[spitfire]

There's a lot of different "versions" of HFT. HFT right now means millisecond response times. While UHFT means microsecond - collocated at the change. 1us=300m fibre distance.

The fact is, you probably don't have the knowledge to pull off UHFT, or HFT. What you're looking for is algorithmic trading, which is where smart, talented people can be very successful.

marketcetera has an open source platform for this sort of thing. You can write your algorithms in Java or Ruby.

To do HFT properly you need large amounts of capital. You'll be shaving fractions of pennies from trades, but doing a lot of them. For your typical trader with $100K or so, this is a losing proposition.

So, as stuck says your best bet is to sell shovels.

Best of luck.

[fuzzypickle]

I suppose that could be where I end up down the road, but right now I am more interested in ( as you correctly concluded that I am without the large amounts of necessary capital for profiting from fractions of pennies) the daily/weekly volatility that is oftentimes seemingly unnecessary and counter to the longer term trends.

[spitfire]

Algorithmic trading.

Hint: backtest on decades of data, not just the friendly waters we have now. Forward test your work for a few months before going live. Use the same amount of capital as you intend to trade with for all your tests.

Also, model with fat tails and decide if you're okay with those risks.

[fuzzypickle]

Testing my theory is the next step, as no theory is perfect and will certainly change over time. Theories come and go, and may even be true for a time, but I need the ability to effectively test/execute whatever theory that may be.

What is your experience in trading, if you don't mind me asking?