Ultimate Day Trading Platform/Information Collector | Hacker News Reader

Ultimate Day Trading Platform/Information Collector

3 pointsfuzzypickle7 comments

Just a quick little concept I've been wrestling with for a few months for use in day trading. As you may be aware, lost seconds can sometimes mean lost opportunity... time is absolutely critical when it comes to day trading.

Many instruction guides/books/blogs I've come across says that if you are to day trade properly then you need to go to a firm to day trade, actually go to work at an office. Reason being there are proprietary software and subscriptions that these companies subscribe to that give the trader good resources. However, this can be a highly flawed mentality because of its closed mindedness. To think that a single proprietary software subscription with its highly technical trend analysis could be the one thing that gives an edge over everyone else is, to put it lightly, is at the very least foolish.

Anyways what I am wanting to do is to do is construct/piece together my own trading framework, not necessarily from scratch, but perhaps a patchwork of sorts.

1. RSS Feed: One that allows me to pull the headlines and the important information, updated constantly. 2. Trading Platform(I am envisioning a Hotkey Enabled, or Macro enabled - could even be done through Windows Speech Recognition for some things): I have come across an Open source trading platform called Marketcetera. 3. Audio/Video feeds 4. Pertinent legislation and their progress 5. Commodity related information (weather for crop, production for oil)

This list can go on, I imagine, for quite a while.

Reason I am bringing this up is I feel like there is a way to bring the necessary resources together in an effective way, and in a way that is more flexible than the premium subscriptions already out there.

I am eager to discuss this with those of you that this interests.

Disclaimer: I am aware of the inherent risks of day trading/taking short positions, this is something I feel I am moderately disciplined in and of course am hoping to not lose everything I have in the process.

You should look into the Bloomberg Terminal and try and build something similar to it. http://www.entrepreneur.com/article/204898 http://www.bloomberg.com/professional/

Keep in mind that the bloomberg terminal costs like $1500 a month. So if you do manage to build something even remotely familiar you should stop trading and sell the machine.

A lot of great resources here, thanks so much.

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.

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.

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.

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?

fuzzypickle: The most successful companies in the trading business are the one's who don't trade! For example, thestreet.com makes most money with their recommendation systems (doubt if they even make a penny with their own trades). So if you are thinking of aggregating things and providing a service, it is a good idea - but if you are thinking of using it to your own algorithmic trading, it may be an overkill. Remember the trading world has some of the smartest people (doing machine learning, statistics etc) and you are competing with them in the zero-sum-game.

Also as full disclosure, I have setup a fully automated algorithmic trading (using IQFeed for realtime feed + IBAPI), but it wasn't HFT - but medium frequency intraday. I shut it down because it wasn't profitable. But if you are thinking algorithmic trading, you need to get into machine learning and statistics space (most of the info is just noise in this space). I have myself thought about the aggregation as a service, but never got around doing anything in that direction and would love to chat if interested.

Reply on news.ycombinator.com