US Trend following system portfolio (Hypothetical)

[jjbinks]

Thanks for the kind words of encouragement guys!

And peter2 you are spot on regarding volatility and many losing trades.
I already filter stocks based on turnover (minimum $500,000 turnoover or 500,000 stocks/day) this is based on Nick Radge's Holy grails book.

Currently I am trying to compare my system to what is described in Holy Grails. I am under the impression the code in the book does not use a market volatility filter or filter stocks by ATR.
Important to note that Flipper system in Holy Grails without a market filter had 950+ trades from 1997-2011 and 40% win rate. The max DD for the unfiltered system was only 35%.

Currently when i retest this I have a 35% win rate and about 1000 trades. When the market takes a turn for the worse or becomes more volatile the losers accumulate very quickly like you said.

As I have mentioned I could move away from the flipper strategy to alternative system which I have confirmed decent backtest results for but I am enjoying the challenge of trying to code the flipper strategy as described by Radge.

Also keep the suggestions coming. I hope to one day move on from the flipper challenge and attempt something new!

 

[rnr]

Hi jjbinks,

"The challenge is good fun isn't it, frustrating at times but still fun."
Hang in there because the fun part probably has a way to go.

"I already filter stocks based on turnover (minimum $500,000 turnoover or 500,000 stocks/day) this is based on Nick Radge's Holy grails book."
The above statement requires a slight adjustment to read based on turnover (minimum $500,000 turnoover AND 500,000 stocks/day).

The above adjustment above might explain some of the variations between the book and your results to date.
If you don't mind me asking:
1) are you including delisted stocks,
2) are you referencing the ZigZag function in relation to the 20% move in either direction?

Cheers,
Rob
(Following with great interest.)

 

[jjbinks]

Hi @rnr

Thanks for suggestions.

1) i have tried 'AND' and 'OR' to filter based on turnover/volume.
To me AND makes more sense but the book to me suggests they use OR so i test with both options.

2)The current stock universe for xao including delisted is 2019. Nick had 1847 stock universe.
Data gets updated and innacuracies get fixed so i dont expect to match radges results exactly but should be close.

I have tested some of the other simpler ideas in the book and get results reasonably close to what radge published so feel my stock universe is not too different to what radge used.

3)not using zig zag

 

[rnr]

@jjbinks

Thanks for the heads up.

Just looking for areas that may be of concern.

Just out of interest are you aware that Richard Dale offers 2 alternatives for Amibroker data used in back-testing?
As well as Premium Data (PDU) there is also Norgate Data (NDU) which was released approximately 12 months ago.

Links to ASF post and Norgate Data

Cheers,
Rob

 

[jjbinks]

Despite the extra week of tweaking. I have not managed to replicated Radge's flipper results.

For the time being I have made some tweaks to filter based on volatility (as per P2's suggestion )

I will start trading Monday. This will be a weekly system.
Starting capital will be $50,000 USD.
Position size $2,500.
Maximum of 4 new positions per/week. So would take 5 weeks to be fully invested if market conditions are favourable.

See example trade below



Green line is the 20% flipper buy line and red is stop loss.

So buys for Monday are
HRS
RNR
MTCH
LAWS

I am using IB paper trading to hopefully make trailing stops more realistic.

Not going to worry about currency at this stage.
As always suggestions/feedback welcome

 

[jjbinks]

No new trades this week.
With 4 trades on Monday portfolio is 20% invested.
Small movements in price in selected stocks have cancelled out so net change in porfolio size is essentially zero!

Russell 3000 was down >1% this week.

Current portfolio

 

[rnr]

Hi @jjbinks

I can see that the second stock in your list has heaps of potential, so obviously a wise choice.

Cheers,
Rob

 

[jjbinks]

Hi all,

@rnr RNR is my best performing stock so far!

Russell 3000 has been trending down last 2 weeks. (Down ~3.5%) . Limiting my buys to only 4 new buys per week has limited losses so far (of course if market was up 5% perhaps I would be wishing i was invested sooner but I prefer the "gentler start").




Will be buying 4 more stocks come Monday.
SEAS
HEI
GWR
CHCT

 

[qldfrog]

Much appreciated this thread jjbinks
And like your 4 per week max buy
Paper trading so far?
Quick question
When you move real $, who do you plan to use for aud/USD exchange and brokerage?
I have a small us portfolio with comsec/pershing but there is no way i would actively trade with their brokerage fees..or even exchange rates

 

[jjbinks]

Good question @qldfrog
I have limited experience trading foreign stocks and others may be able to answer this question better than I can.

Interactive brokers offers low brokerage for US stocks. min USD $1 or 1%

Re currency exchange IB forex is also relatively cheap
0.20basis points or min $2. I believe they have good spreads.
So I think if you apply for a forex account you could transfer say equiv of $100,000 usd for just $2.

I have not tested this and would love if more knowledgeable people could confirm or refute this!

 

[qldfrog]

I thought there were problem to open new accounts from Australia with IB but I might be confused
Interested as well
The bigger size market would make systems more interesting in my opinion
We are really plating in a small world here with the ASX

 

[jjbinks]

4 new buys this week.
MSI
STAR
SAIC
APLS

Porfolio essentially is flat. No suprise give it is only 35% invested and only 2 weeks in.
I stuffed up stop loss for CHCT last week so essential bough and sold same day for a few dollars loss .

Need to be more careful with entering orders!!

Will post more detailed update next week.

 

[jjbinks]

Hi guys,

Another weekly update.

Current portfolio


Total porfolio -0.5%.
Russel 3000 Index +0.8%

A few things to note. Most of the trades are priced above $20 and many above $100. I do wonder if buy smaller cap /lower priced stocks will offer better returns. Will test this!


@rnr continues to be the top performer.

 

[peter2]

Good start considering the tech sell off and that the small caps are on the nose at the moment (ie the Russell index is weaker than the others).

You've got me thinking about the system. I've reviewed the charts of your selections so far. May I offer you some suggestions that could improve your potential profits? I would expect you to test my suggestions to see if they are, in fact, better.

 

[peter2]

Before I continue, with your permission of course.
I'd like to know your definition of the low from which your flipper buy trigger is calculated on.
Is it a yearly low or a six monthly low or something innovative like the lowest low after a yearly high?

This low is pretty significant in your system.

 

[willoneau]

Hi jjbinks, I am interested in your ranking system. Is it much like nicks "bang for buck" or something else?

 

[jjbinks]

@peter2
Suggestions are always welcome. And of course I will back test.

The low is actually any low since the last buy signal. So this could be a week ago or in some cases may years ago.

@willoneau
my ranking system is uses an RSI. To be honest it doesn't add much to return.

p2 and your questions have got me thinking that perhaps I could use time since last low for my ranking?

Will report back!

 

[rnr]

I've posted a daily chart of FMG so that those members who only have access to ASX data can relate this pic to their charts.
Signals (Daily on this chart) are Buy on next bar after blue line and Sell on next bar after red line.
The green line is a 20% high-to-low zigzag.

@jjbinks
I'm not trying to derail your thread with this post.

Cheers,
Rob

 

[jjbinks]

Hi @rnr

I don't think you or anyone else needs to worry about "derailing the thread".
All comments are welcome. Often there is something for me to learn and hopefully others.


The zigzag indicator is forward looking when used for back testing. (you don't know what a top is until you have passed it)
So the flipper code does not use it.

This is a chart with zig zag (in blue) and my "20% trailing buy and sell lines" in green and red respectively. Arrows show sell/buy signals




The buy signal is 20% rise from any low since last sell signal or buy signal.

 

[rnr]

Hi @jjbinks

My apologies for some missing info on my previous post (@#38).
Using the HiLoZigZag set at 20% the Blue line signals when a 20% rise is locked in and the Red line signals when a 20% fall is locked in.

I totally agree with your comment that "The zigzag indicator is forward looking when used for back testing" when used on its own and that is why you need additional code to create the signals as referred to above.

Cheers,
Rob