Step 1) Define your environment (Limitations)

a) Amount of time you expect to spend

b) Amount of money you have to loose before you give up (yes, and exit strategy is advisable)

c) High Frequency / Low Frequency trading

d) Commission rates available to you

Step 2) Define your expectation (Goals)

a) Number of trades per unit time

b) expected commissions

c) acceptable rate of return after all expenses

Step 3) Start searching for methods that satisfy Steps 1&2

@Praveen. Algo trading is nothing that should be attempted in a half baked fashion. Whatever you wind up doing make absolutely sure to test it extensively with paper trading before you start using the algo in real trade. You can lose money really fast in algo trading if you do not know what you are doing.

Maybe some clarification is needed. Maybe I am wrong but to me Algo trading is just a mechanisation of manual trading methods. Backtesting and running forward tests on a demo account is mandatory exactly in the same way as for manual trading. Do you have an existing manual method you want to mechanize? The exception is High Frequency Trading where timing is critical. Hard to test manually. So. What is your starting point?? What is your experience in trading? What kind of market do you have in mind?

That is exactly the point Ingvar, that is exactly the point. You mechanize a trading system, as simple or as complex as you wish, test it and let it rip. You better make sure nothing will go wrong otherwise the outcome may be bad. The level of losses will depend on the trading system you mechanize and how frequently you trade. It is clear as daylight, or is it?

Not quite, but certainly correct in most cases. Automation has a distinct advantage when it comes to a trader that consistently brakes his own trading rules. Automation will give the consistency that may be lacking in a manual system. It is most important to test every system for consistency, for without consistency, profitability may only be temporary.

There are other benefits of "mechanizing" a trading method.

- It is easy to backtest (On patforms that support it)

- Automate Trailing stops

- Trading many pairs simultaneously

- Do automatic add on trades after a previous trade has reached breakeven

To name a few

Algorithmic trading has at least 3 most common definitions, and the synonym to "automated" or "systematic" is perhaps the least used meaning of "algo trading". I've recently held a webinar where I outlined them in as basic terms as possible: https://youtu.be/1XbTiM8F6BE. Hope it adds some clarity to the discussion.

hey guyz thx a lot . will explore this & ll get back to u guyz. .

It doesn't have sentiments, gutfeelings. It automates your market sentiments and heavy maths and statistical analysis, calculations and runs with market speed.

I use IB (interactive Brokers) and their DDE/API with Excel. Very easy to set up and get going.

How I love both questions like OP and answers like that of Michael: both maintain an illusion that success in algo trading depends only on a platform.

hey guys , i planned to get into AFL code of amibroker to do my stimulation, anyone hav any materials on these? thx

I use amibroker/AFL to test my day trading system. But I built my own trading app using C# and IB API. Every few days I loaded amibroker with new minute data and re-test the day trading system and compare actual IB trading results. This is to validate the trading model is still correct and look for optimization ideas.

@Lawrence.

Proves the point made by several here. It is the model (or method) that is important, not the computer.

I have been algo trader for years and I mainly use TradeStation and MultiCharts. Both have easyLanguage programming language, which is easy to use. You can also use NinjaTrader, but the language is more difficult to learn.

ok any material for Afl pls?

I dont think there is a platform that may be better than others. If you have programming experience in a certain lanuage you should search out a platform that is developed in that lanuage. I prefer C# so i currently use NinjaTrader. I would go with the platform you are most familar with so you understand that platforms order execution features and methods. The rest you can learn.

Amazing "advices". None of them suggests any (ANY!) hint on why their platform of choice is superior, or at least what the reasons they're using it are. However looks like the OP doesn't care of it either. What a professional discussion.

Ok guyzs i think its better to recruit the experts and get the job dne

Interactive brokers provides you with an API for automated trading. If your algorithms do not demand ultra low latency (HFT algos do), this would be sufficient.