Backtesting optimization with wrong parameter values

Hi at all,

I'm studying and improving the framework of my EA,
with a basic strategy that I use as an example, currently basic 2ma.

I wonder, and now I ask you too:
how useful is it to do tests with wrong parameter values ???
How do you have instrument values, TP and SL, or even others ...

I show details, for example:
principle of the strategy I am using:
2ma
buy signal: when 1ma breaks bull 2ma
sell signal: when 1ma breaks bear 2ma

So the results to be analyzed would be those based on instrument values such as:
1ma-2ma
5-7
7-11
7-14
7-21
etc ... for example

When it is useful to do tests with instrument values such as:
1ma-2ma
7-5
11-7
14-7
21-7

In some cases, I also got good results, but I believe based on the case, or it would be the case to study the strategy according to other principles ...

As I see it now, backtesting optimization is as good as it is dangerous,
but our profession is based on knowing how to manage losses and get profit.

Can you tell me something about it?
or some good article / post that talks about it.

Thanks at all