Hello everyone, I started a thread a few months ago in the french section. Unfortunately few people contributed. The goal is to use Google Sheet in order to help us build and analyse a portfolio of systems. As we know having 1 or 2 algos is not the key, the risk can be huge. Algos which performed well in OOS and forward testing can suddenly somehow “desync” with the market. It can be painful to watch our expected edge getting hit by the reality of the market.
That’s why I’ve started this GSdoc in order to have a better look at the number.
Unfortunately, I don’t have the time I had a few weeks back.
With the help of the Monte Carlo simulation I hope to somehow get a trigger to kill algo or the entire portfolio in case of too much desync.
I need your help for :
- Providing TrackRecords if you don’t wanna share algos [IS/OOS period/Asset] ( I only have one “ok” system, the others present in the TR are quite like brothers or cousins ).
- Providing ok/good algos and their link in the library ( I started a sorting but time is missing now. Despite good ideas a majority of them are EV- ).
- But most importantly I need you for any suggestions, ideas, critics.
Here are the different modules :
00 – Quick help : Link
0 – Track Record Analyser : The main one
1 – Annexe : Correlation / Overnight counter : Link
2 – Montecarlo : Well just a test at this stage [ It doesn’t need the SheetGo plug-in, you can cancel them ] – Link
3 – My version of the VRT google sheet – Link
Here’s the link for the folder : Link
Soon, we will be able to use the output from the 1-2-3 document and import them manually into the 0 – Main document.
Here a quick overview of the 0-Main document :
CDC : What’s plan in the next versions
TRA : This is where we put the data and can analyse the real number
R-VRT : Manual Input from the Vonasi Robustner Tester [ DocN°3 needed ], not taken into consideration for the moment.
G-All : Graphical outputs summary of the general portfolio, pretty basic yet.
G-EC/Sys : Equity curve (IS/OOS) of each system (only 3 output possible yet, gonna change that later)
RTIME-M : Give a detailed performance for each month ( not fully ok yet ). To get data from another asset you need to collect it on IG Lab and import it using the TPCAC template.
SCR : Overall control panel. In the future it will give a score and a set of checkbox
PATCH : Because risk management is quite subjective, this is the backstage of the checking table located in the SCR. Everyone would be able to share their presets.
TPCAC : Well this is the imported quote of the CAC40 index. It’s necessary to fully use RTIME-M, but as of today RTIME is more a beta functionality.
It could be interesting to run a challenge over a period of 6 / 12 months with a preparation of several months. With no real reward, just to have some sort of creative incentive. Let’s say everyone has 200 gross weight ( calculated over the IS+OOS period / or another kind of metric ) as a “credit”, algos are either published before the start of the challenge or put in forward testing and kept unchanged. Datas from the track records will be shared within Google Sheets. It could be an enriching journey I think.
You can add comments directly in the sheet. You might find some mistakes, don’t hesitate to report them directly with a comment in GS, thanks.
Keep in mind this should just add a layer of confidence on pre-existing “great” algos.
.
.
