200 reputation
5
bio website
location
age
visits member for 2 years, 1 month
seen Apr 17 at 17:17

Nov
1
comment Expected length and depth of drawdown
Thanks for the useful comments. Maybe I should clarify a bit. I am more interested in "analytical solutions" that relate drawdown statistics to portfolio assumptions. Even though this is clearly not realistic, I think it can provide insights into the interaction between portfolio parameters and drawdown, whereas simulation probably gives you a better estimate but it's hard to discern insight from the results.
Oct
19
comment Risk neutral probability in binomial short rate model assumed to be 0.5?
I forgot the exact books, but here they are also simply assuming 0.5 risk-neutral probability: link and link. My impression is that, in the stock binomial model, you assume up and down movement and calculate the risk-neutral probability. In the short rate model, you instead assume risk-neutral probability and match up and down movement to what's observed in the market. But I am not sure if this is the right way to view it.
Oct
19
comment Risk neutral probability in binomial short rate model assumed to be 0.5?
What confuses me is that I have seen two introductory level textbooks that simply assume risk-neutral probability of 0.5 as if it should be obvious, without resorting to SDE at all. That's why I thought there should be a simple/intuitive explanation(not that I don't appreciate your explanation).
Sep
27
comment Fastest solver possible for portfolio optimization
I forgot to mention that I was already using parallel processing. I ran the profiler on my code an quadprog is the one taking the most time simply because it gets called so many times. So anything to speed it up will be very valuable.
Sep
27
comment Fastest solver possible for portfolio optimization
The optimization itself is very simple, just standard mean-variance optimization with standard nonnegativity and full investment constraints. The only problem is that I need to do this daily for say 30 different portfolios each with 100 assets. So it could take hours in aggregate. I have been using quadprog in Matlab with interior-point-convex algorithm. It is very possible that there is nothing I can do to speed it up, but I thought I tried my luck by asking here.
Sep
10
comment Can Hurst exponent be used to characterize nonlinear dependence in time series?
I am more interested in assessing the predictability of time series in univariate settings. For example, if a time series is driven by a deterministic nonlinear dynamic, it can appear totally random if you look at its ACF, even though in fact it's not random at all. From what I understand, the Hurst exponent cannot help to detect the nonlinear predictability in this case, but I just want to see what people's thoughts are.