8,934 reputation
1528
bio website
location Hong Kong
age
visits member for 2 years, 4 months
seen 13 hours ago

Jul
21
comment plotting High Frequency finance data using quantmod
No Joshua, I certainly do not have an ax to grind and neither did I form any sort of impression about who you are or what you do and I actually enjoy to read all your content and appreciate you sharing your thought. I guess it still takes some getting used to that quants are not traders and that some prefer to be touched with silk gloves. My apologies if I rubbed any elbows in the R community.
Jul
21
comment plotting High Frequency finance data using quantmod
@JoshuaUlrich, last comment and just to reply to your last statement. Would you know of any charting library in your chosen language that is cruder, more simplistic, and contains less features than the touted new "Visualization weapons" in R? I do understand the excitement for R's charts given someone comes from a unix/linux environment but all I was saying is that it does not have to be that way. My whole point was that there are better tools out there. Even the open-source ZedGraph (albeit targeting .Net) that was developed years ago makes anything in R 2014 pale in comparison.
Jul
21
comment plotting High Frequency finance data using quantmod
Sure sometimes static charts are just perfect, if the axis labeling works, if the scaling properly works, if the chart series style can be easily adjusted, if axis types can be quickly changed, if the layout can be customized, among many more. Else you end up like 95% of academic papers, sometimes great and valuable content, but so off putting and poorly presented visually that you sometimes have to force it down practitioners' throats. I do not have a beef with R, but I do smirk when each time a war breaks out because a feature of R is criticized.
Jul
21
comment plotting High Frequency finance data using quantmod
I have not attacked anyone nor even criticized someone, certainly not personally. We should recognize we all come from different angles, for some of us even a calm debate is a major stress factor and for others bluntly speaking out what they think does not affect them at all. As long as we do not personally attack each other I think we should recognize each others' styles and let go of the oddities and rather focus on benefiting from our capabilities and knowledge. We could have spent the time more productively discussing what is lacking in R charts instead of minding word choices. My 2 cents
Jul
21
comment plotting High Frequency finance data using quantmod
The relevance to the question is that I recommended to stay away from R if you want to visualize data properly. Simple as that. And I am time and again confused why some in the R community are so extremely sensitive about word choice. I stated facts not opinions and if my usage of "extremely poor" hurts the developers of a static chart package in R then that was certainly not my intention (to hurt them) but I also will not retract it because you can simply not do much with the charts other than look at them and its a major pain to even get static charts right in R.
Jul
21
comment plotting High Frequency finance data using quantmod
@JoshuaUlrich, I did mean interactive, panning/zooming/axis-label-annotation interaction...Joshua, my understanding so far was that this website caters to professionals in the financial sector, and $600 should hardly prevent one from using the right tool for the job (and this is for a developer license not a per-user license). So, just because R does not charge for a retail license does not mean it is the most optimal for each and every task. Visualization is part of the game, maybe a small part, but those who do care should use proper tools imho.
Jul
21
comment plotting High Frequency finance data using quantmod
@JoshuaUlrich, I would be highly surprised if anyone was satisfied with anything less than the ability to chart hundreds of thousands if not millions of millisecond or higher precision timestamped data at high framerates, while still being able to dynamically interact with visualized data for data mining purposes. I know of a few individuals who despite their Linux quant desk setup have a windows machine just in order to run high quality visualization tools for their quant strategy profiling and testing purpose and it makes a lot of sense to me.
Jul
21
comment plotting High Frequency finance data using quantmod
@JoshuaUlrich, I stand by my statement, with all due respect where visualizations came from in R and with due respect to the fact that ggvis or rCharts are a huge improvement over what has been in existence before, they are incredibly basic, static, almost to the point of naive. We are in 2014 and claiming that any .js type charting or packages in R provide satisfactory charting solutions is imho quite an overblown statement. Again, we should look at the industry leaders in this space not where R came from to set suitable benchmarks.
Jul
20
comment plotting High Frequency finance data using quantmod
By the way, R is extremely poor when it comes to graphing (I know I get attacked left and right by some in the R community for saying that). But until now there is not one single library that allows you to dynamically chart large data series. Visualization is a huge miss in R and Matlab as well. Take a look at libraries such as SciChart for .Net/WPF and you know exactly what I am talking about. I was never entirely sure why someone wanted to stare at a static line chart, you get the same information by looking at descriptive statistics of the underlying dataset.
Jul
20
comment plotting High Frequency finance data using quantmod
First off, I do not recommend dealing with high frequency data sets that cannot even display millisecond timestamped data. But you mentioned you want to aggregate the data anyway, so given you are able to group the data into buckets of your aggregation choice, calculate the high and lows in each bucket and also be able to determine the first and last price in each bucket and from there you can easily derive your OHLC bars. But as mentioned you need to be careful with your imprecise timestamps you may have a hard time to extract the first and last prices after you group.
Jul
18
comment ETFs have lower tracking error than Futures?
Agree, it depends on the ETF and actual asset class, in case of VIX futures vs VIX ETF (VXX) the ETF has a much larger tracking error. Case in point, VIX Index advanced 39% at some point while the ETF advanced less than 9%. The futures fared a little better. The high tracking error (to the index) is related to several VIX specific issues but this is an example where generally the ETF tracks worse than the futures.
Jul
15
comment How do I calculate Sharpe ratio from P&L?
Of course it makes sense why would it not, it is important to know whether a high frequency trading strategy has sufficient turnover and generates sufficiently positive returns in the context of risk, taken, whether we talk daily Sharpe or annualized Sharpe.And as said bootstrapping, resampling, drawdowns, nor avg win/loss has anything to do with the question at hand. Anything that generates income (even your salary) is somehow capitalized, hence the ability to calcultate returns.
Jul
15
comment How do I calculate Sharpe ratio from P&L?
That is simply incorrect. Of course do any trading strategies allow for the computation of return metrics. How else do you think some hft houses generate and publish 7+ Sharpe ratio performance metrics? With all due respect but you do not sound like a market practitioner at all. Resampling is certainly not done by professionals in this context. And why do you talk about drawdowns or recovery periods when commenting on risk adjusted return metrics?
Jul
15
comment Sharpe ratio in days with no open positions
By the way your definition of information ratio is incorrect. Information provides a return in excess of a defined benchmark in the context of risk. It has nothing to do with signal-to-noise, not sure how you make that connection.
Jul
15
comment Sharpe ratio in days with no open positions
Call it whatever you want as long as you make the proper disclosure you can define and use whatever risk adjusted return measure pleases you. I merely reflected what from my experience in the market seems to be the accepted practice. You have of course a right to disagree and write up your own answer to reflect how you think the world ticks, which you have done. People care about understanding generated returns in the context of risk taken and if that excludes days on which no returns were generated and everyone knows through disclosures then everyone is happy.
Jul
15
comment Sharpe ratio in days with no open positions
Well you do not generate any returns, excess or not, on days you do not have open posions or close open positions.
Jul
15
comment Sharpe Ratio - my own calculation differs from Yahoo finance, Morningstar
Not saying that your answer is in any way deficient, just wanted to add that little bit of information. Cheers
Jul
15
comment How to compare the results of covered call and protective put option strategies?
the expected cost of a covered call is the expected return of the stock above your break-even (strike plus premium, received) over the expected holding period of the strategy. There are probably many ways to arrive at such estimate, not one single approach is necessarily the only correct one. What matters is that you are consistent across the sampling period.
Jul
15
comment Sharpe Ratio - my own calculation differs from Yahoo finance, Morningstar
I would argue that the inclusion of the risk free rate is a bit out-dated. I know of a number of hedge funds that do not include the risk free rate in their computations (albeit they mention it in the disclosure documents). There is no real risk-free rate, sovereign bills or bonds are anything but risk-free, including bills issued by the US Treasury. Imagine short rates were 20% then the Sharpe ratio with inclusion of Rf would be badly skewed. Sharpe is not a comparable risk measure but measures return relative to its own risk.
Jul
15
comment Sharpe ratio in days with no open positions
I disagree, because Sharpe is a measure of excess risk-adjusted, returns. It does not matter whether competing investments generate return during days you are not taking any risks. Sharpe is not a comparable performance measure such as the Information Ratio, its focus is on its own performance relative to its own risk taken, which is why a lot of practitioners actually leave out the risk-free rate altogether. I am not arguing here whether it should be included or not, but I try to make a point that days on which risk is not taken should not be included in the Sharpe ratio computation.