1,890 reputation
2928
bio website fosstrading.com
location St Louis, MO
age 33
visits member for 3 years, 5 months
seen 3 hours ago

My formal education is in economics. My programming skills and computer knowledge are self-taught. I have an uncanny ability to find answers to questions outside my immediate expertise. I have a deep love of learning and I enjoy building / fixing all kinds of things -- software, computer hardware, engines, etc.


1d
comment plotting High Frequency finance data using quantmod
@MattWolf: I'm a delicate flower and prefer to be handled with white silk gloves (like the Stanley Cup). :P
1d
comment plotting High Frequency finance data using quantmod
@MattWolf: For someone who says we should recognize we all come from different angles, it seems odd to me how hard you're pounding the table about this. I was simply trying to add clarity to your initial statement that, "R is extremely poor when it comes to graphing". You hinted at some of the specifics in the same comment; I was simply expanding on them (I even agreed with you that R sucks for those purposes). It seems more like you have an ax to grind. Or maybe you assume I'm like those you've had bad encounters with. /me shrugs.
1d
comment plotting High Frequency finance data using quantmod
@MattWolf: Still not sure how it's relevant to the question though. OP doesn't have actual "high frequency" data, and hasn't specified a use case, so I'm not sure you can say that anything is "the right tool for the job". Sometimes static charts are the right tool, and those you mention are overkill. I'm not sensitive about word choice. I'm sensitive about over-generalized statements. It's simply wrong to say R is extremely poor at graphing. It's correct to say R is extremely poor at interactively graphing large, high-frequency data.
1d
reviewed Approve suggested edit on plotting High Frequency finance data using quantmod
1d
comment plotting High Frequency finance data using quantmod
@MattWolf: You will get a lot of push-back by saying, "R is extremely poor when it comes to graphing". But no one will argue with, "R cannot chart millions of data points, updated at high frequency, in real-time, while allowing the user to interact with the graph." Nor is it surprising that an open-source statistical language cannot do what special-purpose commercial software can do. I'm not sure why someone would expect it to. I see and agree with your point, but I guess I don't understand its relevance to the question.
1d
comment plotting High Frequency finance data using quantmod
@MattWolf: I misunderstood what you meant by "dynamic". I thought you meant interactive, but it's clear you mean "updated in real time and interactive". I'm not aware of any open source projects that can do that for the amount of data you suggest. We use CQG and Deltix for charting, but many people don't have several thousand dollars/month to spend on charting software, or $600 for a charting library. Also, static charts can be useful for checking data reasonableness (unusual price gaps, timespans with missing data, etc).
2d
comment Difference between Total Long Term Debt and Net Total Long Term Debt
Who is the data provider? Doesn't their metadata indicate the differences between the series?
2d
comment plotting High Frequency finance data using quantmod
This question appears to be off-topic because it was cross-posted and answered on StackOverflow.
2d
comment Finding Trailing Twelve Months on a month between two quarterly reporting periods
It is basic. Companies do not provide data for the interim months, so the value is either 1) unknown, 2) equal to the prior quarter end, 3) some approximation. Which of those 3 you choose depends on your analysis. Even if this weren't basic, it's off-topic because questions about accounting/corporate-finance are not related to computational/mathematical finance.
2d
comment plotting High Frequency finance data using quantmod
@MattWolf: It's not accurate to say "R is extremely poor when it comes to graphing", but I would agree that it's not well-suited to dynamically graphing large data series, since the default graphics device uses a static, pen-and-paper model. The fairly new packages ggvis and rCharts are steps in that direction though.
Jul
17
comment How is PnL calculated
Welcome to quant.SE. Unfortunately, your question is both too basic and too broad for this site. You question would be more on-topic if it summarized what you already understand about the calculations and asked a specific question about the unclear part(s).
Jul
16
comment ETFs have lower tracking error than Futures?
1) isn't the ETF error larger (0.0012959 > 0.0006794)? 2) I would expect the ETF tracking error to be lower. It holds the actual index components and has fewer transactions than the futures index. The futures contracts must be rolled periodically, while the ETF only needs to transact for index constituent changes.
Jul
15
comment How to construct a deterministic trading model based on a loess (local regression) model?
Sounds like you have a hammer and are searching for a nail. The data should inform the model choice. It's a mistake to choose a model and then search for data to apply it to, because "all models are wrong".
Jul
7
comment Isn't a perfect economic system always in debt?
This question appears to be off-topic because it is about economics, and is too broad.
Jul
2
awarded  Curious
Jul
1
comment Implementing A 50/50 Prediction Model Strategy
A model that's right <50% of the time can be profitable. Many trend-following trading systems are only right 20-40% of the time, but are profitable because the winning trades are so much larger than the losing trades.
Jun
28
comment Combining BHHH and Levenberg Marquardt
Related: Does R support switching between optimizers like STATA does?
Jun
14
comment Is the market really Normal. Is Implied Volatility Historically Correct?
It's still wrong. 10.99*sqrt(20/250)*195.6 = 608. 195 +/- 608 != 201/189.
Jun
13
comment The option values are different from two r package - foptions,rquantlib
Yes, it's in Fortran. You could compare the current version with archived versions of fOptions on CRAN. Maybe one of the older versions had a pure-R implementation.
Jun
13
comment The option values are different from two r package - foptions,rquantlib
The source code for both is available. Have you looked at it?