Why share experiments that fail?

We often see a polished result. We rarely see the dozens of ideas that did not work.

But what if the failure is the most useful part?

SHING CIVITAS is an ongoing attempt to understand how artificial societies behave. A resident might fail to learn, a world rule might determine an outcome without anyone choosing it, or a beautiful graph might conceal a weak assumption.

This archive is a commitment to document those questions honestly.

What will a research record contain?

Whenever evidence is available, an experiment entry will include a research question, hypothesis, model and code version, experimental conditions, main metrics, results, limitations, and reproduction materials.

What does this not prove?

Even a perfectly repeatable simulation does not automatically explain actual human societies. We need to distinguish results inside our model from claims about reality.

The record begins with curiosity, and stays open to correction.