Programmer looking for some help

[Optitrader]

Hello,

Some background information before I post my main questions...

I have been reading around the subject of trading for a few weeks now. I am a .NET developer by trade and (like many before me) I fancied the idea of building a virtual trading platform to back-test systems based on genetic algorithms. After an initial splurge of activity, I successfully created a virtual stock market, and a virtual broker. I created the generic algorithms and back-tested some simple strategies using 10 minute tick data from livecharts.co.uk. After a few moderately successful simulations, I noticed some strange results and I traced these back to the data which I have since concluded is totally useless. The data does not match the data from other sites like Yahoo and contains other junk data. After numerous web searches and searches on t2w, I have been unable to locate a cheap source of data for UK stocks. I am based in the UK and I would prefer to trade companies I am familiar with. Having a very small capital to invest, I am reluctant to spend hundreds of pounds just to get a few years of old data, this seems excessive.

So, my questions are:

1) Does anyone know a *very cheap way of acquiring historic data for FTSE 100 stocks ideally in CSV and ideally 2-3 years worth.

2) Would any fellow developers (or TA traders) be interested in some sort of joint venture, sharing resources, strategies, ideas etc?

 

[Optitrader]

Update: I have found a supplier of cheap FTSE100 stock data, http://www.regisdata.co.uk/.
It works out at 50p per stock, per month for 10minute data which is reasonable.

 

[Optitrader]

Update2: The site I posted no longer deals with FTSE Stock data, so I am back to searching the web again.

 

[kiwiBandit]

2) Would any fellow developers (or TA traders) be interested in some sort of joint venture, sharing resources, strategies, ideas etc?

Hi Optitrader,

I started out with a similar idea 2 years ago and am trading live daily with a .NET automated platform + 4-5 algos now. I would be interested in joining something if you get it together. People are fairly protective of IP though so perhaps start off as just talking. I'd be interest to see how many .Net independents there are in the same boat + success rates.

For daily[edit] data I use Yahoo and made a bot to download years worth when required.

Can check these guys: http://www.dukascopy.com/freeApplets/exp/

JB

 

[Optitrader]

Hi Optitrader,

I started out with a similar idea 2 years ago and am trading live daily with a .NET automated platform + 4-5 algos now. I would be interested in joining something if you get it together. People are fairly protective of IP though so perhaps start off as just talking. I'd be interest to see how many .Net independents there are in the same boat + success rates.

For daily[edit] data I use Yahoo and made a bot to download years worth when required.

Can check these guys: http://www.dukascopy.com/freeApplets/exp/

JB

Hi JB,

It is nice to hear I am not the only one going down the .NET route. I would be interested to hear more about your project. I like the idea of a Yahoo bot, but I understand the Yahoo figures are only available at 15 minute intervals and with a 15 minute delay. Is that correct?

I was aware of the dukascopy website, but when I looked at their intra-day figures compared to the Yahoo finance EOD figures I found many discrepancies. have you come across any issues with them?

I can understand people not wanting to give away their edge, but I think it would be good to share some other resources like libraries, data, links, basic strategies etc. For example I have found an excellent library for indicators if your interested, it is available here: http://ta-lib.org/hdr_dev.html.

Cheers
Optitrader

 

[kiwiBandit]

Have you checked the differences aren't just stock splits?

 

[Optitrader]

You could be right, I don't know enough to be sure. In some cases the data was repeated and had gaps. Also the volume information is limited...example below:

LLoyds.L

26/11/2009 15:40:01 514 90.42 90.46 90.06 90.84
26/11/2009 15:50:01 605 90.46 90.12 89.78 90.65
27/11/2009 08:00:01 26 90.12 88.81 88.65 90.12
27/11/2009 08:10:01 1 88.81 88.82 88.81 88.82
27/11/2009 08:20:01 5 88.82 88.81 88.81 88.82
27/11/2009 08:30:01 4 88.81 88.81 88.81 88.83
...
16/11/2010 07:30:01 33 87.86 87.84 87.84 87.86

Maybe I just hit my download limit.

 

[kiwiBandit]

- Sometimes they divide volume by 10 or 100 to save on zeros.
- The volume after hours is generally small.

Check on Google for splits, can find dates where they've had splits and normalise the data if you need to. Or just run models using the new share prices & incorporate splits into your model since that is the real scenario.