Searching for Edge or Improvement

[Rapunzel]

Hi folks, hope you are well. Just a quick question to understand how you approach building trading systems.

Assuming you train, validate and test in different sets of data. Do you thin itis better to;

1) Search for a target win rate across each data set or
2) Search for a target edge improvement across each data set.

Appreciate your responses, thanks.

 

[stepandfetchit]

IMHO: (1) Find (and understand) your edge; (2) improve exploitation of that edge, (3) go to step (1) {repeat}
I think quantifying/understanding your edge may prove illusive!

 

[zghorner]

This is all you need to know.

 

[easymon1]

Hi folks, hope you are well. Just a quick question to understand how you approach building trading systems.

Assuming you train, validate and test in different sets of data. Do you thin itis better to;

1) Search for a target win rate across each data set or
2) Search for a target edge improvement across each data set.

Appreciate your responses, thanks.

Why limit your exploration? Both and then some...

Search term: Optimize
https://www.dogpile.com/serp?q=Optimize

Fly it while you build it
https://www.dogpile.com/serp?q=Fly+it+while+you+build+it

 

[JackMorgan]

Tweet of the week

 

[easymon1]

Tweet of the week
View attachment 290628

Is that how you improved your edge J-Mo or are you posting this wisdom in a thread other than that which you intended?

 

[JackMorgan]

I figure that advice is an edge, no?

Speaking of edge, dont play monkeypox scams . Very terrible loss for all concerned just now

 

[easymon1]

I figure that advice is an edge, no?

Speaking of edge, dont play monkeypox scams . Very terrible loss for all concerned just now

Ya don't say, please tell us more...

 

[zghorner]

Tweet of the week
View attachment 290628

 

[murray t turtle]

Hi folks, hope you are well. Just a quick question to understand how you approach building trading systems.

Assuming you train, validate and test in different sets of data. Do you thin itis better to;

1) Search for a target win rate across each data set or
2) Search for a target edge improvement across each data set.

Appreciate your responses, thanks.

%%
Long story short;
books help on that treasure hunt.