Search results

What's with all the nastiness in this thread? You all are gonna drive off yet another person who actually makes their living trading.

I can't say if the TWS scanner has made me money or not. What I can say though is it's stopped me from entering trades where the risk-adjusted upside wasn't nearly what I thought it was due to incorrect or mistaken assumptions on my part. That was very valuable for me.

It's not magic, if that's what you're implying. It just optimizes in the sense that if there's a strategy with break-even at points (x, y) and one with break-even at points (x-a, y+b), it's going to show you the latter (simplifying a lot here by ignoring variables like downside risk etc., but...

Interactive Brokers TWS has a solid strategy scanning tool in the Strategy Builder. It's pretty straightforward: you put in your forecast and it optimizes available strategies for you given the current pricing data. It's great, especially for a free tool. It feels pretty limited though -- it is...

http://ocw.mit.edu/courses/electrical-engineering-and-computer-science/6-003-signals-and-systems-fall-2011/ You have to learn to crawl before you can run.

What you folks are describing with the "curve fitting" is exactly why you have a separate training set and test set. It is not statistically correct to validate your learned parameters (the aforementioned "curve") against the same set you learned them from.

Sorry to disappoint, but I've never heard of the book.

Err... "Cool in a crisis? UW Emergency Management seeks volunteer responders"? Not sure what that has to do with anything. I'm just talking about basic statistics concepts that you'd learn in the first week of any college-level machine learning class.

How did you create your model? I think a lot of people do backtesting wrong. In machine learning there's a concept of a training set and a test set -- you'd *never* take the results of the model on the training set as any indication of the model's strength. Instead you have to see how well it...

Think of it this way. If it was a regularly profitable strategy, why would anyone ever sell options? It would always be a losing proposition. Hence the price of options increase until that is no longer the case. As it is right now, it'll be generally a losing strategy unless there's a black...

Err, no they wouldn't. By definition they make money on the trade, just less money.

I think we'd need an actual lawyer to answer this question. Because you're totally right -- from the broker's standpoint, if it's not illegal, this is a completely obvious move. I can only think of one reason not do it if it's not illegal and that's training your clients to find the break-even...

We just spent two pages teasing the exact description out of you, and I'm still not sure you exactly have it. If you want the code, you are welcome to pay my consulting rates.

Forget about the *how* to do it. Just focus on actually defining the problem completely. Then once you have it defined, implementation is straightforward. This is literally just picking data out of a database and doing some division.

This is still not a well-defined problem. You specified that a 10%+ drop from a given date represented such a gap to be recorded, but what if the previous date was higher? That would also be a 10% drop and would result in a gap of the same number of days + 1. Basically, you need to define how a...

Hi syswizard, the problem you posed is still ill-defined. Does the 2000 represent a peak (ie the prior day is < 2000), or is it entirely arbitrary (ie the prior day could be > 2000)?

Assuming you mean figure 3, that depends on how exactly you define "10% correction". How would you define it?

This might be an unpopular opinion but stop being a leech. Rather than pay for a proxy service, how about just paying for the data service? The limits are not there arbitrarily, it actually costs them money to provide a free service.

Is there a data feed that has these particularly well? Right now I'm just using the feed that comes with IB and they don't even go back very far for *unexpired* options! I don't need to go back beyond a few years or so. I've looked at DTN IQFeed in the past. Do they have a good offering for...

@DarthSidious I think you may have confused me with someone else. I'm already using the trial, and I was looking for someone else's experiences to help me understand the shortcomings of the backtesting system. The documentation is pretty solid for most of the program I agree, but for the...