What kind of statistic software is that ? That is a coolest tool ever posted on ET. I had no idea there is something like that available. Can you post link ? Thanks, Walter
It's Ensign's tool. What if you run those stats through a Neural network and added say hourly functions or even 5 minute routines instead of daily stats, added some rules like standard tech stuff (RSI maybe and MACD) as well as some esoteric stuff like the first chart I posted here (whick is root-2 derived and Gann's So9) and you would have one model I use as described here http://www.elitetrader.com/vb/showthread.php?s=&threadid=25533&perpage=6&pagenumber=6
i thought when the dow was in the 7k range prechter was predicting a crash to 4k.when the guru of ew is almost always wrong how can one take this stuff too seriously? http://www.siliconinvestor.com/stocktalk/msg.gsp?msgid=19466294 Robert Prechter - DOW below 4,000 by end 2003 April 2003 Blast from the past Courtesy Forbes- http://www.forbes.com/2003/04/09/cz_jd_0409watch.html Robert Prechter, editor of the Elliott Wave Theorist, is about as bearish as they come (as you might detect from the title of his 2002 book, Conquer the Crash: You Can Survive and Prosper in a Deflationary Depression). Prechter views the stock market as a meter of mass psychology and social mood that fluctuates in a series of (Elliott) waves found throughout nature. The primary trend in stocks since January 2000 reflects anger, fear, negativity and social polarization, but occasionally yields to brief periods of optimism, such as the rally in stocks that began a week before the war.
I bet I can top that as the coolest tool. This one is a tool of numbers and the mind. Take a swing low like 9/30 and subtract 10/15 from it. 15 days in between, sqrt that to get 3.8729, add 2 and resquare to get 34.49. Now you have a time target of 34 or 35 days from 9/30 to look for a swing high. Add 34 to 9/30 and you get 11/03 a swing high. 10/15 to 10/24 is 9 and this formula gives a date of 11/9 (which is a sunday so round up to 11/10) and the pivot date was 11/11. This can be shrunk down to intraday time frames but this is not going to be explained here (and never will) because it's a HUGE success. Another techinque is low to low sqrt dates. Oct 10 2002 to March 12 2003 is 153 days. Multiply sqrt 2 gives 216, added to 10/10/02 is 5/14/03, very close to a swing top. Multiply Sqrt3 is July 2, a day late of a swing bottom (sqrt 3 has VERY big implications from a theoretical sense) and sqrt 5 give Sept 17, the day before the swing top Sept 18. This as well can be shrunk down to intraday but that's something that I'm not going into here, or anywhere for that matter. Just a sample of what's possible.
I did a lot of research on different predictive methods and I am a firm believer that predicting is a valid method. Accuracy of intraday called tops or bottoms is an issue here if this was to be used commercially. It was shown here , on ET , that even small 10-20 min window is not accepted by general public I would like to see your intraday method(s) in action . Would you consider posting a few tops or bottoms for some well known issue ? QQQ, SPX or MSFT come to mind. Thanks, Walter
No, I have no desire for commercial application of my methods. Traders trade. But if you want my advice on finding a precise timing tool try something like what I did. I have three neural networks that exchange data and results; one is pure statistics of price data, dates and times , number 2 is a radar knockoff that finds patterns inspired by the work of RL Pimmel, and the third is geometric technical analysis ( sqrt's, So9, Fib, squares and such) and the timing of all three is almost perfect. Still needs some work but I'm getting there. Of course the poors mans methods could have worked just fine today, see the stats below.13 Mondays in the last 60 days statistically averaged a closing high after starting at the low in the first half hour, and achieving a higher high on 5 swings to get you long. Seemed like a no-brainer to me.
Off topic and not a personal attack, first hand... But saying you have a "model" seems more sophisticated than saying, "routine" or "system"... Again... no offense... just my random thought... In a sense, I only have one esoteric "model"... Everything is connected and is a whole rather than parts of the whole... (not saying I know "All"... just Law of Vibration... )
Thanks, I was not suggesting that you should sell your signals. I am curious how accurate can you be. Something along the lines like " Do you get 80% of your calls within 20 min ?" I am using geometry and some patterns but did not know that these kind of statistics are available so I want to thank you for your help with that. Walter