Interesting software

[gbr128]

Have been thinking recently about software for trading, and was wondering what any readers of these boards would consider a "killer app" that's not readily (or cheaply) available at the moment.

I'll begin by saying that I've been trying to develop a back-testing system for home-grown strategies but as usual with such hobby-projects have never quite finished it off. The idea was that you could run a strategy with N sets of parameters and then compare their respective performance. I wanted to get this working to get confidence in a strategy before commiting real money.

 

[scpope]

gbr,

This is the sort of thread which will soon be full of alsorts of witty! remarks, so given that, I will attempt a serious answer first.

You need to actually trade for a while and think, what would make this easier/more convenient. From my own perspective, I have a simple daytrading volatility breakout system which seems to work reasonably well. The problem is that I can only daytrade on odd occassions as this is a hobby!/second business. I am trying to find the time to learn the extra vb/java required to automate my trading so I can daytrade every day. An app that enabled a trading system to be auto implemented would be useful and I don't know of one currently.

Unless you know different?

Stephen

 

[gbr128]

The plan always was to be able to code a strategy up in Java and then use that for testing, but also for the exact same code to be plugged into (for example) the IB API and make the actual trades. Personally I'd be too scared to let it run unwatched but it would be a nice thing to have.

Is the Vol-breakout you mentioned anything to do with the one described in the Bollinger on Bollinger Bands book?

 

[scpope]

Never read the book, its based on several ideas gleaned from bulletin boards and joined together.

 

[Rognvald]

SCP
It is my undestanding that unattended trading can be implemented through the current Tradestation package. It can also be implemented through a combination of IB (Brokers)/Investor Realtime/Ninjatrader.

IMHO thise facility while superficially attractive and tempting could be (and probably is) very risky for a number of reasons

 

[gbr128]

The Bolly one was dead simple, once the band-width reach a six-month low watch for the price moving outside the bands and then try to get in on the breakout. Was hoping to work out if this worked in practice by using objective testing on the computer. Was also interested in seeing if buying a straddle would be a safer way of making the bet (Bollinger refers to head-fakes, which appear to go North and then break out properly Southbound). Would need an option pricer to do that properly but haven't got 'round to working out how to use the one with the Hull book from Java, or writing my own.