- Search Energy EXCH
- 52 Results
- ngawang replied May 8, 2009
SNR & Delay — Absolutely correct. I have the option in the code to either subtract the mean or both subtract the mean and detrend. I had the detrend option off for that calculation. FYI Ngawang
- ngawang replied May 7, 2009
SNR & Delay Again — Actually that synthetic price was Price = 20*Cycle + Trend + Noise If you change the formula to Price = 1*Cycle + Trend + Noise Then the delay comes down considerably -- to more like 7 or 8 observations before SNR hits ...
- ngawang replied May 7, 2009
Delay in SNR — Hi Krzysztof, Here's an attempt to answer your original question about the delay in Signal to Noise calculations. The top half of the chart shows a synthetic price which cycles with period 20 for the first 460 observations. ...
- ngawang replied May 6, 2009
Amplitude — Hi Killbill, I have good news and bad news. The good news is that probably nothing is wrong. The bad news is that probably nothing is wrong. Amplitude calculated by the Goertzel algorithm isn't designed to equal the amplitude of ...
- ngawang replied Apr 29, 2009
Strange SNR Chart — The seemingly strange SNR chart I posted before comes from assuming there are multiple dominant cycles in the data where your signal only has 2 cycles. Thanks for the test signal. I'll check to make sure the results make ...
- ngawang replied Apr 28, 2009
Missing Link — Krzysztof, Attached is a shot of the signal you posted a while back along with the signal to noise ratio the way I'm computing it now. In case it's not clear. The blue line is my attempt to replicate your 20 & 100 period sine ...
- ngawang replied Apr 14, 2009
Missing Link & S/N ratio — Hi Krzysztof, Thanks for those observations. I haven't found anything better yet but am definitely interested in variations on the theme that might do a better job. Thx Ngawang
- ngawang replied Mar 26, 2009
Thread Continuation — The concern about replication makes sense. I'm all for continuing the discussions, but I don't see ho you do that in a public forum without revealing things you would prefer not to reveal. If you transform into a private ...
- ngawang replied Mar 26, 2009
Wiener Filters — Is anyone familiar enough with Wiener Filters to answer this question? I found the following source on line and was reading it to develop a better understanding of Wiener Filters... url The author breaks Wiener Filters down ...
- ngawang replied Mar 20, 2009
What's the missing link? — Hi CB, This is great news that your strategy is paying off so handsomely. I hope that continues in all market environments. I'm getting to the point where I have almost all the pieces of the puzzle together to begin ...
- ngawang replied Mar 13, 2009
One Typo — Bk = -i*(yspec - yspecs)/2; Otherwise, I think it ties out. FYI Ngawang
- ngawang replied Mar 13, 2009
Correct Fourier Transform of Real Series — Meant to post this a while back. The correct schematic for the Fourier Transform of a real series is as follows: x(n) is a real series of length N whose Fourier Transform we wish to compute. We can do ...
- ngawang replied Mar 9, 2009
Goertzel — dizzyd, Check the exchange between Barhopper & CB. It clarifies pretty extensively the correct algorithm for Goertzel. I think you'll find that your definition of imag should change to: imag = s1 * np.sin(2.0 * np.pi * omega) This ...
- ngawang replied Mar 5, 2009
FFT Resolution — Barhopper, Thanks for looking at that. Padding with 0s is infintely more straight forward. I was trying to do something from my notes, which were not complete and probably not 100% correct. I'm attaching as a PDF the section ...
- ngawang replied Mar 4, 2009
Never Mind That Last Post — The symmetry of FFT should guarrantee that Goertzel and FFT are equivalent. To prove it, I think we need to create a new series xprime, which is half the length of x. xprime is a complex valued vector where x was ...
- ngawang replied Mar 4, 2009
Resolution Advantage for Goertzel over FFT? — Hi Barhopper, (and anyone who has thought about this problem) You recently proved that the Goertzel and FFT algorithms are identical. I can replicate your result if I use the following Matlab code, ...
- ngawang replied Mar 3, 2009
Causal / Non-Causal Filtering Idea — First of all, congratulations to CB on the great trading results. Hope this continues. I've been a bit swamped & so haven't been able to test that last idea about filtering forwards through the data and ...
- ngawang replied Feb 18, 2009
Causal vs Non-Causal Filtering / Noise Reduction — There have been a few posts by people objecting to using non-causal filtering or non-causal methods of any kind because they assume that you have that look into the future when filtering past ...
- ngawang replied Feb 17, 2009
Missing Data & Close to Open Transition — Hi Barhopper, I messed up your original quote when writing that last post. It was probably clear, but just in case, this is the post I was referring to... Thanks, Ngawang
- ngawang replied Feb 17, 2009
Close to Open Transition — Hi BarHopper, I may be derailing the purpose of your questions slightly, but I see another place where this question comes into play, and that's where you don't have 24 hour markets as we essentially do in FX. I ...