[ScottMcNab]

Rob Giles wrote:

thanks Scott,

Yes I can see that an investment in getting the wrinkles ironed out of new systems by Nick & Craig in the future is money well spent.

My post may have come across in the wrong tone Rob…I shouldn’t type before coffee. I remember feeling really pressured for time during the course…just seemed so much to try and master. When I re-read what I wrote it sounds like I think I have the answers..nup

[ScottMcNab]

This is something I used to see more in R1000 but occurred last night in SP500…was in trading hours (250 shares 3 sec after open)(was not a spread or special trade)…would also be avoided with primary exchange only data….this has to be impacting backtests on MRV systems

[ScottMcNab]

I have been trading ASX and US MOC for a year now….I am about to launch rotational systems on these markets but have subscribed to the extra 3 months to get Craig to check code and Nick to improve/advise on implementation…..hard to imagine I would ever have sufficient knowledge and experience such that getting their input isn’t worth the 800 bucks or so…. absolutely worth getting as much out of the 6 month mentor course as you can but take the pressure off yourself to feel that have to know it all and have every system running in that time….2c worth (at most)

[ScottMcNab]

Nothing sus with those equity curves !

[ScottMcNab]

July -0.02%
Pretty quiet

[ScottMcNab]

Interesting article Julian..many thanks

[ScottMcNab]

Yeah..the devil is in the detail….I would not do it outside of mentoring program as my chances of not missing a crucial factor would be zero…not sure if Nick trades ETF’s….if not then that is probably the first red flag

[ScottMcNab]

I think there was some talk a while back on the forum about using ETF’s in Amibroker backtests…based on what you have posted Julian perhaps ETF’s may offer us exposure to commodity/financial markets in a manner we could learn more easily using Amibroker backtesting techniques that have applied to stocks ? Ideally, we would take an existing system (eg rotational) and apply it to a basket of ETF’s and voila….

Will add it to the list

[ScottMcNab]

Yes…was something I looked at a little while back….I did a backtest and subtracted 1 tick from the buylimit price and found quite a significant impact on ASX MOC but not on US …suggesting many of the successful trades on ASX had buylimit at the low of the day….regarding the more important question as to what this means going forward….not sure sorry…if my ASX MOC was a stand alone system I would be more concerned but as it is just using the equity from the US system overnight then as long as ASX system remained positive I thought I would let it run and watch it.

Perhaps it is just a reflection of the lower liquidity and smaller universe such that a similar effect would be seen regardless of the buylimit price selected ? I did not test this further at the time. Does make me wish Norgate would include LSE or larger Asian markets but from what I’ve read online that does not seem to be likely in the immediate future (might have been something written in Norgate section of Amibroker forum)

[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

[ScottMcNab]

Using 5% of volume for the day as a cutoff for intraday systems (assumption is that over this % is going to impact price) even the US stocks seems struggle to have the liquidity once reach 8 figures…most funds would have aum of 50m+ I’m guessing…course I have probably the least experience or understanding of these issues of anyone in the entire mentor course so my posts are only meant as topics for possible consideration

[ScottMcNab]

Capacity plays a role in the systems we are designing too I think….20% for a 6 fig or small/mid 7 fig account possible without moving the markets but these systems would probably not translate such that large funds would be able to implement them and are therefore not using them…that is the best explanation in my mind as to why these work…it may have been quoted in here before where Buffett ponders the return he would be able to achieve with only one million dollars …hope that helps clear the hurdle Rob ?

[ScottMcNab]

Do they keep the trades open until they go above break even (no transaction costs)….if don’t go into the black then trade is classified as still open ?

[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]

Hi Trent, I was trolling through some old folders and found a MOC system that made me think of your post. I abandoned that line of investigation as it would have required me to monitor the market during the day to get the entry….required price to open in a certain position and then required change from the open of a certain percent or atr multiple

While there was nothing special about that system as it was, it does open up some additional coding options if you were able to monitor the market and place the entries during the day…maybe could try a weekly system so only had to do that for one day a week rather than all of them

I imagine it would have to achieve significant returns or add sufficient diversity to current systems to make it worth it though

[Author]

Posts