The optimal window seems to be defined as thus: any time when there is a large number of powerful, high-level noders hanging around.

But aha!, there are tricks to ensure even greater success:

When you're satisfied that you've done all you can for the little guy, it's time to release your node into the mean streets for evaluation. Perhaps the best marketing tip of all is just to write good nodes. After all, look at the world of computing - no one buys operating systems just because of good marketing, do they?

(Whew, I think a few more nodes about noding and I can finally earn my certificate of nodology)