31–40 of 120 posts
So... anyone want to share some of these algorithms for predicting stock prices?
That's been going on since the late 1970s, when personal computers first became available. It worked back then, because few were doing algorithmic trading at the time. Now, everybody is. There's new interest in systems which look at external data sources such as news items, rather than just at market data.[1] They have a chance of picking up something before the market reacts. But that's already widely deployed; abou…
Your example of analysing new items is a good one. If you can spot a better way to predict stock or currency movements by a cunning new way of processing news feeds, then you could potentially profit from it without having to rely on speed of execution.
People have been using computers to do stock analysis, as you say. But this doesn't make it a solved problem. There's money to be made if you can do it better than the others. Of course, the competition is very tough.
Interesting article but for a different take on a statistical approach to the market, curious if any peeps on HN are into volatility trading? From what I understand a lot of the "DIY vendors" cater to the equity crowd, meaning people who build their models on technical indicators (MACD, RSI, advancers/decliners ratio, Fibonacci golden ratio retracement, MA); you build your model of some combined signals, back-test it…
Any pointers gratefully rcvd
Are there any free or cheap sources for historical stock data in a computer-friendly format?
Having said that, it's hard. Most people simply don't have the bandwidth to do it properly. Parsing daily Yahoo prices and having a R script or two somewhere will absolutely not put you in the "I do successful algo trading" camp. Or running TradeStation with a handful of pair strategies or whatever.
So I differentiate between them and people who write more serious automated algos. I'm sure in the first camp there are some that make profits, but as someone said here already, It's usually a zero sum game.
Who are the more serious DIY traders? Hard to define but if they do most of those things I'd say they are in the small camp of home traders that often make money.
- Willing to fork up the money for good historical intraday data (10 years of US intraday data from Nanex cost about USD 30K). I'm super serious about this, if people even blink an eye here, I'd walk away. They don't understand how important data is.
- Build a proper security master database, or buy a security master feed. This is a project of blood sweat and tears. Every day securities need to be resolved. Again if building this out of, e.g., Bloomberg data, quite some money has to be invested.
- Have a sophisticated back-test system where they can simulate models against MANY traded stocks concurrently. Meaning, they can back-test making decision tick by tick, or second by second, against a large universe of instruments, for example 3000 US stocks. This can be slightly eased by doing some daily preselection, but usually good strategies still scan hundreds of stocks real-time.
- Have the ability to execute this stuff live. IMHO, the article is right to mention Interactive Brokers for this, they are one of the best in this kind of setup.
(Only now are we where most people try to start) - Have the stats, math and dev chops to find and implement a profitable model/strategy. I have to say though, while still difficult, there are many untapped possibilities to make some money. But once decided on a general direction the "time to market" is still often at least half a year. Many give up before that...
There are some more things I'd look for but this is the main list I'd say.
Here's the problem with trying to create your own trading system. How do you back test it to know that it works. If you back test over the past 5 years then you are only testing your model against a huge bull market. If you back test over the past 20 years then I'm not sure it helps much as the market of 20 years ago didn't really have any of the major market drives of today's markets, HFT's, huge numbers of hedge fu…
Here's the problem with trying to create your own trading system. How do you back test it to know that it works. If you back test over the past 5 years then you are only testing your model against a huge bull market. If you back test over the past 20 years then I'm not sure it helps much as the market of 20 years ago didn't really have any of the major market drives of today's markets, HFT's, huge numbers of hedge fu…
well, giving the fractal nature of price movement, i'd guess one can backtest against last several days/weeks taken at very high resolution and scaled out to represent months/years :)