Introduction

[cb79]

Hi just joined the site and thought I'd introduce myself. My name's Chris, from Palmerston North. My background is in IT but I'm studying a GDip Bus Studs with a focus on quantitative finance and econometrics. I'm interested in automated trading and machine learning. I am in the process of designing an experimental ATS. I will post the documentation up here in case anyone's keen to have a look.

[stoploss]

Welcome, all the best

[Snow Leopard]

Are you just going to design an ATS or do you intend to implement it as well?

I do a fair amount of trading/investing related software when not distracted by the demands of others.

I like to have the final say on anything my software recommends.


https://www.youtube.com/watch?v=PcvT6Zqa2Ps

Best Wishes
Paper Tiger

[mccollr]

Keen to follow your progress. Been an interested observer of such things for many years. Seems to be options for the US market, but not so our local market. Good luck.

[cb79]

Welcome, all the best

Thanks!


Are you just going to design an ATS or do you intend to implement it as well?

I do a fair amount of trading/investing related software when not distracted by the demands of others.

I like to have the final say on anything my software recommends.

Best Wishes
Paper Tiger

Yes I definitely intend to implement one when I've got a viable design. I am working with Python and trading via Interactive Broker's API. I am also getting historic/fundamental data from there.

Do you mean in terms of making sure the software hasn't made a horrible mistake? I'm not sure how I'll handle that yet, but I want to avoid emotional decisions if I can. I don't want to interfere with the system but don't want a Knight Capital scenario happening either!


Keen to follow your progress. Been an interested observer of such things for many years. Seems to be options for the US market, but not so our local market. Good luck.

Thanks, yes NZX is not big enough but I'm targeting ASX instead. US markets would be contenders but there's more red tape and inconvenient timezones so I thought I'd stick with Oz for starters. Only problem is of course whatever conclusions I can draw about stock prices are useless without forex predictions too, a whole new layer of complexity.

[Snow Leopard]

...
Yes I definitely intend to implement one when I've got a viable design. I am working with Python and trading via Interactive Broker's API. I am also getting historic/fundamental data from there.

Do you mean in terms of making sure the software hasn't made a horrible mistake? I'm not sure how I'll handle that yet, but I want to avoid emotional decisions if I can. I don't want to interfere with the system but don't want a Knight Capital scenario happening either!

Thanks, yes NZX is not big enough but I'm targeting ASX instead. US markets would be contenders but there's more red tape and inconvenient timezones so I thought I'd stick with Oz for starters. Only problem is of course whatever conclusions I can draw about stock prices are useless without forex predictions too, a whole new layer of complexity.

I will not start a religious war over your choice of programming language, we tend to be wary of python [click me :) (http://www.nst.com.my/news/2016/11/192884/100-kg-python-captured-baling-after-feeding-villagers-goat)] round here.

Rule number one for historical price and volume data, understand what your provider provides - how do the handle dividends, splits, re-namings etc, etc & etc.
I grab raw data, have any anomalies flagged to check manually, keep it, and make all compensations in my code or data.

Rule number one for fundamental data - don't believe anything!

You will want to extensively test anything you create before you let it loose with your wallet.

Best Wishes
Paper Tiger

[Toulouse - Luzern]

I am interested in this subject too.

[cb79]

I will not start a religious war over your choice of programming language, we tend to be wary of python [click me :) (http://www.nst.com.my/news/2016/11/192884/100-kg-python-captured-baling-after-feeding-villagers-goat)] round here.

Rule number one for historical price and volume data, understand what your provider provides - how do the handle dividends, splits, re-namings etc, etc & etc.
I grab raw data, have any anomalies flagged to check manually, keep it, and make all compensations in my code or data.

Rule number one for fundamental data - don't believe anything!

You will want to extensively test anything you create before you let it loose with your wallet.

Best Wishes
Paper Tiger

No offence taken, I am just as opinionated as the next developer... What are your issues with Python and what is your weapon of choice just out of interest? I am open-minded and flexible although I do have a linux/open source bias. My proper job is Business Intelligence, all Microsoft.

Is there a lot of discussion about this here then? I didn't notice much when I was scanning through. That would be nice.

Yes, my major concern with the data is the availability of failed companies, I think the provider I'm using doesn't give this.

It's more academic at this stage, just putting into practice some stuff I'm learning in my studies, yes it would be thoroughly tested if I ever got it tuned well enough to use for real.

[Snow Leopard]

If you want a criticism of Python, the programming language, then it is the reliance on indentation to define the scope of conditional statements, etc.
But I have never really used it for anything non-trivial and so have no basis to make an objective judgement.

These days, depending upon the environment (and the paycheck ;)) then I prefer C# if it can do it and C++ when it can't. I do a fair amount of php for websites & a lot of domain specific stuff.

There is very little discussion about 'this' here - it may be you and me - and I am off on my travels in a week. :)

Best Wishes
Paper Tiger

[cb79]

That is a bit unusual to begin with, but on the other hand you're never scratching your head over a forgotten semi-colon. It can be a bit clunky e.g. with nested iterators but then again there are ways round it like breaking out into a separate method. It is quick to use and there are existing packages for just about everything. I've used C# a fair bit at work and prefer Python for its speed and flexibility although I am biased as my Python coding is much better than C#. I have done a fair bit of PHP in the past which is OK for website stuff, although I would be keen for an excuse to play with Django. I never got into C++, I'm really keen to but am put off by the perceived steep learning curve. Plan is to eventually code resource-hungry algorithms in C++ instead of Python. End of the day it's just picking the right tool for the job I guess.

Is there a board that's more appropriate to talk about 'this type of thing'?

Safe travels.

[Snow Leopard]

...End of the day it's just picking the right tool for the job I guess...

You might want to look into the R programming environment [https://www.r-project.org/].


...Is there a board that's more appropriate to talk about 'this type of thing'?...

I make no recommendations, except stick with ShareTrader.

Best you go a googling on what you actually want to achieve, get stuck on, etc and follow the leads. See if anything suits you.


Safe travels.

Thanks for the reminder to update my location with https://www.safetravel.govt.nz/

Best Wishes
Paper Tiger

[cb79]

You might want to look into the R programming environment [https://www.r-project.org/].

Yes I went through a lengthy comparison of R and Python a while back, in the end opting for Python. People I know who use R speak highly of it.

[Meextr]

Hi I will also introduce myself, I'm from Rotorua and have been interested in shares for a long time but have only now just started buying again. I last bought in the early '80's in the UK during the government privitisations plus a few other shares. Got out of it when my finances were utilised bringing up children. Now an empty nester with a bit spare cash looking to invest again. Have gone for a few safe companies, but look forward to following this site for inspiration. Also hope to give some inspiration at some stage.