Looking for someone who has a few minutes to point me in the direction I need to go. I have a few trading strategies i'm using that need optimization. One in particular can be used across the option market and stock market. My problem is gathering the data and compiling it and then optimizing it. It needs intraday data and not daily. I'd like to tailor it to the option market specifically. However, if I think I can get a reasonable sample set using particular stocks if needed. It's a very basic strategy with a very basic entry and exit. But with 2 to 3 signals per strike per symbol per expiration I'm looking at an enormous amount of data. On top of that I have no clue where to get the data unless I manually extract into excel by hand. I have access to almost all historical option charts via my software but im not sure if this would help. Simply put, I'd like to gather some statistics over an unbiased sample and optimize this strategy even further but I have zero how to compile the option data into a useable format that can automatically output a rough estimate on the stats after I input the additional trading variables I'd like to test. As I said, Excel is the only way I see this happening at the moment which is very time consuming. I can enter this by hand but my main concern is the amount of time it will take. Anyways I was wondering if anyone has been able to come up with a layman's solution to this problem. I'm not a quant and don't have enormous resources. Otherwise if anyone has come across an analysis software that can at least speed up this process that would help tremendously. I have about $5000/ year to pay for software. Thanks
Genetic optimization can be used to optimize a large set of parameters. It can easily overfit though.
If you can see it on a screen, you can block it and copy it into Excel. Years ago I did this with some phenomenally large blocks of data doing similar option analysis.
The specifics of your situation matter in how you solve your problem. But here's my solution based on my assumptions that I made up about your situation: I think you have no choice but to buy intraday data. I don't know if it will cost more than 5k/year. After that you have to do it the hardway in excel or in R or some other language (like Matlab). I personally use Excel even though it runs slow. The development time for me is much faster. Is it possible your signals per strike are correlated? For example, if you think a stock will go up, all your call strikes will show a buy signal.
That makes sense. For the most part yes they'll be correlated. Deep OTM and deep ITM no. But i'll filter those out.
Long ago I would grab data from Yahoo Finance. Just hold the mouse button down and scroll down to highlight a pretty clean block, then right click and paste into Excel. If you're pretty handy with Excel it's easy to clean up pretty quickly.
Man, that's a lifesaver. I really appreciate it! This may be a dumb question but did you just use line charts when importing or would it recognize other chart styles like candlestick.
I only grabbed data, never charts. Once I got everything formatted and lined up in Excel, I made my own charts there.