Creating Algo for MORE Intraday Trades

profitlocker said:
I recommend Multicharts with IB API. I think it's a great combo. The hardest part is getting it coded right
More...
I agree with this. Going with Excel and either the DDE or ActiveX API is risky..and unreliable.
Go with Multicharts and the IQFeed real-time data feed...another great combination.
 
Trader Nomad said:
Hello all,

First post. I've been a FT trader for 5+ years now but am completely new to the automated trading scene and have no experience with C++ or writing code for anything for that matter. I would like to develop an algo for use on IB's API.

This particular intraday strategy works well if I can apply it to ALL the stocks I have chosen to apply it to that day (typically 6-12 stocks) but I can usually only get entries and exits properly in 2-3 stocks doing so manually. If I can get proper entries and exits on every stock, the law of large numbers would overtake Murphy's Law (strategy has been back tested thoroughly).

I want to be able to manually input each stock, appropriate share sizes, and be able to cancel the orders if the bid/ask spread gets too wide or other things happen that make the trade risky or not worth it. The only variables or inputs are to do with price (no volume or any other indicators).

The strategy is fairly simple and I would think developing the algo wouldn't be too difficult but again I know almost nothing about coding. Should I go the route of learning the process myself or hire a dev to create the algo for me? Any advice/opinions would be appreciated.

Cheers,
TN
More...

Try freelancer.com (formerly rentacoder.com) to find a coder.
 
sprstpd said:
Just bite the bullet and learn some Python. That way you can customize the way you want, you will learn a useful programming language, and you will broaden your horizons.
More...
Python is not for starters.
 
Simples said:
Wondering how much discretion goes into your trading decision, are you mechanically following your own rules or are there room for interpretation beyond spread, volatility, etc? Point is a trading bot will take all the signals, even "stupid ones", thus applying murphys law to the nth power, not reduce risk. This unless your backtesting includes everything for execution, avoids hindsight biases and can be followed 100% mechanically. Making such systems takes alot of experience both with programming and trading the market. One will go through eureka moments thinking youve cracked it, then discover the parts that doesnt work. This can go on for years. Id recommend to start simple and build from there. Excel is as simple as you can get it, but still very flexible. When your needs exceed excel, maybe python is achievable after experience with excel. Best to do it yourself if you can, otherwise need a partner that wont abandon you.
More...

No room for interpretation beyond spread/volatility. An easy example: If a stock has a bid/ask spread of $0.30 and I'm looking to make $1.00 from the trade.. the risk/reward isn't worth it. A loss would be more than expected and a breakeven likely wouldn't be breakeven due to that spread. An example of this was EL (estee lauder) this morning. The bid/ask spread was like $0.50 on a $80 stock

The strategy is very mathimatical and binary, so I don't see how Murphy's law would apply so long as the algo follows the strategy correctly.

My backtesting accounts for commissions, poor/no execution, and slippage. I've experienced a lot of this and know any trading system has to account for these inefficiencies that happen in the real world.

Thanks for the advice. Followup question: How different is developing algos on excel vs. python? I would be concerned that any skills learned on Excel would not be transferable to Python. Also as others have said, reliability is an issue with Excel.
 
Question for those of you who are well-versed in Python: How long did it take you to get to a level competent enough to write an algo for your trading that works? Unfortunately, I took Finance in Uni and not programming.
 
Trader Nomad said:
Question for those of you who are well-versed in Python: How long did it take you to get to a level competent enough to write an algo for your trading that works? Unfortunately, I took Finance in Uni and not programming.
More...

Experienced programmers learning a new language is not the same as someone new to programming, learning a new language.

Based on some quick research I just did, I would recommend you learn C#.

EDIT:
Looks like you may have found a way for Python to work with IB. In that case, learning Python would be quicker than learning C#.
 
You must log in or register to reply here.
Back
Top