Backtest results change when I stop using variables
- This topic has 56 replies, 3 voices, and was last updated 4 years ago by
.
-
Thanks GraHal
You’re a man after my own heart. I’ve already started on the beer 🙂
I’m not 100% sure what changes you’ve made. Could you attach the code please?
I know Nicolas has written a trailing stop algorithm. I may try that or just tweak a fixed stop.
Thanks again for spending so much time on this. I’ll let you know if I can fix it based on the sideways market of the last 21-24 months.
1 user thanked author for this post.
I’ll let you know if I can fix it based on the sideways market of the last 21-24 months.
That version of mine makes over 10K in the past 24 months … so just keep enjoying the beer and making the money! 🙂
The code is attached to my post as an .itf and to see the differences enter your version and mine on the link below. Takes less than 1 minute (after you have got in a routine), you don’t need to sign up / in and it’s easier than doing a spot the difference check. I uses it often.
https://www.diffchecker.com/diff
1 user thanked author for this post.
Sorry
Missed the file link – thought it was the name of the picture files. (Novices – they’re hopeless, you know!) Thanks also for the diff checker. I had a quick run at your code on tick by tick (on True too) and am now very much looking forward to Monday!!
I think I’ll be able to afford wine too now 🙂
Plagiarism is a wonderful thing
Thanks again
1 user thanked author for this post.
I had a quick look at it and yes the robustness test results are good. Attached are my 20/5 robustness test results.
The entry conditions do seem like they have the potential to be curve fitted and analysis of so many variables and combinations of variables to confirm whether they are or are not curve fitted could take an awful long time.
There is also the concern that during the test period the DJI has basically been a bull market except since the beginning of 2018 where it has been going sideways in a very volatile manner. The majority of robustness tests show the equity curve as being under water for 50% to 75% of the first part of the test period before taking off in a very bullish market and then going sideways again in a very volatile sideways market. This concerns me. Would I want to trade something that goes sideways or is in a small loss for 7 to 8 years of a 9 year test period – probably not.
At the end of the day we must always compare our equity curve to buy and hold on something like the DJI. We either want to seriously beat it or get very close to it but avoid all the major draw downs on the way or we want a strategy that just consistently and repeatedly makes money and more money than we could get if we invested in something much lower risk than spread betting.
It can seem sometimes as GraHal says that my views are opposite to what others want to hear about a strategy but it is always a good idea if you are going to risk your own hard earned money on a strategy to take one step back (or three or four preferably!) and try to find as many reasons why you shouldn’t be trading it before counting all the reasons why you should be trading it. Sometimes one big reason can stand out and be a singular reason as to why you should not trade a strategy and sometimes a whole bunch of small reasons can add up to be enough reason not to.
In the adrenaline rush of the excitement of feeling that you could suddenly be rich if this strategy works out it is sometimes difficult to step back and with a cool head look for reasons why it is actually a rubbish strategy. I have created strategies in the past that have kept me awake at night with excitement but then after two or three days of hard testing I have decided to never ever trade. Don’t be in such a rush to be rich or you could find that you just rushed towards being poor!
1 user thanked author for this post.
Find exclusive trading pro-tools on 
