Help! I dont know much about computers, and Jack wants to sell me his old computer for $2500...... is this a good deal???? http://ed-thelen.org/comp-hist/vs-cdc-6600.jpg
Jem has correctly posted that during the time periods he references several (we both used 4 it turns out) feeds were required. Among the four feeds some are fairly fast and some are fairly slow ones. Logially the slow ones are used for coarse and others are used for medium and fine. I regard the Stretch/Squeeze as a precision tool. And, therefore, it requires a fast feed. I use a fast feed. It is probably clear to most that that screen shown has no feeds from Qcharts. As a "prove in" on perfomance (between thanksgiving and Xmas)and in isolation. It was possible to "tune" the "noise band" and the outer trading band (in combination with a detector for volatility compression and expansion) to extract over 11 points in a three point range (under 150 minute segment)( I may have said 190 before). Using S/S alone and not just visually, approaches the limits determined and posted by MAK. 3x the H-L is not a limiting case for precision trading, in other words. MAK is correct at the 7x level. Others use 6x in the literature. I apologize for not making it clear enough that index trading requires feeds other than Qcharts for fine sweeping. Ordinarily in the Tucson area we use the Sacramento feed for Qcharts and keep it referenced to a clock on that display screen (not shown) when using several feeds, anyone knows what is what. See next paragraph as well. As you can see sweeping is required to collect a data set. In trading, a person grooves by using NOW to trade and another time related to NOW. For me it is a time which is slightly in the future and therefore, I am NOT doing anything that is in REACTION to price. Price, to me, for trades is something I find out later on prints. I never know until I am told what the price of a reversal was. I was gong to reply previously and now I am glad I didn't. I apprciate jem's viewpoints and the reasons he uses to knock what I do and suggest to others. We, meaning a lot of very hard working people, are making a continuing effort to ameliorate the problem in the first sentence. As jem has explained he is a decade ahead of me and at the same time what I post is bullshit. I gather he has tidily solved the problems we are currently working on. I do not use only the feeds he assigned to me and I recognize that he doesn't either. It is obvious a combo FA/TA single feed is required. For PVT which is a coarse feed, unusual volume works quite well on the Qcharts platform. Script makes it possible elsewhere. If anyone knows how to bring up old Qcharts displays that were saved (stopped use of) a few years back, let me know by PM. I need to access, without updating, many saved past Qcharts files. Think in terms of quarterly duration anotations of coarse components.
so is it your position that you did not tell people you used qcharts for the data for your you indicators? What exactly were you using for your 4 data feeds a few years ago when you were making your posts.
You're so delightfully funny. I find particularly entertaining the reference to Qcharts as a feed. I may not be using the most elegant metaphor, but to me you sound like a former European electrician who used to service power supply for the instruments in one of European the labs, who upon stumbling on "Scientific American" for the first time, fires off the letter to the editor, rushing to warn the public that all progress made in American science is a snake oil, because American instruments run on 110V vs 220V in Europe, and therefore are way too slow. So everything they have found so far is wrong, and therefore they are not even real scientists. Referencing yourself as an expert with the air of implied authority is a very nice touch: "I know, what I'm taking about - I used to repair power supplies". Two things come to mind about somebody like that: he's talking about the subject which is way outside of his area of expertise and he's obviously not bright enough to even realize how ridiculous he sounds.
Hi jem, My position is similar to yours. Both of us use the best that we are aware of that is available. There is a "trading forum" running on how weird I am. It was begun by a person who took the liberty to post a pm I sent him in response to a post similar to those you make about me and how I trade and what platforms I use. I have never been concerned about data timing in terms of its path and arrival. Nor have I spent any effort in improving the paths. I try to use the best available info at any time. I do spend hundreds of dollars a month for my inputs. And I do have one inconvenient problem of recovering data from a source I no longer use. (I am writing five books and their illustration needs a little tweaking here and there.) You content that I am full of shit and I can't get done this and that. What is conceivable to you is that what I have been doing is not a possibility for technical reasons related to information, its delivery and its prossessing and intrepretation all in a timely manner. For what it is worth many others are in the same boat I am and I am sure that were you to examine their efforts, you would draw the same conclusions. Please stick with your guns on these matters. I have had many discussions with people about these kinds of things. Some are trying to learn to do better and some are trying to fault me. Those kinds of people learning do not deal with "the your full of shit" contentions. So FYI this is how I deal with market data and the timing of its processing. You ,as anyone is, are able and welcome to draw any contentions from my viewpoint. I feel 100 millisecond sweeps are sufficient in markets. I feel that at least 5 to 7 degrees of freedom are required to have adequate data. From that point on in about seven stages, which are sequential in nature, about 70 degrees of fredom result. There is no connection to 7 and 70, it just happens to be the case. The outputs, which are data for trading platform use are about 5 to 7 degrees of freedom. This involves about 100 TS's and each would be characterized as multi page in terms of development input. Just as hardware uses repeated circuitry, so does logic of managing information and creating derivatives of input information. As you read about how weird I am, you will discover that most people take my measure using a CW template that they have created. I do not feel the CW gets the job done so I have gone to a different world and its unique algorithmic variations. I do not fit under the scutiny of the CW template nor does my output of IP and TS. This advantage will not go on for much longer. A 50 cycle is as good as it gets for uniqueness. To not embrace probability took a proof. To turn to certainty took many proofs. Information theory, today and recently, broadcast these proofs. Occasionally, I have given citations. I have not been able to find or be found by like minded people. today, fortunately they are found throughout the spectrum of inquiry. So the easy filtering days are gone or at least numbered. What I mean is that being isolated was an easy filter and it is not there as much any more. There are many who use the "Cray" persuasion. I hope they persist for reasons of personal advantage. More and faster is your theme as well and I am full of shit since I don't go there and I get different results than you do. One eye opener of using certainty is how at different times what is going on in the market is different in importance. In driving it may be recognizable that different skills are used at different times. How probability is used is very divorced from how certainty is used. I don't measure users of probability by applying the template of certainty. Ideas from probablitiy based trading do not work very well in certainty based trading. One algorithm uses entries and exits the other uses reversals. In reversals an exit is synonomous with an entry, at that moment, in the opposite direction. This post may not reveal anything to you about how feeds and their arrival bear on certainty trading. Maybe it will. Price at 100 millisecond intervals can be seen to be quite "spread out" in time. On Qcharts, data comes in pulses every so often. depending upon the Q charts source chosen or automatically switched to by Qcharts feeders, the pulsing is more or less frequent and offset in time. Other feeds like tic chart feeds, DOM feeds can show how different the Q charts feed is. So could the DTM feed by satelite coming into twin 36 inch discs. Worden bros feeds long ago were limited to EOD data and to this day they do not provide FA data. What is the speed of indexarb.com? I don't care about that either. Drift indicators that I use are based on 10 millisecond sweeps. All the 100millisecond sweeps are supplied on a 10 millisecond basis in series/parallel. Gating is what stretches out the sweeping of 70 degrees of freedom. All these rep rates are faster than the market moves in trading with certainty. For probability trading, I would imagine they are insufficient with respect to probability mathematical requirements. As you see our posting was intened to build minds. New minds arrive and some things are repeated. For the Utube guy (S&P500) to get the SCT stuff he uses it had to come from subscribing to a programmer's blog who, in turn used the certainty rules of the journal. It is only a piece as we see and not certainty trading as a whole. Determining the cardinal elements of the certainty algorithm is binary. We do not always deal in opposites either. It is built as an onion and therefore timing is determined to be an advent over and over BUT no trading platform activity results. Coming into one of 20 to 40 actions a day (times 5 when partial fills are required because of market capacity limitations), is not as you describe when I use my algorithm. Certainty is always maintained and ACTION behaviorally (meaning the ATS sends signals to a trading platform) is not as you say dealing with a momentary spike of VERY LITTLE DURATION that isn't visible to the naked eye. Many tic formations are possible. a tic is a pair of values and NOT a price point. Two groups of traders deal with a tic. One group deals with one value and the other deals with the other value. the presence of group members is what determines the TIME FACTOR of a market. Cascading, easily measured, is an example of imbalance in the two tic groups. One group is eaten up rapidly by all of the electronic forces hitting the same spot, electronically speaking. As the cascade goes on, data is certainly poor interms of feed and arrival and processing. Fortunately, this is NOT an end effect of cascading. Only the moments after cascading are important in certainty trading. Binarily, cascading is either ON or OFF and certainly intermitent (and stretched out in time) occasionally. That was a long time on the T&S. There are gaps in seconds but who cares. What is important is to trade at the capacity of the market in partial fills doing each turn whether fast(at the end of cascading) or slow.
Nice evasion of a simple question. To the other jackolyte. I am not an expert. I am experienced. An experienced trader does not have to be expert to know malarky when he sees it.
I tested buying the 0 to 7 turn on 1000 of the S&P1500 stocks from 2000 to 2005 -- a total of 5000 stock-years -- using spydertrader's code for scoring. Here's the equity curve I got: http://www.elitetrader.com/vb/attachment.php?s=&postid=1278986
Here is the equity graph. It shows X trades and a Y loss. This is about 16 bucks a trade over each hold period when, supposedly, the cycle ends as 4 goes to 3. We know you didn't follow the cycle shown but what does the 16 buck loss per trade represent? Please tell everyone why you only used the 0 to 7 and NOT the 4 to 3.
Because your model is so BROKEN that the transitions from 4 to 3 are VASTLY outnumbered by the 0 to 7 transitions... enough to make the average trade last for YEARS. Which means that your diagram and concept of the price, volume relationship are at odds with reality.