Rotational Systems

[ScottMcNab]

Is it possible to move relevant posts in “graveyard” and “Saids ranking method” so all under the heading of Rotational systems Nick ? I have found info very helpful and may make it easier for future members to find ?

[ScottMcNab]

I have been inspired by the posts regarding rotational systems. I am hoping to learn more regarding monthly rotational type systems.

Are backtests on monthly rotational systems done using just 12 time points a year ? This would mean that 20 years of backtesting still only equates to about a year of backtesting data points used with a daily system ? My daily systems can look great for 5 years and then go to hell in the 6th !

I changed my rotational system to weekly to look into this further and was surprised to find that there was not very much change..they seemed very robust….for example on NDX (from 2000 to present) one system had CAR/MDD of 14/18 for monthly rotation and 12/14 for weekly (I am constructing something for smsf so not swinging for the fence). So my next question is whether trading maybe a weekly version on the NDX and a monthly version on SPX and/or XAO seemed crazy ? Combined with a MOC system, this would provide a nice diversity. Is there something I am unaware of regarding the tendency for these systems to be monthly or is it that they are just looking to capture longer type trends?

Thanks in advance

[ScottMcNab]

Anyone using Monte Carlo with rotational systems ? Assuming not trading these on margin ..I was getting pretty pleased with my systems until I looked at the MC…had to reduce the postionsize with corresponding moderation of CAGR

[TrentRothall]

I believe because you are always taknig the top ranked stocks MCS aren’t as necessary. But probably good to know what would happen if the past was slightly different??

[Nick Radge]

Quote:

Anyone using Monte Carlo with rotational systems ?

There is zero selection bias so trade skipping MCS is not required.

However, to test robustness, you could do as Trent suggested and do some stress testing using parameters and price variances.

[JulianCohen]

Out of interest how many people are rebalancing the positions that you are not rotating out of at the end of each month?

I know we haven’t tested for this when running the backtests but I just thought I’d ask if anyone has tested it, if they know how as I assume it means using the custom backtester.

[Nick Radge]

I’ve only ever done it once since 2015.

[ScottMcNab]

Rotational systems seem reasonably robust to parameter changes…3 biggest influences I have found on metrics of the backtest are:
1. index selected for index filter (if used)
2. position sizing
3.start date (go 4 years before 2000 or 4 years after and you will feel much better on the US systems)

[ScottMcNab]

This site has some interesting ideas with rotational systems…for example they use a shorter period for their ATR calculations than they do for the ROC

Monthly Sector Rotation

[RobGiles]

When backtesting a rotational system, I should get the same performance metrics for a given date range every time I run the back test right (as there’s no SB)?

[TrentRothall]

Should be the same trades everytime, yes

[RobGiles]

Thanks Trent

[RobGiles]

Nick
Going over my notes from our various hook-ups prompted the following question, why do you continue to trade the ASX growth portfolio when it has the following inherent disadvantages:
1) Selection Bias
2) Trend systems need a trigger to generate an entry and there are times when stocks are trending up but don’t close above the trigger, hence you miss the move. A Momentum system eliminates both those issues.

Reason I’m asking is that I’m assessing whether there’s still a place for a trend following system in my quiver, once I build a momentum system.

[JulianCohen]

You could set your trigger to be “high greater than” as opposed to “close great than”

Worth a test to see the difference.

As to selection bias…I have a lot more selection bias on my Russell 3000 WTT than my ASX All Ords WTT but I cut it back with price limiters and a momentum based ATR filter that gives me good results.

As an example I run a US WTT (RUI) and a momentum system (SPX) in tandem but on different markets. The WTT is far out performing the Momentum system at the moment. That could be because the Russell 3000 is doing better than the S&P, or it could be because the WTT system is working better….at the moment. It might change but I’m happy to run both as they both work well and add diversification to my portfolio.

[RobGiles]

Julian Cohen wrote:

You could set your trigger to be “high greater than” as opposed to “close great than”

Worth a test to see the difference.

As to selection bias…I have a lot more selection bias on my Russell 3000 WTT than my ASX All Ords WTT but I cut it back with price limiters and a momentum based ATR filter that gives me good results.

As an example I run a US WTT (RUI) and a momentum system (SPX) in tandem but on different markets. The WTT is far out performing the Momentum system at the moment. That could be because the Russell 3000 is doing better than the S&P, or it could be because the WTT system is working better….at the moment. It might change but I’m happy to run both as they both work well and add diversification to my portfolio.

Thanks Julian
Is there anywhere in the forum that describes in general terms what the WTT system is? I’m assuming it stands for Weekly Trend Trader?

cheers

[Author]

Posts