John

1,626 reputation

312

bio	website
	location
	age
visits	member for	1 year
	seen	2 hours ago
stats	profile views	153

	1,626 reputation	bio	website		visits	member for	1 year
312 badges		location			seen	2 hours ago

summary answers questions tags badges favorites bounties reputation activity

146 Comments

suggestions reviews revisions comments badges posts accepts all

Mar 1	comment	Optimization procedure for entropy pooling I have noticed that taking extreme views can tend to cause problems (the effective number of simulations being low is usually a good sign). One simple way to check the results is to assume it is a normal distribution and take the views using the analytical formula and see how the results change.
Feb 26	comment	Market weights for Black-Litterman It can be difficult to obtain market capitalization for some types of assets. Instead, you can use the weights to an arbitrary benchmark portfolio. That would be like backing out the returns that would result in you investing in the benchmark portfolio if you don't have any views.
Feb 16	comment	Computing the Sharpe Ratio @Freddy I was referring to when people create portfolios by maximizing expected Sharpe ratios (not my preferred way, but people do it). It's not done the same as evaluating a portfolio ex post.
Feb 15	comment	Computing the Sharpe Ratio I think the point that @Freddy has emphasized is the ex-post evaluation use of Sharpe ratio. He would presumably give a different answer for optimizing a portfolio using the ex-ante Sharpe ratio.
Feb 8	comment	What continous adjustment methods are firms using for futures backtesting? What do you think about using optimization to drive the transactions as compared to rules like these?
Jan 31	comment	How to work out weights for a portfolio based on an inverse ratio with positive and negative values? You mean when EBIT<0? Then just take EBIT/EV instead.
Jan 31	comment	What to do with linear regression or regression splines outside of the training range? There seems to be answers your link. You aren't satisfied with them? This topic is more relevant over there.
Jan 23	comment	Robust-Bayesian optimization in Markowitz framework His book has appendices that show the derivation for robust Bayesian optimization (available at symmys.com)
Jan 18	comment	How to simulate one-minute bars data from one-day bars? You're not able to get one-minute bars? Why not simulate one-minute and then construct simulated one-day bars using that information? At least that way it is consistent.
Jan 17	comment	When hiring a quant, how can I protect my IP? @Freddy, the only other case I recall off the top of my head is Eric Falkenstein and Telluride, but I'm sure there are others.
Jan 17	comment	When hiring a quant, how can I protect my IP? NDA may not prevent the theft, but it can provide recourse (though they are more effectively used by larger organizations that can afford to pay lawyers for an extended period than a one-man shop likely would). Nevertheless, I'm still impressed by how much you(@Freddy) had thought it through.
Jan 14	comment	Regime switching in mean reverting stochastic process I wasn't talking about that specific class of models, but you may need to get more data.
Jan 11	comment	Regime switching in mean reverting stochastic process This is closely related, but I'm not sure whether that R package can fit autoregressive models that you are looking to fit.
Jan 10	comment	Regime switching in mean reverting stochastic process You might want to learn more about regime-switching. Also, try to fit one of the models using that Matlab package to get a better sense of what you're dealing with.
Jan 9	comment	What data transformations to use in regression of credit spreads on equity prices? That's effectively PCA on a correlation matrix, so what I said probably wouldn't apply.
Jan 9	comment	What data transformations to use in regression of credit spreads on equity prices? @Freddy 1) If both the equity returns and spreads are in percent, then the equity returns will dominate the first PC (if done on the covariance matrix). 2) I used to create factors when some variables were stationary and mean-reverting, but I have been less comfortable with that approach over time. Have you noticed any downside?
Jan 8	comment	How to fit ARMA+GARCH Model In R? In practice, it is often easier to just make the AR part of the ARMA long enough so that it encapsulates however much MA the series is. This way you can just use normal regression methods instead of relying on numerical methods for ARMA.
Jan 5	comment	Yield of a risky bond @Freddy I see your point, but then what he really would want to know is, not what the yield is and how to calculate it, but how to price the bond. A broad topic, to be sure, but the OP did not provide sufficient information in the question to be sure that's what he was wondering about.
Jan 4	comment	Is this comment right about subadditivity? Your interpretation is the correct one. VaR is not subadditive, but Expected Shortfall/CVaR is.
Jan 4	comment	Yield of a risky bond To find the yield to maturity for zero-coupon or coupon-paying bonds, the calculation is the same whether the bond is default-free or not.