First, all of this is pretty standard knowledge in quantitative finance, and so there is little reason to believe this effect hasn’t yet been arbitraged away
I’m a little confused here: even if everyone was doing this strategy, would it not still be rational? Also what is your procedure for estimating the (μ,σ) for real stocks? I ask because what I would naively guess (IRR) is probably an unbiased estimator for something like (μ−σ2/2,σ) (not a finance person; just initial thoughts after reading).
So my understanding is that the more everyone uses this strategy, the more prices of different stocks get correlated (you sell the stock that went up, that drops its price back down), and that reduces your ability to diversify (the challenge becomes in finding uncorrelated assets). But yeah, I’m not a finance person either—just played with this for a few months...
Well if you take simple returns, then the naive mean and std gives (μ,σ). If you use log returns, then you’d get (μ−σ2/2,σ) - which you can use to get μ if you need.
I’m a little confused here: even if everyone was doing this strategy, would it not still be rational? Also what is your procedure for estimating the (μ,σ) for real stocks? I ask because what I would naively guess (IRR) is probably an unbiased estimator for something like (μ−σ2/2,σ) (not a finance person; just initial thoughts after reading).
So my understanding is that the more everyone uses this strategy, the more prices of different stocks get correlated (you sell the stock that went up, that drops its price back down), and that reduces your ability to diversify (the challenge becomes in finding uncorrelated assets). But yeah, I’m not a finance person either—just played with this for a few months...
Well if you take simple returns, then the naive mean and std gives (μ,σ). If you use log returns, then you’d get (μ−σ2/2,σ) - which you can use to get μ if you need.