1,166 reputation
726
bio website twitter.com/shabbychef
location San Francisco, CA
age 42
visits member for 3 years, 8 months
seen Feb 5 at 18:32

matlab/stats/linux nerd. proud, sleepless, new parent. HCSSiM alum. faking it as a quantitative analyst at a small quant fund in San Francisco.


Jun
1
comment How high of a Sharpe ratio is implausibly high for a low-frequency equity strategy?
This is one thing that burns my butter: Sharpe ratios published without units! I cannot tell from Exhibit 2 shown above whether the SRs are monthly, quarterly, or annualized. It matters! (Although in this case, not terribly).
Feb
26
comment How are risk management practices applied to ML/AI-based automated trading systems
@wonghang at the current point in time you have to decide how to deploy your money over the next time delta. You can use all the data available to you to both select the best model and estimate its performance. If you do so, your estimate of performance is biased upwards 'by selection.' You can instead partition the data into two sets, one for selection, the other for estimation. This increases the chance of making a selection error and increases the standard error on your performance measure. Representation length can easily be confounded by introns, BTW ...
Aug
1
comment How high of a Sharpe ratio is implausibly high for a low-frequency equity strategy?
@Freddy I don't believe it either, and find it absurd. The number is from a backtest performed by a third party. My job is to convince someone that this figure is suspiciously high. If I just tell them I think it is too high, it is my word against someone else's. This is why I am looking for a published account that, presumably, has been vetted and is representative of achieved performance.
Jun
30
comment Can social media be applied to algorithmic trading?
The Bollen paper is very dubious; it has been thoroughly discredited here: sellthenews.tumblr.com/post/21067996377/noitdoesnot
Nov
9
comment Can social media be applied to algorithmic trading?
evidently Derwent is trying to reproduce the results! They are all over the media for this because Twitter is trendy. Not only is it free advertising, it may end-run around rules about hedge fund advertising. While I agree with you that the results are very unlikely to be reproducible, it doesn't much matter if the hype gets them clients.
Nov
8
comment Can social media be applied to algorithmic trading?
This is brilliant advertising for Derwent (hedge funds typically cannot advertise in the US), but the original paper by Bollen et al (now famous for being famous) is laughably inadequate. (just plot the empirical distribution of the 49 p-values they quote, for example.)
Nov
8
comment Can binary model lead to non-normal distribution?
you would have to violate the conditions of the Central Limit Theorem. For practical purposes this means having a non-finite variance to the binary jumps.
Sep
20
comment How can one compute the Greeks on VIX Futures
hmm. I think I get it. Although the expectation is of the future value of the basket, and the derivative is with respect to spot basket value, right? In that case, I think I can use the 'Delta method' (after genuflecting before some regularity conditions) and get a good approximation. (Well, I view Delta method as Taylor's theorem plus expectation magic.)
Sep
7
comment Should Sharpe ratio be computed using log returns or relative returns?
that really makes no sense to me. If the AUM of the fund changes (investments/disbursements), or there is a split in the stock, or a large change in nominal value, you cannot compare dollar returns from one time period to another. Did I misunderstand your comment?
Jul
22
comment Should Sharpe ratio be computed using log returns or relative returns?
what exactly do you mean? returns with dollar units?
Jul
22
comment Should Sharpe ratio be computed using log returns or relative returns?
This is a good point. For my purposes, I have the daily (or even higher frequency) marks because I am looking at equity portfolios.
Jul
22
comment Should Sharpe ratio be computed using log returns or relative returns?
daily returns as percents or in log returns?
Apr
28
comment Predicting Price Movements on a Betting Exchange
Do you mean Albert Kyle?
Apr
13
comment How does UBS hedge its exposure to XVIX ETN?
XVIX tracks a spread in VIX futures prices. Any tracking 'error' attributable to the futures prices not matching spot VIX should not be attributed to XVIX failing to track...
Apr
13
comment How does UBS hedge its exposure to XVIX ETN?
It's an ETN, not an ETF. they do not have to publish holdings.
Mar
23
comment How to perform risk factor calculation?
if you perform e.g. a en.wikipedia.org/wiki/Gram_schmidt normalization before the regression, you will have an orthogonal design matrix. However, it is difficult to interpret the resultant regression coefficients.
Mar
23
comment How to perform risk factor calculation?
This is only the case for simple linear regression with a single factor (i.e. CAPM), and not for multiple linear regression. If computing a multiple linear regression were that simple, there wouldn't be the vast mountain of literature on the topic (do a google search for 'solving normal equations' )
Feb
20
comment How would you test the hypothesis “There are no idiosyncratic returns available in the market”?
It seems for either of the two proposed tests I will have to correct for the correlation of stocks returns to each other (beyond just via the market) and possibly make some correction for multiple hypothesis tests. I was hoping for a more consolidated approach that looked at all stocks returns simultaneously. The graph presented with the commentary is supposed to be illustrative, I guess, but I am not sure what it tells me.
Feb
6
comment How to combine various equity measures into a single measure (vector magnitude)
What is the purpose of combining them?
Feb
3
comment How are risk management practices applied to ML/AI-based automated trading systems
@Lirik: no, I am thinking about what happens after 6 months of paper trading with mediocre results. Does one give up on finance and become a plumber? Or does one fiddle around with the algorithm, the data, etc? You always have only the data you have today, when you are deciding what to trade tomorrow. If you have any choice and the historical data guides that choice, you have datamining bias.