US Trend following system portfolio (Hypothetical)

[Sir Burr]

@peter2
3. Risk aversion and Market filter
-->I am not particularly risk averse. Happy to push boundaries (whether re number of trades/ capital etc.)
-->20% flipper with market filter is relatively risk averse system. When market trends lower you get out of the market very quickly!

3. Pick stocks trading close to 2 year highs. (this helps reduce DD without market filter)

Just curious if your using a market filter?
Flipper I use on the ASX pulls the stop tighter when the market filter goes negative and no buying.

 

[jjbinks]

Not using a market filter.
Initial code had market filter but I removed it as it did not improve profit or reduce DD.

I personally don't think tightening stops makes huge difference in long run. (i.e. 10% vs 20% stop loss as suggested by holy grails)
Using market filter to stop entries when market is down does improve results when i've played around with it but currently I only will stocks close to 2 year highs and when I do this market filter is not that helpful

 

[jjbinks]

Update on flipper
So

This is the current holdings.
As you can see only one sell (that was a mistake).
With 20% trailing stop stocks don't get sold very quickly.

One of my concerns about trading this system IRL was how to manage risk of currency fluctuation.
One way to hedge would be to use an ETF such as AUDS. (https://www.betashares.com.au/fund/strong-australian-dollar-fund/)
Curious to see what people think about this.

Overall the Russel was up 1.9% and system followed along.

 

[jjbinks]

russel was up another 1.6% this week.
Portfolio was ~0.3% down

 

[jjbinks]

Portfolio down this week by >1% compared to russel 300 which was up by >1%.
Had 3x sells with stocks hitting the 20% trailing stop

 

[jjbinks]

Up 1% this week compared to Russ300 which was down slightly.

Currently on 75% invested after sell off last week
My current trade filter is stricter than what I started with so having so not getting any buy signals recently. Could also be due to market down turn.
Will need to do a bit of back testing to sort this out. (Don't have the time right now)

 

[jjbinks]

Portfolio fairly flat whilst russel was down ~1% this week.

My US data subscription comes to an end in a few weeks and this little experiment will also probably come to an end.

At this stage I will probably not renew my subscription as I do not have time to back test and tweak strategies to make it worth while. I do feel the US market is definitely tradeable and perhaps I will return to it!

 

[peter2]

@jjbinks Thank you for making the time to post your research. I'm grateful for the glimpse of a system trading the US markets. There's a lot more opportunities in the US markets for all types of traders compared to the ASX.

@Skate 's Hybrid and CAM strategy would work just as well in the US markets. There'd be no problem filling 40 - 50 positions.

 

[Sir Burr]

View attachment 97705
At this stage I will probably not renew my subscription as I do not have time to back test and tweak strategies to make it worth while.

Hi jjbinks, can you expand thoughts on this?

Why does the system need further tweaks and backtesting? Why not leave that as a yearly thing?

I don't know about you but to me $700 is a fair chunk of cash for the data (Platinum) required for backtesting. It's good data but still a lot of cash. I've just taken out 6 months data as I have not had time to backtest Flipper in the past month. Planning to do it soon then drop back to a lesser subscription when it runs out.

 

[jjbinks]

Hi @Sir Burr ,

I took a 6 month subscription of platinum which as you point out is a fair bit of cash. It is great data and I'm sure I will probably look at it again.

RE: Tweaking.
I was not planning on tweaking the flipper code further.
Testing the flipper was really meant to be a step to then back tests more systems!
I would not trade the flipper because the results on Russell 3000 are not good over a longer testing period.

See: https://www.aussiestockforums.com/t...em-portfolio-hypothetical.34661/#post-1024112

Even with some volatility filters/price filters I CAGR is only about 5-8% depending on testing period. (So effort not justified IMO).

Would be keen to hear how your results are when you get around to it


Cheers
jjb

 

[jjbinks]

@jjbinks Thank you for making the time to post your research. I'm grateful for the glimpse of a system trading the US markets. There's a lot more opportunities in the US markets for all types of traders compared to the ASX.

@Skate 's Hybrid and CAM strategy would work just as well in the US markets. There'd be no problem filling 40 - 50 positions.

Although my trading experience has been relatively brief and IMO to brief to draw any conclusions. When I back test simple systems the over the last 10-20 years the return vs drawdown is always better when trading the Australian market.
This makes me wonder whether more opportunities are always a good thing. I definitely agree that it will suit some type of traders better but perhaps not all. I've come to the conclusion that you need to tailor your system to the market.
Although the same principles may apply it may work at different time frames (Both the signals you use to identify trades/aswell as trading frequency) and you may need slightly different filters to best suit the market.

 

[aus_trader]

Good effort in testing this strategy jjbinks, keep up the work in testing strategies that can outperform.

 

[Newt]

One other thought jjbinks about market noise, "maturity" and the suitability of trend following versus mean reversion....

There was an excellent Better System Trader podcost with Perry Kaufman which looked at the sorts of strategies that outperform in new markets, and those in older, mature, much more noisey markets. Essentially the US would be the ultimate example of a mean reverting, noisey, mature market where trend following is harder (definitely not impossible). Likewise it might be that MR in the US markets makes more sense the Aus.

http://bettersystemtrader.com/010-perry-kaufman/

 

[sir elements]

I was not planning on tweaking the flipper code further.
Testing the flipper was really meant to be a step to then back tests more systems!
I would not trade the flipper because the results on Russell 3000 are not good over a longer testing period.

I was looking at the 20% flipper strategy and was curious if anyone here had applied it (hence the search and finding an old post). I'm not sure if you were applying it correctly? When you bought some of the stocks they were already trending and shouldn't have been buy signals. The way it was applied here seems to be more like what Nick Radge talked about in Chapter 7 of Adaptive Analysis with 20% zones.

I have attached charts to show you what I mean. The arrows are where you bought, but the buy signal was well before the buy points.