Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Politics : Formerly About Applied Materials -- Ignore unavailable to you. Want to Upgrade?

To: Robert O who wrote (50777)	8/27/2001 2:17:15 PM
From: mitch-c	Read Replies (1) \| Respond to of 70976

Re; - max-pain discussion, statistics - this gets deep! To allow your range (in order to offset 'noise' effects?) is too grand. I think Ben's idea here is novel, but allowing +1/-1 of closest strikes doesn't pass the Z-test. Umm. This bugged me for a week; I thought about it last weekend, and I think I know why. Stock prices are a continuous spectrum ... or at least a close discrete approximation of continuity, given that the steps are small compared to the value. Therefore, statistical tools (such as a z-test) can give accurate predictions (correlation, standard deviation, etc.). However, the steps in option strikes are at best 5% of the stock price, and at worst 10%. ($2.50 below $50, $5.00 above.) Worse, as the time value approaches zero, the distribution of each option price approaches its linear asymptotes, and is highly discontinuous (therefore non-integrable) at the "knuckle" exactly at the strike price. So, using statistical tools designed for a continuous (therefore integrable) distribution on a discrete, discontinuous distribution may generate precision, but not necessarily accuracy. I don't think the data can generate a prediction more granular than what it is derived from (example - coin flips). So squeezing a prediction of a closing price to the closest nickel from data that is granular at 50 times that is futile. However, when you look at the premise behind the max-pain procedure, I think you may have a point about how tight a predictor it is. Specifically, allowing the low range to go down to the next strike lower might be too generous; instead of a +1/-1 range, a +1/-0 or +0/-1 range may better reflect the spirit of the max-pain premise. I prefer, though, to use a looser range - because I am more confident that anything outside of it represents an opportunity. (The idea is analogous to the difference between a 95% confidence interval - 2 sigma - and a 99% confidence interval - 3 sigma.) Thanks for the challenge to think through what had been more of a call on instinct than analysis ... I'm actually happier with my range choice than I was. (Assuming, of course, that someone doesn't shoot down my vaguely remembered statistical analysis ... <g>) Comments and critique welcomed. - Mitch