No use trying to reason with Jack and his cabana boy Spytrader.. they are like Woodie and his CCI.. they are pushing a religion for their own fame.. the fact is does not work doesn't deter either of them
Moving out of the probabalistic arena is not arcane. How many degrees of freedom do you advocate for maintaining a data base. If you are able, please reflect on this as a general question since we use different methods. I am oriented to channels and their information content for price. Also I have taken probability off the table in the data base I maintain. Price is only one of three legs (for stability) of my "stool". You are using an interesting handle. For me, the data offered by the markets is more than sufficient to eliminate any unknowns. I feel that data provided at this time precludes the use of probabilities. So the question of boiling tripe is not on the table for me. I feel non probabalistic data input, processing and a resultant data base provides "an accurate reflection of the critical variables associated with the problem." Optimized trading is "canned" for me. By dealing only in NOW, relatively speaking, for any number of degrees of freedom, a finite set of possible results emerges. This table is canned for me. The connection of the data sets to everything else is also canned and is a serial process involving finite sets where, certainly, I could eliminate intervening steps to get to always be trading in exact synch with the market more simply. The reason that I do not skip steps is that I have a requirement to update break points as an iterative refinement process. This has to do with the capacity of any given market. Five or six times you have raised a question you are coping with how to present to readers. Raw data (having degrees of freedom) is continually delivered for processing continually. The results are in a data base and are used for many purposes by processing. At any time there is a status with regard to the corresponding decision being made (In my case automatically). For me 20 to 40 actions per day result. they correspond directly to the change of market direction at the time of the market direction change. All other times are more important since they provide for the "harvesting of capital" by being in a "harvest" orientation. Money is made and the accumulation has a velocity. The mechanics of periodically "banking" the harvest are done when there is no price movement. 20 to 40 times a day the harvesting slows to 0 and a banking occurs and the harvesting begins again. The data sets for these activities fall into three interrelated legs. The leg arrangement is dictated by the market and the critical named variables. As has been pointed out to readers, there are many many ways to process data into useful stuff. the tripe category is not used by those who are making a lot of money. There are many ways to make money and, for me, I feel optimizing taking all that is available is the reasoned approach. It does not, therefore involve probabilities. There is an implication about any strategy that uses probabilities. It implication is that something is unknown and it is being dealt with by using a less than optimum approach. A great deal has been written about this by people who have very strong views of markets and their operations. My view is that at all times the programming of the data has to be fully adequate to "know that it knows". This is out of the consideration of a kind of system that deals with uncertainty and the extent of uncertainty. An uncertainty system, to me is, prima facia, too risky. The tripe issue is not part of uncertainty or the "know that it knows" real time operating systems. A systemmic approach provides a structure, operations within a structure and results of the operations. The variously colored box approach. Market data is all that can be used to input the system. within the system any number of degrees is possible but not necesarily useful. Dealing with only the useful is what is required. this is the "contraction" function of info gap theory as we all know. Non tripe sufficiency is the definition of utility in trading the markets automatically and continually. the box needs to use sufficient degrees of freedom at all time to not have uncertainty and to "know that it knows". Trading markets is done using continuous inputs and the system results are either used or not used as part of the next iteration. You say in several posts one way or another: "That the set of initial conditions which one uses to formulate the problem is an accurate reflection of the critical variables associated with the problem." "conditions" is a big assortment of items for you. For me the system is the critical condition. We have differing systems. There is a large chance that you judge other's stuff using your stuff as the standard. A better way would be to use the application (the market) as the standard. The market is there. It offers. The problem is formulated, simply, to obtain the offer. Trading involves being part of the market and not affecting it if you want to have a high money velocity. The sacrifice needed to optimize the money velocity does mean that you will be having an effect once you activity becomes significant. Testing that is a priori also. As things settle out, screen displays of direct market data become unnecessary and only a given range of degrees of freedom need to be displayed at a given time. The market determine what and when. It is all non tripe. what traps people in the conventional orthodoxy it that as they approach the markets they do not know what is going on. as time and knowledge, skills and experience all advance, the unknown theme still prevails and by then it is a case of the safety in the familiar and "enough is enough". Making 2 or 3 million on 6,000,000 trades is enough and apparently it is a kind of human limit kind of thing when operating in the reaction modus of the fighter pilot. For me the grass limits how many sheep I can run, instead. I am going the pastoral route. So you have figured out that applying statisitcs is getting the results that are found all over the place. As you look at this, you do see something. We both see that the markets are huge and they offer almost beyond the imagination. You may be concluding that there is something wrong with the picture. Here is a snapshot. Three times this year a trader has pulled 100,000 in a day using 30 contracts as margin. (66 points a contract) "That the set of initial conditions which one uses to formulate the problem is an accurate reflection of the critical variables associated with the problem." The problem is to see what the market offers and then have a system that ectracts the offered.
Mr. Hershey, While I am still mulling over your comments from yesterday and as yet am not ready to ask you any particular questions let me respond to your most recent post (5-21-07 11:11 AM). My "boiled tripe" post was directed to "PointOne" and attempted to make the point that no matter how pretty the findings of a statistical "experiment" happened to be, if the initial conditions were found to be wanting then the result was meaningless. Nothing more or less. As to your question of how many degrees of freedom are needed to accurately define the market condition, I quite frankly don't know but would think that there must be at least three, e.g., time, price and volume. You are quite correct that we use different methods but it may well be more a question of how one skins the cat as opposed to whether or not we are dealing with comparable cats. While I do not front-run the market per se, I do try very hard to stay out of the chop and that is what my method is primarily about. In a sense I guess it is front-running because no trader wants to enter the market at the spot where the big guns have decided to turn on the Waring blender. lj
Thanks so much for your response and your comments on where you operate. I'm getting a lot more data than you are. I didn't realize that from your posts. I certainly agree with you that sidelining when the risk goes way up is a very good idea under probabalistic conditions. When I look at the progression of a trader through the various levels of knowledge, skills and experience, the logic, circuitry and coding of the various stages does not vary greatly in structure. I use a paradigm that keeps all traders in the market all of the time with one exception at the lower levels; I find risk to be highest (for the trader because of lack of knowledge, skills and experience) in a much different place than you. That is to be expected. There is another thread involved with work (having a job type of thing) and a choice is put on the table regarding being a developer or doing what leads to doing development (modelling). The discussion doesn't come around to topics related to making money to any extent. There was another thread about coding a while back and I learned that "snippets" were used as building blocks to take models to codes and that was development of a way to make money. Cat skinning must be the result of getting the software built. Being fearful of ensuing chop is an interesting place. It may be that the developer people haven't done too well on the snippet business in that place. Before getting to chop there are loads of internal formation Q's that come along. I did all the snippet work and looking back here and there, I notice that most people learning to skin cats in various ways are getting caught in information overloads and such. On the other hand people who are paid money to model and those who develop models thrive on information overload and have conflicting signals all over the place. Both results seem to define "chop". The two best articulated screw ups on this come from a couple of buddies ET buddies not buddies of mine) who use differing approached and get screwed the same way in chop. One gets conflict and the other shuts down unknowingly. You better double your degrees of freedom as a starting point.
Thank you Mr. Hershey for your comments and in particular; "You better double your degrees of freedom as a starting point." Indeed I do make use of more than three degrees of freedom but at this point I have not settled on an exact number. One simple way of looking at how these 3 variables interact is to view them as tensors with each affecting how the other performs, not unlike the fabled 3 body problem in quantum mechanical simulations. Which is to say that what ends up being required is an approximate solution (a first guess if you will) and then taking that initial result and iterating it. The process is not statistical and relies on how one's guess performs in the market, i.e., do I reproducibly make money with this guess. That is why for what I do, I place such importance on having reliable data. I do not find it useful to try and figure out how the market movers think but rather focus on what it is that they do. My methodology is very much in the "now" time frame. lj
I'm touched and amused to read this. (I think you meant to say you were addressing trefoil's post above mine.)
liyoung, et alia: a few things. All of you need to read <a href="http://www.bartleby.com/141/" target="?">Strunk's</a>. You all take 10 words, big ones at that, to say what you mean, when three or four will do. Secondly, this is for the point, if you don't know what a frequency distribution is, or a regression, you should take Stats 101. I don't know how to be any clearer without revealing exactly what I do, and I ain't doing that for free for you guys. Thirdly, point whatever, I suppose you should be proud of knowing what a noun is. I'm happy for you. Me, I turned down a fellowship from the University of Chicago in Linguistics for no other reason than that I was tired of school. Which just means I know my <a href="http://en.wikipedia.org/wiki/Declension" target=";">declensions too.</a> My point, point whatever, is the same as Occam's. Look it up.