Unlike this website, which has scores of references to Jurik, a search reveals only four references at wilmott.com. Most of the four mentions refer to the book he edited, Computerized Trading. The fourth is a thread named "smoothing algorithms", and only the OP talks about the Jurik moving average. All of the 50 replies studiously avoid discussing Jurik. I find that hilarious.
Why are quants afraid of Mark Jurik?
- Thread starter kut2k2
- Start date
- Status
Are there any other types of studies that show merit to Jurik?
I've seen a proprietary version of his algorithm reconstructed, but it wasn't anything unusually special, it ended up being something like a 7 tap FIR.
What do you find especially useful about his work?
And is he known for anything outside of relatively smooth/low-lag filters?
I haven't followed his stuff much, but he comes up an awful lot in DSP related filter forums. Part of the reason he's not likely discussed is that most of his stuff (as I understand it) is proprietary, outside of the reverse engineered FIR I referred to-- it would be hard to substantiate the usefulness of any of his work.
I've seen a proprietary version of his algorithm reconstructed, but it wasn't anything unusually special, it ended up being something like a 7 tap FIR.
What do you find especially useful about his work?
And is he known for anything outside of relatively smooth/low-lag filters?
I haven't followed his stuff much, but he comes up an awful lot in DSP related filter forums. Part of the reason he's not likely discussed is that most of his stuff (as I understand it) is proprietary, outside of the reverse engineered FIR I referred to-- it would be hard to substantiate the usefulness of any of his work.
It's pretty well accepted that the JMA is the best smoother. He even offers a money-back guarantee if you can come with an algorithm that performs as well; the details are posted at his website. He also offers a variety of other interesting indicators and trading tools. I believe it's well worth any technician's time to go through his website, even if you don't plan to buy anything.
I don't know the formula for the JMA but I'd bet it's more complicated than a 7 tap FIR.
The JMA can be a proxy for price. Instead of using raw price which fluctuates wildly and can/will produce many account-draining whipsaws, using the JMA as the fast MA in a cross-over strategy will cut down drastically on whipsaws while maintaining a low lag at the same time.
Dunno where all the millionaires are ... maybe they figure the less they say, the less competition they'll get.
Myabe, other posters discuss the book and not the software sold separately. In a thread dedicated to a book it is quite reasonable to discuss what can be learned and used from the book and not the background of the author or any unsubstatiated marketing claims. (Read Dilbert comics regarding why engineers, and quants, usually stay on topic.)
The above doesn't mean I specifically doubt Jurik's filtered moving average. It's just people who create and test their own filters are extra cautious about everything that is "proprietary".
On a separate topic, filtering is one of the most important things in higher-frequency auto-trading. In signal processing filtering is one of the key instruments. Sometimes the precision of a measurement can be improved by an order of magnitude by applying the right filter. Same can be extrapolated onto trading signals.
On ET pimping the JMA, obviously.
HMA supersedes it, anyway, and is non-proprietary. You can_do it in Excel, if you like.
http://en.wikipedia.org/wiki/Arnaud_Legoux_Moving_Average
Sources please. (Frankly I think I can come up with results better than your "reverse engineered JMA".)
No, I don't work for Jurik, I've never met him or even bought any of his products, nor do I own any stock in Jurik Research. But the man definitely thinks outside the box and that makes his products worth a look. Just passing it along.
- Status