1,412 reputation
818
bio website quantdevel.com/public
location Elgin, IL
age
visits member for 3 years, 5 months
seen Jun 27 at 1:26

Quantitative developer, writer, instructor, private trader, so-so musician


May
8
comment How 'High' is the frequency in HFT?
@novice Thanks. Yup, those qualify as fast-moving markets, and your observation is good. Since I wrote this answer, I've seem more traders doing VHF trading on smaller budgets. In the past, I saw only organizations with substantial budgets -- for gear and people! Maybe the dividing line is between successful market takers (like you) vs. successful market makers. The VHF market makers I know do have big rigs.
May
8
comment How 'High' is the frequency in HFT?
@novice What market (or markets) are you trading?
Dec
17
comment Recreate Positions after downtime. Suggestions requested
I agree with Louis. If my computer was trading and it crashed, I would immediately cancel all open orders. That's just good risk management.
Dec
10
comment Portfolios from Sorts
Thanks for bringing this to our attention. The approach is very interesting. (PS - I fixed a bad link to the second paper.)
Nov
29
comment Maximum Likelihood using a Kalman filter for two factor model
Have you considered using R rather than Matlab? In R, there is already an implementation of the Schwartz-Smith model, including parameter estimation: cran.r-project.org/web/packages/schwartz97
Nov
27
comment Is price gaping the major risk that market maker has?
@MattWolf Thanks. Your point is well taken, and I am likely over-generalizing. I suppose I'm reacting to the OP's question: "[Are market gaps] the main worry of the market maker?" My opinion is that inventory management and liquidity issues are much larger problems than, say, competitors, connectivity issues, or price-model risk; and that market gaps are a cousin of those larger problems.
Nov
22
comment Is price gaping the major risk that market maker has?
In that situation, we say the market "traded through" your bid. It's still a problem in inventory management: you acquired the inventory, then the market moved against you. You are describing a special case of the general problem.
Nov
22
comment Optimizing stochastic functions numerically
This sort of difficult portfolio optimization is handled quite well by the PortfolioAnalytics package. The authors wrote an informal tutorial which explains the what and how.
Nov
15
comment Risk and Reward in practice
@Matt +1 Thank you for putting the human face on risk management. It usually gets lost in the numbers. I think many quant traders don't appreciate that dimension.
Nov
15
comment pairs trading detrend the spread
When you say "make the spread stationary", do you mean construct a feasible spread trade; or do you mean post-process the spread data to eliminate the trend component?
Oct
22
comment Control for bid/ask bounce in high-frequency trade data?
@Ilya That's not a typo. When there are many buyers, they repeatedly hit the ask price, lowering the available ask size. So we weight the bid less heavily, and the weighted price rises. Likewise, when sellers arrive, the weighted price falls.
Oct
22
comment How 'High' is the frequency in HFT?
@user2763361 Thanks. That's an interesting perspective. I trade daily data. If 5-minute bars are ultra low frequency, what would you call daily bars?
Oct
14
comment Time-series similarity measures
@vonjd Thank you for bringing this to our attention! Distance correlation appears to be a wonderful statistical device. I have incorporated it into my weekly scan for related markets.
Aug
12
comment How 'High' is the frequency in HFT?
@fermat: Thanks for asking. Actually, I never said it could not handle 1-minute data... I just said that handling 1-minute data was "somewhere in between". I'll stick to that vague statement. Your actual success will depend upon the trading latency required by your trading algorithm. (And remember: your computer has gotten faster... but, then, so has everyone else's.)
Oct
24
comment Why is the Drawdown measure not used for portfolio optimization?
Of all the professional investors I work with, not one of them uses simple mean-variance optimization. In fact, every one starts with a maximum drawdown parameter to bound the space of acceptable solutions (among other constraints), then applies some optimization within that space. As for assuming a parametric distribution of returns, I've never encountered anyone willing to make such an assumption.
Oct
24
comment Searching for pairs-trading in sub O(n^2 t) time
Common sense tells me there is no connection between hogs and Microsoft stock. I tried trading several of those pure quant pairs. After losing enough money, I decided that common sense does have a role in quantitative finance after all.
Oct
16
comment Statistical significance of trading systems that use indicators with long lookbacks
Your basic assumption is flawed. Quantity of historical data is independent of serial correlation. I work with trading firms whose model parameters are estimated from years of data, but the resulting signal has a half-life measured in hours.
Oct
12
comment How to account for bid/ask spread when backtesting?
Assylias is right. If you trade exclusively with limit orders, then the bid/ask spread is irrelevant. Instead, your slippage comes from "trade-through", when the market moves through your limit price before the broker can fill your order. That, in turn, is driven by the market's liquidity and volatility.
Oct
11
comment Trading a synthetic replication of the VIX index
@Strange Thanks for contributing your real experience. All in all, I have concluded that replicating the cash VIX is too difficult. I am sticking with VIX curve trades using only the futures.
Oct
11
comment Trading a synthetic replication of the VIX index
@Jared Thank you very much for your thoughtful and complete reply. These are excellent observations. Thanks, too, for bringing the variance futures to my attention. The "if they attract liquidity" thing is a problem. Does any contract but the Big VIX really trade on the CFE?