The author is completely full of shit when he makes outlandish comments like you don't need to understand math which he willfully admitted he didn't. Stat arb requires a tremendous understanding of math to model the risk. Stat correlation as some pointed is pairs trading but that is not saying anything. Most math guys find themselves in the world of stat arb because it's far easier to model "relationships" between two products vs an outright product. A correlation trade is synthetically the same as an option and they need to be modeled that way. If you are trading convergence you are synthetically selling naked puts on the spread between the two stocks. We all know how that works. You will earn consistent small premiums most of the time then get killed every now and then. If you bet on "divergence" you are "paying" the premium to the other guys who are betting on convergence. So you will have the profile of an option buyer who will take consistent small losses most of the time and reap huge rewards every now and then. Just so you guys have a little heads up here, modeling options on correlation is one of the hardest things you can do in the financial world. It's extremely complex. Yeah I get it, you can show mom and pop how easy it is to pick up those nickels most of the time and leave out the part about the math. The pill goes down much smoother that way. It's usually a moot point anyway because retail margin does not make this strategy margin effective. Essentially putting up twice as much capital to earn a fraction of the return. And then there are the commissions. Don Bright (may he rest in peace) was smart to provide this niche in the prop world where this strategy was at least effective margin wise. Miss ya Don.
For quite a large variety of reasons, I think, including the following ... (i) There are probably more people trying to do it than you imagine; (ii) There are also HFT's doing it, which might make it hard to compete; (iii) It doesn't work, for most people who try it; (iv) Most people lack the numeracy/statistical/mathematical skill-set to do it, albeit that some are gullible enough to imagine that that somehow, magically, "won't really matter" when they try; (v) There's quite a turnover of people each trying for a while to do it, until they gradually come to appreciate many of the difficulties and drawbacks not mentioned in that article; (vi) As Mav so accurately and succinctly summarized the position in his excellent post above, "the author's full of shit".
i did alot of pairs trading with Don/Bright, 10 years ago or so. If i were still there, i might still be doing pairs... I would guess it's still profitable, with the proper approach/research, etc. I don't think i would trade pairs on my own. On my own, semi retired, i'm content with mostly options [those were some really good days, trading with the Brights] marc
It does. Do you think option is about penny and bulldozer ? No. Of course. It's an analogy. What do we compare ? Payoff ... You get killed or survive, a bit better than before.