Got any ideas for a comp sci / market project?

[mix_master_mike]

Unfortunately, I cannot get my stats requirement out of the way in the winter '06 semester here at UofM. Rather than take 13 credits, I've opted to look into doing a 1 or 2 credit Directed Study with a Computer Science prof.

I've had a little time to figure out what I'd like to do related to the markets and programming, but there's so many things out there that I'm having a tough time narrowing it down. Of course you've got the basic things that have already been done - simple backtesting of well known technical strategies, what have you. This is definetely a possibility, but its all already been done and such programs are pretty accessible.

I was also thinking of perhaps digging up historical options data and looking at volatility and movements and how it relates to the future price of a security. This could be pretty neat. Also perhaps develop some new ways of charting options prices.

Another project that I could do with very easily accessible data would be to look at INET data for something like the QQQQ to produce realtime graphs and look for patterns of not only the price/volume but also the bid/ask prices and sizes. Find some patterns, test them.

If anyone has other cool ideas they would be greatly appreciated!

[calven]

Don't know if this qualifies...but here is what I would be interested in....

1) How reliable "chart patterns" are... IE - Cup & Handle, ascending triangle, horizontal channel, bump and run.

2) How other world markets affect the US markets....IE if the chinese market is down does this affect the US market the next day. What is the correlation?

3) Contrarian indicators... Gold rising markets rising / oil prices rising how do these commodities affrect stock prices...to what degree?

4) Affects of differant analyst sentiment on stock movement...

5) Relationship of short interest on stock price.... high short inerest = stock moving lower?
How much so?

These are my ideas....

-cal

[agatto2]

Calvin them are some great ideas , i have these i would like to throw in

1 - stock jumps price and volume above 100% , then settles for next month or so down to original price, then eventually they pop and run for a good profit. I would like to know the average time interval(when the stock starts to rise again) , and the eventuall average profit. Ive been trying to write my own program for this but am having a hard time finding stock scaning tools that will let me do this.

2 - Bankrupt companies. They drop like a rock, do a dead cat bounce, ( usually for a good profit) ex; look at enron down to micro pennies, then bounced to over 10 cents a share, is there a pattern as far as percentage drop , percentage bounce.

tony g

[dyc100]

Hey, awsome, I am also a comp sci student and was also thinking of doing something of this nature, we should hit it up,lol!