Quant Finance Question

Does anyone have any knowledge or experience in using “Reversal Factors”? If so, please describe their use, and calculation. I’m only finding vague references and loose definitions on the google. Thanks!!

Check out reversion

Yeah, I figured it related to the reversion of prices over time. A strategy that makes money by trading on the over reactions to events and other information. However, I’m not exactly clear how one would quantify this to create some sort of decision rule. Would it be based on a variance of price from the estimated fundamental value of the firm? Or what?

CF_AHHHHHHHHH Wrote: ------------------------------------------------------- > However, I’m not exactly clear how one would > quantify this to create some sort of decision > rule. Would it be based on a variance of price > from the estimated fundamental value of the firm? > Or what? You can get as complicated as you want. One simple reversion approach would be to buy something that was down yesterday and sell something that was up.

I totally agree with you and I get that. But what I’m trying to find out is: are there any specific techniques/ideas for determining when something was “down” enough yesterday to compel you to believe it would go “up” today. Because for you to say something is down, you would have to be comparing it relative to some base value. So what is the base value? And once you know that, then the question becomes: is there a way to easily quantify this for a number of securities on a daily basis? Clearly this idea isn’t revolutionary and its been done before, I’m just looking for a little guidance or a few examples if they’re out there.

CF_AHHHHHHHHH Wrote: ------------------------------------------------------- > But what I’m trying to find out is: are there any > specific techniques/ideas for determining when > something was “down” enough yesterday to compel > you to believe it would go “up” today. The quant funds would have you believe that those techniques exist. > Because for you to say something is down, you > would have to be comparing it relative to some > base value. So what is the base value? And once > you know that, then the question becomes: is there > a way to easily quantify this for a number of > securities on a daily basis? Down is simple. Just look at yesterday’s close and two days ago’s close. But what you want to know is – will that “down” revert? This is a different story and depends entirely on what data you’re looking at. Some time series are inherently mean reverting, and there are thousands of quantitative researchers out there that do this for a living. > Clearly this idea isn’t revolutionary and its been > done before, I’m just looking for a little > guidance or a few examples if they’re out there. Yep, ideas like these are covered under the umbrella of statistical arbitrage. Google and Wikipedia are a good place to start. For some specific ideas, you can look at French’s page: http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html

Thanks man I appreciate the input!

Glad to help! I would appreciate hearing from other quant finance people on this forum, if they exist. :slight_smile:

I’d post on Wilmott if you want quant people’s input

.

.

1morelevel Wrote: ------------------------------------------------------- > I’d post on Wilmott if you want quant people’s > input why? paul used to post here often, urging people to sign up for the cqf

Wow, i just looked up how much the CQF costs. I mean I’m sure its worth it but… Jeebus. http://www.cqf.com/apply/cqf-costs