Terry’s Journal

[BenOsborn]

Impressive Terry

[AnthonyFuentes]

I just went through your journal from start to finish — incredible results! Even more, there was a lot of useful information shared here.

Terry, at the top of page 14, you said:

To make it through my MOC/MR back testing process and be traded live, I use the Monte Carlo Simulation results as follows:

1. The average (of the MCSs) CAGR must be greater than 20% and greater than the average Max DD
2. the worst case (of the MCSs) Sharpe Ratio must exceed 1

Would you be able to share, are these statistics from the default backtest in AmiBroker or from a MCS test that’s run using optimization?

Thank you.

[TerryDunne]

Hi Anthony,

Thanks for the post. The first thing I should say is that I got lucky and started at a great time, so I immediately had some momentum.

I usually do my initial back testing from 1/1/2010 to 1/1/2020. I may choose to optimise some parameter or other, but I really want whatever optimised value I choose to be in the middle of some range of good values (i.e. all of them work).

I also want the draft system to be at least OK on the 2000-2010 and 2020+ timeframes and to also work on both the Russell 1000 and the Russell 2000 (this is the MOC systems than the MR systems, where I haven’t been able to crack one on the Russell 1000).

When I think I have something that might be tradeable, I run a Monte Carlo simulation, again on the 1/1/2010 to 1/1/2020 period. I do two things with the output:
1. I check whether the average of the MCS runs makes my minimum benchmark, which is that the CAGR is greater than 20% and that the CAGR is greater than the max draw down. I no longer use Sharpe Ratio as I have been won over by the Ulcer Performance Index, which has a denominator based on the depth, frequency and duration of draw downs (although I should confess that I’m not over the maths on this).
2.I also use the MCS output to allocate capital to different systems. Right now, if the systems wouldn’t be allocated more than 4% of my capital, I won’t trade it. Context here is that I have 15 systems, I had a higher % allocation requirement when I had fewer systems.

As I mentioned, I started at a really helpful time, so I’ve been lucky – I hope you are lucky too!

Terry

[TerryDunne]

One other thing…in my opinion, the true measure of success is longevity. Someone like Nick has been doing this with great results for a thousand years and many of the people here have been churning out great results year after year for ages. That’s what I aspire to do.

[AnthonyFuentes]

Thank you for the insight Terry, that is helpful! I like the idea of having a benchmark that the system should hit. I expect that removes some second guessing about whether or not a system is worth trading live. Curious, is there a reason you use 1/1/2010 to 1/1/2020 for your backtesting period? I ask because I’m looking to shorten my testing period as AmiBroker can be dreadfully slow at running MCS.

I think it was Seneca who said “Luck is when preparation meets opportunity.”

And yes, nice thoughts on seeking longevity. I’m in the same camp as well.

[TerryDunne]

There’s no science to it. I first started back testing in 2020 and ended up with a pile of MCS’s for that decade. I use the same time frame for new systems so I can compare apples with apples.

[JulianCohen]

I heard your comments on the Bensdorp Book on the Group Call.

In RealTest, Marsten has coded up the various strategies…I don’t know for certain if they are all accurate as I haven’t read the book, but knowing Marsten I’d say they are.

There’s 6 strategies, 4 long 2 short, with various time frames.

As a portfolio results are CAGR 8.76% MDD -28.27% since 2014 to present.

If you want full metrics I can send them to you.

 
 

[JulianCohen]

I heard your comments on the Bensdorp Book on the Group Call.

In RealTest, Marsten has coded up the various strategies…I don’t know for certain if they are all accurate as I haven’t read the book, but knowing Marsten I’d say they are.

There’s 6 strategies, 4 long 2 short, with various time frames.

As a portfolio results are CAGR 8.76% MDD -28.27% since 2014 to present.

If you want full metrics I can send them to you.

The best performer is the Short Mean Reversion High Six-Day Surge strategy

Sorry about the screenshots, I can’t ever seem to get them to work well on this forum for some reason

 

[TerryDunne]

Hi Julian,

Thanks for posting this (and for taking the trouble to get the screenshots to work), it’s certainly food for thought, not very palatable food either!

I have been more interested in the concept than his actual systems. For example, I wouldn’t ever be able to have a short only S&P system. Also, Bensdorp doesn’t use filters at all, which is a real problem for me…

I’m really grateful for your post as it gives me a justification not to take money out of higher performing systems and put it into poorer performing systems , thanks again!

[TerryDunne]

October 2022

MOC 1000 12.20%
MOC 2000 16.18%
MR 2000 0.00%

Total 9.42%

[TerryDunne]

November 2022

MOC 1000 5.27%
MOC 2000 15.18%
MR 2000 0.00%

Total 6.98%

[TimothyStrickland]

Nice on both MOC’s Terry

[TerryDunne]

December 2022

MOC 1000 -7.70%
MOC 2000 0.35%
MR 2000 -3.89%

Total -3.74%

[TerryDunne]

2022 Year

MOC 1000 92.07%
MOC 2000 124.02%
MR 2000 -10.64%

Total 46.58%

For the coming year, I’d like to decide whether to launch weekly swing systems for the ASX and TSE and also to fully document daily processes and the process for closing my positions if I get run over by a truck.

My children have expressed interest in learning what I do, but I’m still trying to decide whether they are actually interested or just like the idea of making $$$ without having to work too hard. If I think there is genuine interest, then I’d also like to provide them with some training. I’m not too confident abut this as I can imagine a whole pile of events and circumstances where it could go pear shaped.

Good luck to everyone for 2023!

[BenOsborn]

Wow impressive results Terry

[Author]

Posts