Propietary information is ok. You don't have to give away the how's or why's or even the name. The "anon" prediction model. You have to be able to use it to predict and predict accurately.

Someday, if you publish or want recognition, well, you might have cough it up or be ignored. Others have to be able to reproduce the results for it to be an empirical method based on science. You also run the chance that someone else may come upon the same methodology and get credit for it because they revealed the secrets first. The "marc/berkeley brat/beer/onions model"!!!

It's a tangled web...