Dilemma: A small private institution wants to develop an algorithmic trading system. They want to minimize lead time to develop the algo, but also want to minimize start up costs. The institution has one programmer in house but would like to use an algo platform to start with so they don't spent lots of lead time "reinventing the wheel". The algo will have to be able to handle high volume. The institution also wants the flexibility to choose the brokerage firm and clearing house used to handle the trades. In researching this, it seems like the institutional-quality algo platforms are products by Portware, FlexTrade, and Progress' Apama. When these platforms are described as "broker-neutral", does it mean that customers choose who handles and clears their orders, rather than the software tying the customer to having to use a specific broker and/or clearing firm? I know these platforms can be basic and very customizable by the customers, or customized up front by the vendors. Would getting a basic platform from one of these vendors be able to fulfill this institution's needs, and if so, about how much would it cost to purchase it? The Portware, FlexTrade, and Apama sites don't really give any pricing info. It seems like these are the main players for the non-retail customer. Have I missed any others that might be able to do the job?
I have read up on Apama and they are broker neutral as far as i can tell... and they have a flexible algo front end setup so the trader can change one or more basic parameters as well as more advanced actions... Much better, more cost effective, for a a mid tier shop to buy a software component, a good one after doing research dilligence, then it is to develop one... countless, nay many countless hours have been spent by top teams to produce this product... i think it is worth a consideration... as well as the possibility of complete project failure on the part of a mid tier firm programming shop completing the project at all or completing it and not getting the results they want... major out the door over the top costly... in energy time and money... best... cj... HAVE STOP <img src="http://www.enflow.com/p.gif"> WILL TRADE
I think you might consider it more important to do Total Cost Analisys on all the trades you make over a week/month and thereby bench test a number of Algo platforms... that way you would be way ahead of the curve instead of possibly trying to re-invent a very complex wheel by building it from scratch... example company that offers this... http://www.gtanalytics.com/Methodology.aspx?tabindex=-1&lbindex=2 my 1.5 cents... cj...
And another just in case you are not already bewildered by the choice. http://www.aegisoft.com/trading_athenatrader.htm
Actually StrategyRunner would not be a bad choice...they appear to have licensing agreements with many different brokers. The platform supports both gray and black box set-ups and it's datafeed is fast. The trading system interface is DLL-based, so it is very fast....however, VB6 or C++ are about your only language choices if you are building from scratch.