We need to use some more 3x5 cards.
First I will explain what to do with the cards already made up.
In automation, a database is used. "Fields" represent specific data elements which make up the database.
The stack of 3 by 5 cards represents all the elements of the database.
just as Keynes defines paradigm theory, coding theorists define the "normalization" of fields in automation.
the pile of 3x5 cards is the RAW data. It includes the data from the market and all the added degrees of freedom of the system (About 70).
divide the piles into the First Normal Form.
divide each of the First Normal Form piles into the Second Normal Form.
divide each of the Second Normal Form piles into the Third Normal Form.
From this work you see how to interconnect things by using primary keys in each pile and "wiring" together the system of "tables" with commands which build the dynamic analysis table and how the analysis leads (See MADA) to the dynamic profit segment "table".
You see noise, anomalies, and flaws.
what is done ordinarily is ADD trader fields for risk and money management. They preserve capital mostly by eliminating the use of capital.
The alternative I used was to deal with these items by writing out more 3x5 cards.
Then I placed the cards appropriately.
Here is where the cards go:
1. finish the lateral set of elements system
2. Iteratively refine the gates and kills for the EE's.
3. Do not use RTL's within laterals.
4. when an AND of EE's occurs, then do not assign a P1 at the EE unless the prior EE is a FS and a FS is in the AND.
In the last few years, for every day of trading, the maximum length of a lateral is 36 bars.
It is very easy to conclude using Science and maths, that anyone who sees noise, flaws or anomalies is using an incomplete system. As we see here, the OP is going to create an incomplete system. He decided to compensate by using risk management and money management. This means he has a trading performance yield in a certain range.
One year at TradersExpo, Worden Bros was processing trading systems on their new drag and drop. They assembled the PVT quickly but could not use a Universe. They substituted the NAZ 100. The backtest performance had a Sharpe Ratio of 60 plus.
Obviously, there was something wrong. Worden immediately erased the work; but they did repeat their D and D work only to find they were exactly wrong again. This time they saved the system I gave them.
The PVT system is complete. when used with a good Universe it really makes money.
When you setup SCT using SQL you double capital every one to three days.
When automating, it is important to complete the work. then there is no programming for risk or money management required and all the capital can be put to work. All of this is done by paying attention to principles.
Last edited:
