Quote from Trader666:
Why is buying your "0 to 7 turn" (per page 8 of your paper on the P, V relationship, ATTACHED) no more effective as a buy signal than flipping a coin?
You may wish to look a little closer.
Get three coins to start of with; one for each of the three variables.
As you flip and record the scores you will see that the scores appear in a random order for you. Do as many flips of the est of three coins as you wish. You will find that there is no order to your appearance of the 8 possible scores.
Also check out the people who flip coins to prove things. Some, as you, flip one coin. Others flip sets of coins to deal with more complex problems. Thomas Bayes and the "frequentists" (two diffenrent groups of people) did what they called "probability" and they applied it to many different things where its application was warranted. It is interesting and part of information theory is in relation to what they logically determined.
I am not a user of their works.
In computer design, there is an aspect that is often used. It supports calculations. For those who have computer design experience, you may recal the bistable multivibrator. Gangs of these were used in computers. the genral subject was counting. For example to multiply, repeated addition may be done. n-1 additions of the multiplicand provide the product. The multiplier is counted down one step at a time and one less times (n-1).
As an IBM employee, I was surrounded by 1's and 0's and all kinds of logic for solving problems.
We did not flip coins at IBM. We did thinking. IBM imortalized two things: "THINK" and Kotex Blue.
I used "binary vectors" to define the PM of the HS. "ing" is a suffix used on words to express a binary vector. There is also a prefix. In and de were used for volume. A coin labelled in and de can be "ing'ed". the price coin is labelled continuing and changing on opposite sides. The third coin is labelled accumulating and distributing on opposite sides.
With three coins, you think that each is flipped the same number of times to get the randomness. It is also possible to flip the coins in a different way that you didn't think. If any way is chosen that is a way you didn't think, then the same random answer is achieved for having random scores.
As your back tests show, there is almost no trading in any backtest you do.
Lets say you flip the labelled coins until your p<0.05 or <0.005 or <0.0005.; one thing is for sure, you will not know the average value of the flips. This answer involves THINKING so it is not possible for you to gain the answer.
you have done 1000's of backtests and you have published one primarily and a couple of others as we know. Your primary posting is statisitcally insignificant in its posted results using any statisitcal test deployable.
The scoring is designed to describe the stock cycle using the variables of the cycle: P, V and A/D. Most people do not know what these variables describe in a market. The variables are direct and independent variables over time.
Traditionally people have seen the cycles of the markets and, for whatever reason, they have given them names. I associate scoring with those market names on the fractals those names are applied to. Scoring follows the convention of the ST, IT and LT nomenclature for market cycles. Coin flippers do not see the named cycles of the markets.
There are many many words used to describe all of the market's activities. Maybe you use these terms even if you are a coinflipper, who knows. Your posts have no substantive content.
Scoring is a tool useful in recording on logs as part of the trading strategy when using the paradigm and its applications to the different instruments traded in the markets. On the DAS it is the right most columns of the M section of MADA that goes across the chart.
You relate flipping one coin to trading stocks (you do not trade anything else you say). From this strategy of yours you get a statisitcally insignificant result through backtesting since it is akin in some way to you, to your results of back testing a creation you exptrapolated from a paper you fouind somewhere in that form posted (the original has an author and a publication date and is not indented).
So you ask a question. It is this: Is there any diference between two statisitically insignificant back tests? I an answer is the standard answer for this question that is rarely asked by anyone. The never will be any different answer to the comparison of tests (any number) which share the same statistically insignificant result.
Applying a perfectly random signal to any market will always get a statisitically insignificant result. This is a text book answer that is found in all text books.
You have proved that entering on any signal and exiting in five days, yields a statisitically insignificant result.
Ask yourself: What is the minumum dollar return Z for Y trades involving X capital that yields a minumum statistical significance. Obviously a table containing the three variables results. One axis is Y, the other is X and the cells are filled with Z's. When you look in that table your value 20 bucks is not there; it is too small to appear in the cells where X equals 500,000 and goes to 100,000 on all of those columns and where Y goes from 1 to 24,000 on all of those rows. These diagonal cells that are pertinent all have values greater than 20 bucks. Lousy back test to vary capital that way and vary the round lot size for given instuments. and there was little trading over the years. Buy hold five days and sideline for weeks. Not too swift at all.
