Know how to program, don't know how to trade. Where do I begin?
- Thread starter ET873
- Start date
http://wl4.wealth-lab.com/cgi-bin/WealthLab.DLL/browsechartscripts
You can start there. There's only 1897 chartscripts, or you could save time and just read the top 25.
You can start there. There's only 1897 chartscripts, or you could save time and just read the top 25.
Totally not true.
I know how to trade but not program... Where do I start?
OP, what do you write/what languages do you know? What is your startup budget? I'm learning C++ and trying to get (MUCH) better at VBA. It's no joke and a serious investment - and I know how to (lose money) trading.
It'll cost about $20k-$30k startup for data feeds & equipment so I'd suggest that you start with plain old trading and get a sense of the markets. Once you have a successful strategy then the programming part will be easy.
Sort the NYSE by daily volume...
Look at the Bottom Third...
And find a niche to specialize in.
When you can make money manually... automate and scale up.
The absolute last thing you wanna do...
Is compete in latency wars with the Big Players...
That means avoid anything very liquid and sexy...
And about 95% of the posts here are total BS by kids with 10K.
One of my skill is the programming (PHP, VB, JAVA) also.
1) I think you should get a good trading/testing SW (amibroker, metatrader, etc.).
2) Learn some TA (chart and bar patterns, indicators, support-resistance, moneymanagement, trailing stop, using different time frames).
3) Practice to use the SW API, and functions. Write your own custom function library. Maybe you have to code some pattern recognition, support resistance level calculations, trailing exits, etc.
4)Bact test your programmed ideas (system) for buying and selling on historical data and made conclusion about statistical data (good system or wrong system). Systems should treated separatly for every stock / commody. No general system can exist.
1) I think you should get a good trading/testing SW (amibroker, metatrader, etc.).
2) Learn some TA (chart and bar patterns, indicators, support-resistance, moneymanagement, trailing stop, using different time frames).
3) Practice to use the SW API, and functions. Write your own custom function library. Maybe you have to code some pattern recognition, support resistance level calculations, trailing exits, etc.
4)Bact test your programmed ideas (system) for buying and selling on historical data and made conclusion about statistical data (good system or wrong system). Systems should treated separatly for every stock / commody. No general system can exist.
avoid system pitfalls:
-curve fitting,
-lags on real trades,
I think good systems considers:
-trending or ranging the entire market
-S/R levels
-recent volatility
-managing risk/reward ratio below 1.
-impact of volume
-confirming the system is profitable on min. 200 out- of sapmles data.
-curve fitting,
-lags on real trades,
I think good systems considers:
-trending or ranging the entire market
-S/R levels
-recent volatility
-managing risk/reward ratio below 1.
-impact of volume
-confirming the system is profitable on min. 200 out- of sapmles data.
Go over to quant land and you'll see that programming skills aren't that important; it is math skills. You can't fake math.
Programming is just a specialized subfield of math. (Or is math a specialized subfield of programming..?)
Programming skills are definitely very important. There are many algorithms where the naive implementation runs in days or weeks or longer, and the good/non-obvious implementation does the same thing in a few seconds..