that is the way to run a business even if it depends on open source.

I want to capture mindshare first and market share second. I cannot pull this off alone and thus want to encourage other developers to pitch in. I plan to make money on add-on products, support and services but so can everyone else.

Just offering the software for free won't bring people to my platform. I have to offer something unique and compelling which is what I'm trying to do. I'll give away some by giving away the platform and I will make some on services, support and add-ons. This is the open source business model and a way to build a trading war chest. I'll eat my own dog food and use my own software.

Do you have anything of substance to contribute re: topic of this thread or my post that started it?