Why are questions about "portfolio optimization" considered offtopic?

I noticed a trend where questions about portfolio weight estimation and optimization tend to either be suggested for transferring over to quant SE, or even closed as off-topic.

Here are a few examples:

1. Efficient Portfolio Optimization Through Simulation (note the comments)
2. Markowitz portfolio mean variance optimization in R (note the first comment)
3. Alternative Applications of Portfolio Optimization (closed as off-topic)

However, when one reads the literature - the problems discussed in these topics are of clear statistical nature, involving issues of sample size, uncertainty estimates, and model complexity. Example: "Portfolio of Automated Trading Systems: Complexity and Learning Set Size Issues".

So I am just curious - why is it considered to be off topic by many?