Ah see, that's an interesting way of doing it. I haven't seen a single app that handles specifically corporate actions like that so I'm not sure what intradaybill was going on about that this is very common. Thanks I'll look at it, even though I'm probably already decided on how to do my version. Oh I didn't know that was him. I read about that system. However I've never used it since I don't think NeuralNets are very practical for trading applications.
Out of curiosity, is there a good free (or cheap) source for corporate action data (splits/divs/takovers)?
http://getsplithistory.com/list?c=A This is for splits , but would require a lot of work to make usable.
I build a basket database once to walk a supposedly optimised ETF basket backwards. Lots of moving parts .. and .. turns out it's hard to track an index without all of the index.
I use only the historic price data that is more recent and greater than say, $ 1 / share. That still might provide about 25 years of price data.
+1. Exclude stuff under a certain price, but keep in mind that you might be excluding stocks that never split but were actually trading that low too.
Regarding the stock splits handling, please have a look at the following post: Real World Equity Trading Please let us have your opinion.