Yeah. It seems like if he didnt make so much money running a brokerage firm he'd be a step below Tim Sykes.
Good article. This is what a lot of people don't understand. Whenever you make comparisons in data you want to avoid two floating variables, in this case price and volatility. You want to make one a constant and analyze the second one. So a true comparison would evaluate implied vols at EXACTLY the same stock price. In a perfect world you would have stock XYZ that traded at a constant price and then you could perfectly price vol. The problem is when you go back and look at data on a stock and the stock has been all over the place. You can't possibly make an accurate comparison. For example, take a look at the implied vol on GOOG at $1000. What if GOOG's stock price dropped to $100 over the course of a year. What is the implied vol going to look like now. It won't be a 20 vol like it is currently, it could be a 50 or 60 vol stock now. Does that mean vol is expensive? Good lord no. In fact, it's probably cheap. You have to normalize price. And vice versa. I see guys make this mistake all the time on the upside. Take a stock that is $50 and over time works it's way up to $200. Well, implied vols are going to slowly drop in relation to price. If the stock had a 90 vol at 50 it might be trading at a 30 vol at 200. Well, a relative comparison is going to make the vol look awfully cheap when in reality, at a 30 vol, it might be insanely expensive. The second thing you have to normalize is the overall market vol which one can simply use the VIX for a dirty calculation. Obviously the market is trading different now then it was in 2008. But one has to make allowances to what the overall market is doing. In this tape, implied vol looks very cheap everywhere when in reality, it's probably over priced across the board. In 2008 implied vols probably looked very expensive when in reality, even stocks trading at 3 times normal implied vols where actually dirt cheap. Vol is very dynamic. There are so many things that need to be evaluated then simply "ranking" it over a 52 week period without regard to price and environment.
I would not advise buying/selling vol because it's in it's n-th percentile. However, money in options is made by betting that market prices/vols are incorrect. The trick is to find the little inefficiencies that exist out there. Everyone is aware of the level of IV/HV, you need to find reason why everyone else is wrong. As an aside, I see that guy in my neighborhood a few times a week. He needs a haircut.
Mav, I find it hard to digest that vol isn't already normalized to the price. With the range of values it uses, 0 to 100, it already shows normalization. So, there's a normalization with a time dimension? And, what for does this take? A third order greek? FYI I'm still trying to work up the math to understand Sinclair and others. If it is so relevant, why do data vendors /analytic services not offer this? Your post will be a notepad on my desktop for quite sometime.
The issue is not whether vol is normalized to price but whether the trader doing the analysis understands that the two vols are different. Let's use a subscript to describe vol in TSLA. We have vol(40) which is TSLA vol at $40 a share and vol(200) which is TSLA vol at $200 a share. Yes of course the market is pricing vol accordingly but are you? Vol(40) and vol(200) are not the same thing. Therefore using a "ranking" structure to compare them is absolutely pointless. And no, software vendors are absolutely clueless in part. On the flip side vendors market stuff they can sell and simplicity sells, complexity doesn't.
the worthlessness of IV ranking doesn't apply to vol cones correct? if implied vol of a one month option is 50, and i see that 50 is the 90th percentile for one-month realized vol over the last two years, that is worth something, right?
The total price number shouldn't matter for vol, should it? If GOOG did a 1:10 reverse split then vol would still be 20 the next day after the split.