where would you point a noob to automated trading that has never written any code before. Tradestation easy language??
Easiest way to start coding an automated system
- Thread starter Autospreader
- Start date
Maybe try Quantopian or QuantConnect and see if either of them seems natural to you. There will be a significant learning curve for any automated system, regardless of which language you write it in or which system you use.
I would agree Quantopian is very good. It has an excellent API to Interactive Brokers that works fine with paper trading account or your margin account. 100's of example trading systems and a university level collection of webinars on algo trading.
Quantopian is built around open source technology and python.
If you are new to trading you should check out the following classes:
Coursera.org "Computational Investing, Part 1" by Tucker Balch from Georgia Tech. This is Excel and Python based introduction.
https://www.coursera.org/learn/computational-investing/home/welcome
udacity.com: "Machine Learning for Trading" also by Tucker Balch
https://www.udacity.com/course/machine-learning-for-trading--ud501
coursera.org, edx.org and udacity.com all have other courses in python if you need more time getting up to speed.
All are free and very well done.
Good trading!!!
Quantopian is built around open source technology and python.
If you are new to trading you should check out the following classes:
Coursera.org "Computational Investing, Part 1" by Tucker Balch from Georgia Tech. This is Excel and Python based introduction.
https://www.coursera.org/learn/computational-investing/home/welcome
udacity.com: "Machine Learning for Trading" also by Tucker Balch
https://www.udacity.com/course/machine-learning-for-trading--ud501
coursera.org, edx.org and udacity.com all have other courses in python if you need more time getting up to speed.
All are free and very well done.
Good trading!!!
Thank you for your information,
I have been a discretionary hand trader for 3 years and I just want to automate some of my basic strategies to scan for opportunities and maybe eventually execute the trades with an algorithm
Not sure if this it what you're looking for, but this is what I use. Sort of semi-automatic.
https://code.google.com/p/yahoo-finance-managed/wiki/enumQuoteProperty
Though it says updated 2011, the codes still work today.
I run Applescript on Mac Numbers, similar to Excel, and with a command, I created, it updates the request automatically. My actual trades are all done on a discount broker site.
Python
Trading Technologies offers ADL, Algo Design Lab, which provides DIY algo development. No coding required. https://www.tradingtechnologies.com/customize/adl/
@patrickrooney - any ADL inputs for strategies that uses S/R (Floor Trader Pivots)?
Many of the standard tools use walkforward, it makes the coding quite complex depending on the scenario, there are specialist quant tools as people have mentioned. The main problem is "easy" and "automated" are not compatible with automation.
Automated trading has one problem, getting rid of the false positives and false negatives, when you do find a way you invariably get rid of the winning trades and include the losing trades. The key is to find the 'exact' balance to exclude the false positives and include the false negatives. Given the millions of combinations of signals it's not 'easy', and then when you do find it the strategy only lasts a set period of time, trading goes in waves and the length depends on the wave you catch.
Basically automated strategies are caught out by fat-tail events as you need perfection to balance winners vs false positives and losers vs false negatives, fat-tails will make sure that those using automated strategies are punished! This is because when people seek perfection they 'forget' the simple things outside of normal standard deviations, it makes the algorithms too complex.
It's not impossible, it's as difficult as discretionary trading but the effort is put in to perfection, the platform you use then becomes more personal choice and what fits your style and experience. The best is to get your feet wet with a standard platform and then increase your knowledge once you have created some test strategies watching how they work in realtime, for that you want to ideally use low timeframes, it increases the false positives/negatives but also reduces the testing time, backtesting is very hit and miss.
Automated trading has one problem, getting rid of the false positives and false negatives, when you do find a way you invariably get rid of the winning trades and include the losing trades. The key is to find the 'exact' balance to exclude the false positives and include the false negatives. Given the millions of combinations of signals it's not 'easy', and then when you do find it the strategy only lasts a set period of time, trading goes in waves and the length depends on the wave you catch.
Basically automated strategies are caught out by fat-tail events as you need perfection to balance winners vs false positives and losers vs false negatives, fat-tails will make sure that those using automated strategies are punished! This is because when people seek perfection they 'forget' the simple things outside of normal standard deviations, it makes the algorithms too complex.
It's not impossible, it's as difficult as discretionary trading but the effort is put in to perfection, the platform you use then becomes more personal choice and what fits your style and experience. The best is to get your feet wet with a standard platform and then increase your knowledge once you have created some test strategies watching how they work in realtime, for that you want to ideally use low timeframes, it increases the false positives/negatives but also reduces the testing time, backtesting is very hit and miss.