After spending nearly 10 years on this subject the day came learned only one single logic that produces results.I think because i knew enough that in some strange way i was ready for it.This is what i think now,back then there was a lot of frustrations.
And that's mainly because if everybody goes one way i go the other(not necessarily opposite).I understand short comings of this and I do like to fit in and "be one of the guys"( i learned from living in USA),but in solving problems i have this strong tendency to look at solution in unique ways.
Yes ,around that i build trading method,but that wasn't easy by any means.
Only because logic was solid i proceeded forward.
First test did not look so great. I piled up signals and all of them were in a draw down and if i only rely on short term backtests i would not put any value to it and possibly throw it away.Only understanding that math behind was solid and logic was sound i started working on putting together a method.Basically it came from understanding that i if somebody told me better logic i still would come back to my own.In there was an answer to some difficult questions,it felt like it was mine and it was good.
This is why i view that from backtesting alone it is very difficult to see something meaningful even if it was there.The believe in yourself has to be borderline crazy.Yes the results of the first test aren't great,but no,this will not shake me down.It was time to work on it
There was no string them together,just building a method to the a single logic
kind regards
-----------------------------------------------------------------------------------------------
I started trading options the first trading day of January 1998.
I learned to create programmed mathematical mechanical trading systems. After several years I discovered the key to success for each system, was "forward testing each system for 1 year minimum."
What I mean by this is: Lets say your system generates a trades signal for tomorrow,
including the exact BL (Buy Limit) Price, SL (Sell Limit) Price & STP (Stop) Price.
Computer Output Example:
Signal Date: 01-23-2018
SYSTEM: M25
CALL=40-66-80=L5 PUT=10-10-10=L2 CODE: UA5W2=WP-2 RELIABILITY 6/7 (86%)
OUTPUT:
DIA FEB258P B1=1.35 SL=1.75 / B2=1.10 SL=1.55 STP=0.90
If I was simply forward testing the above signal, I would have to follow it tomorrow
with a Real Time Live 1 minute chart to make sure B1=1.35 occurred.
If it did occur, I would then bring up a Time and Sales Report to make sure an Ask Price of 1.35 was available.
Then after seeing 1.35 Ask on the T & S Report, I am now watching for SL=1.75 occur on the chart and confirm it with the T & S Report.
If the option tanks after B1 (BUY 1)=1.35, Then I have to watch for B2 (BUY 2)=1.10 with the Option Charts and T & S Report.
After the B2=1.10 Ask occurs on the Option Chart & T & S Report, again the same routine of
watching the Option Chart & T & S Report for either the new Sell Limit of 1.55 or the
Stop of 0.90. This can go on day after day for a Time Limit of 4 days.
Its time consuming and a pain to test that way, but in the couple of decades I have been doing this, its proven to be 99% accurate.
On the other hand I have found Back Testing to be about 70% to 75% accurate.
That means a Back Tested system generating a W/L of 90%, is actually about 63% to 67% accurate. There is simply too many unknowns in Back Tested data!