Index filter for Rotation/Momentum

This topic is empty.

Viewing 12 posts - 1 through 12 (of 12 total)

Author

Posts
May 14, 2021 at 6:14 am #102106

DeanWalton
Member

This is my first post from a newbie still doing course so a little nervous in posting. I am currently working on my ASX Rotation/Momentum strategy for ASX100.

The returns are not what i was hoping for but still ok – about 16% return and 25% MaxDD. MaxDD is a little high and the CAGR is a little low. I’m rotating every 2nd week as returns slightly better than weekly.

I want to do a weekly or every 2nd week as the monthly is a little nerve racking for me if everything goes pear-shaped quickly at the beginning of the month like Mar 2020!

At the minute i am using a 75EMA on XAO as my index filter. This tests best so far. It is my understanding that Nick does not use a MA for his index filter and his returns are far superior to mine.

So i was wondering if anyone could lead me to an area to explore as alternative to MA filter. I have looked at ROC > n days ago, RSI > n days ago, MACD, Directional Movement Index (pdi), ROC > 0 on XAO and other indexes, Cumulative 4W highsLows and ROC on it.

I’m going slightly mental and quite frustrated as i am probably missing something very obvious.

May 14, 2021 at 6:35 am #113275

JulianCohen
Participant

Hi Dean,

First thing I would say is that 16% and -25% are not out of the ball park by any means. Especially for the ASX…depends on the time frame of course but if that is 10 to 15 years then you are not doing badly.

Personally I have always used MA on the Index as a filter for trend following purposes. I’m not sure what Nick is using on the ASX but I’m pretty sure that’s what he uses on the US markets, so it is not a bad solution. Have you looked at running a monthly as well as a 2 weekly to see what the cumulative returns are? That could be a solution that looks outside the box you are currently searching in.

As for other indicators, there are a few….try looking at the ATR of the Index, if volatility is too high in the Index then exit. I also look at Stockcharts and go through all the indicators they have listed there, testing everything I can find. Don’t always try them in they way they were intended to be used, in other words try something different. It could give surprising results. I just read a Cesar Alvarez blog where he used ROC * HV

I tested it and it didn’t work on my system but that doesn’t mean it won’t work on yours.

Try changing the universe….Try the ASX 100 only, or the ASX 300 without the ASX50, or any other combinations you can think of.

Hope that gives you something to work with

May 14, 2021 at 8:17 am #113279

GlenPeake
Participant

Hi Dean,

Welcome aboard!!

From my experience, when I started testing a monthly rotational on the ASX100, my stats looked pretty similar to yours to begin with, so you’re in the ‘ballpark’ etc….. I’d suggest taking a look at a Dual MA on the XAO.au for your Index Filter (i.e. a slow/fast moving average crossover filter etc)…. This is what i went with and it really smoothed out the equity curve / MDD. It might work good for your specific system. You could also test the Index Filter against the ASX100 symbol XTO.au to see if that makes a difference….

In terms of the weekly/fortnightly system, you could consider running a monthly system in 2 tranches…. i.e. divide your capital into 2 parts i.e. 50% split and run 50% at the start of the month and the other 50% as a weekly (or fortnightly etc)…. Have a read of Nick’s ‘Signal Luck’ article for further details here:
https://www.thechartist.com.au/diluting-signal-luck/

If you’re not 100% sold on a monthly strategy also have a read of Nick’s Momo article where he compares various weekly/monthly strategies etc
https://www.thechartist.com.au/momentum-with-52-week-high/

Other links FYR:

My Starting stats:
https://edu.thechartist.com.au/kunena/progress-journal/413-glen-s-journal.html?start=100#8889

Dual MA Index Filter comparison:
https://edu.thechartist.com.au/kunena/progress-journal/413-glen-s-journal.html?start=100#8992

Rotational System Resources:
https://edu.thechartist.com.au/kunena/progress-journal/413-glen-s-journal.html?start=125#9033

Cheers
Glen

May 15, 2021 at 12:25 am #113281

DeanWalton
Member

Thanks Glen. I’ve read a lot of those links already. So am i correct here – the dual moving average when you say “searched” this is done via trial and error and optimisation.

The issue i have had with dual moving average filter was the slowness to get you back in. But maybe i was using too slow.

Can you optimize two parameters like dual moving average at once? I have only ever done one.

May 15, 2021 at 12:34 am #113280

DeanWalton
Member

Thanks Julian. I have not read this blog about ROC*HV – will try it.

I came across to the mentor program heavily using something called SuperTrend which i was using to trade in Bullcharts (software package). This is an ATR which sits above and below price/index letting you know when to be and out of the market. All the backtests we have done so far has proven to be less then good. It looks so good on my old charts I can’t understand why it doesn’t work when i backtest it….but as Nick keeps telling me – that is why i am doing the course to find what works and what doesn’t.

I have experimented with different indexes but will keep going. Maybe my expectations and reality need a check.

May 15, 2021 at 3:13 am #113282

GlenPeake
Participant

Dean Walton wrote:

Thanks the dual moving average when you say “searched” this is done via trial and error and optimisation.

Can you optimize two parameters like dual moving average at once? I have only ever done one.

Yes to both questions Dean.

It could be that the Dual MA Index Filter works well on my Monthly system, however when applied to a weekly or fortnightly, maybe not so good….possibly depends on the other parameters for the specific system I guess etc….

May 15, 2021 at 4:51 am #113283

JulianCohen
Participant

Dean Walton wrote:

I came across to the mentor program heavily using something called SuperTrend which i was using to trade in Bullcharts (software package). This is an ATR which sits above and below price/index letting you know when to be and out of the market. All the backtests we have done so far has proven to be less then good. It looks so good on my old charts I can’t understand why it doesn’t work when i backtest it….but as Nick keeps telling me – that is why i am doing the course to find what works and what doesn’t.

Funnily enough I saw the SuperTrend system on the RT forum the other day. I tested it and discarded it straight away. I always test everything I see on different inter web sites….you never know what might have the secret ingredient to turn lead into gold

Seriously though, doing this course teaches you how to test correctly so you can discern the wheat from the chaff. The Supertrend system looks good at first glance, seems like a sensible idea too. But it just doesn’t work well enough compared to other ideas….you already have one that is probably twice as good as that Supertrend system and you are not happy with it yet as you are sure it can be improved upon…that’s worth stepping back and thinking about.

May 20, 2021 at 9:33 pm #113284

OliverBenson
Member

Hi Gents,

I read the same Cesar Alvarez post about ranking. I have found in testing my MOC short system that Rank = 100+(ROC*HV) tests really well (better than 100+ROC or 100+HV). My long MOC ranks by 100+HV(20) – also plagiarised from Cesar Alvarez.

I’ll have to test 100+(ROC*HV) on my live long system now (although I feel like I shouldn’t really be messing with it seeing as I’ve only just started trading it…).

In general I have found that the ranking method has a huge impact, probably more than anything else. But my experience is very short lived and I have only ever looked at MOC systems.

May 21, 2021 at 5:57 am #113291

JulianCohen
Participant

Welcome to the rabbit hole…leave your sanity at the entrance.

Nah only joking….constant testing is part of the fun…for example I am testing your ROC*HV idea again…just in case I missed something

May 25, 2021 at 3:44 am #113276

MichaelRodwell
Member

Have you got dividends turned on?

June 8, 2021 at 9:08 pm #113277

Anonymous
Inactive

I think those returns are quite acceptable for a low-touch rotational system with minimal optimization

I’ve explored lots of different index filters. Index MAs are probably the easiest and most logical form, though I’ve found that breadth works quite well too for my systems. You can explore things like:

1. XXWeek Highs vs Lows
2. % of stocks above/below their respective moving averages
3. Advance/Decline ratios
4. Other oscillators

Amibroker has a great deal of extra data for the indices which you can play around with; they just involve a lot of trial and error.

Probably worth noting that there is some timing luck & likely to be some curve fitting with a monthly/bi-monthly strategy, so for the sake of a 1-2% extra returns it’s not the be-all and end-all

June 10, 2021 at 8:04 am #113278

DeanWalton
Member

I ended up using a SuperTrend indicator for my macro filter. Thanks for everyone’s input. It works for me in testing – now to do it for real!
Author

Posts