95% ASX100 success rate

[RobertoHood]

Your 12 weeks history yielded only 23 signals, or <2 per week. This week and next week you have 6 signals per week.

Any reason as to why?

I have been thinking the same and have no idea why, the same patterns are showing up more often and Im not sure what it means.

Next week will tell the story i guess skc.

 

[systematic]

Is this the first time you've live tested it, after 4 years of development?

 

[RobertoHood]

Is this the first time you've live tested it, after 4 years of development?

Hey system,

I haven't been developing THIS particular system for 4 years, but more after 4 years of working through developing systems this is where i have ended up.

My starting System, when i first started trading has been posted in the "Potential Breakouts" thread. That was where i started, now as i evolved so did my system which is the one i am using today. (which when i stumbled upon it, it wasn't long before i started this thread due to my excitement!)

I have been trading it for the past 2 weeks.

Next week i will post some PnL, with confirmed Buys and Sells to prove i'm actually trading it.

 

[RobertoHood]

Alright guys here's what i have for next week;

AST
BHP
IAG
TWE

I have removed NVT and TAH as they are on the fringes of the BUY scale.
These 4 sit firmly in BUY range

 

[RobertoHood]

Your 12 weeks history yielded only 23 signals, or <2 per week. This week and next week you have 6 signals per week.

Any reason as to why?

To add to this question skc , as i now have a clearer head

What this system MEASURES makes sense logically if i were to show you... But where the BUY signals happen and the where and why the patterns form make no sense at all, they actually seem counter intuitive.

And the way this System come about , was because i was looking to measure certain components to add to my trading strategy and the way i had my spreadsheet set up , i was clearly able to see a very strange pattern.... To the point where i got my girl friend to go through the spreadsheet (she knows nothing about stocks) and by telling her the pattern to look for , and with out seeing the next weeks %Gain , she was able to go through and say which stocks would end the week up and which wouldn't..

All worked BAR, ASX.AX Jun 1 - Jun 5 which ended the week -2.28

 

[RobertoHood]

Hey guys,

Spent most of the weekend rebuilding my system , i found a number of errors and decided to rebuild from scratch.

TWE , WPL , BHP

These are the stock picks for this week

 

[tech/a]

Did it return similar figures to those shown by version (1).

Interested to know how and what you did in the testing process to
be able to be confident that this version is More accurate.

What do you look for when evaluating the robustness of a method.
What "figures" are important?
Your only using Excel?
Still with only 3 mths of data?

 

[RobertoHood]

Hey guys,

Here are this mornings BUYS, i was a bit slow off the mark

BHP 630 @ $25.510

TWE 2870 @ $5.596

WPL 485 @ $32.64

 

[RobertoHood]

Did it return similar figures to those shown by version (1).

Interested to know how and what you did in the testing process to
be able to be confident that this version is More accurate.

What do you look for when evaluating the robustness of a method.
What "figures" are important?
Your only using Excel?
Still with only 3 mths of data?

Hey tech,

I had created a template , which i had planned to use to test other markets but because i had copied/pasted the template to blank worksheets, some of the formulas were take data and calculating from the wrong cells.

Only made a slight difference from my previous version in some stocks, and huge differences to others.

But also squashed alot of the BUY signals i had previously got for this week, once i ran the data through the new sheet.

 

[CanOz]

RH, do you need more data to test with?

 

[RobertoHood]

RH, do you need more data to test with?

Yes , i am going to start running data from other markets through the template and see if i can find anything interesting

 

[CanOz]

Yes , i am going to start running data from other markets through the template and see if i can find anything interesting

Perhaps i missed something prior, but i was wondering if you wanted to test back further than three months on your current universe but you didn't have the data?

 

[RobertoHood]

Perhaps i missed something prior, but i was wondering if you wanted to test back further than three months on your current universe but you didn't have the data?

I have used 3 months as my own personal general rule CanOz (which ties in with the system), i could definitely get the data if i wanted to do so.

 

[CanOz]

I have used 3 months as my own personal general rule CanOz (which ties in with the system), i could definitely get the data if i wanted to do so.

Sorry, i was under the impression you didn't have the data going back further than 3 months. reading though the thread you really havent addressed the reason why you've back tested on so little data? From my POV a walk forward (what you are doing now?) would only occur after:

1.) a successful back test on a minimum or 10 years of EOD data
2.) a successful backtest on an additional 5 years of OOS data

At this point, you would have the confidence in your system to walk forward at least 1 year.

I'd like to see Howard jump in here, but i think even my requirements above are likely too short. I know one system trader that everyone else knows and i think he tests his EOD systems on 30 years.

So i guess i am just trying to understand why you are not following the accepted test regime for EOD systems? What is the point to move directly to a (live) walk forward?

Again, sorry if i have missed this answer.

CanOz

 

[skc]

Hey guys,

Here are this mornings BUYS, i was a bit slow off the mark

BHP 630 @ $25.510

TWE 2870 @ $5.596

WPL 485 @ $32.64

FYI, both WPL and TWE are due to report on Wednesday 19 Aug.

Unless your strategy has been designed to take into account of potential wild swings from reporting... might be wise to step aside.

BHP reports next week on 25 Aug so I guess the position would be closed by then.

 

[Wysiwyg]

A backtest report would be handy to show these purported outcomes? Something like the following --- Only 52% right though!

 

[RobertoHood]

A backtest report would be handy to show these purported outcomes? Something like the following --- Only 52% right though!

View attachment 63886

View attachment 63887

Hmm Interesting Wysiwyg...

 

[Wysiwyg]

Hmm Interesting Wysiwyg...

I could not stomach a 45% drawdown and back testing over long periods should be done on stock lists including de-listings, mergers etc. otherwise the test list is bias toward surviving.

 

[RobertoHood]

I could not stomach a 45% drawdown and back testing over long periods should be done on stock lists including de-listings, mergers etc. otherwise the test list is bias toward surviving.

That back test survived the tech bust and the 08 recession!

Looking at total return from an initial $10,000 investment , i know which of the two i would want to place my money.

 

[Wysiwyg]

That back test survived the tech bust and the 08 recession!

Looking at total return from an initial $10,000 investment , i know which of the two i would want to place my money.

Oh I can get a fantastic back test result too.