I've written about this before; I don't think it is possible to develop a reliable 'rating' for how good a unit is. I prefer to use instinct and intuition - I KNOW that No-Dachi are good flanking units, that they are weak defensively etc. by experience, not by reading statistics.

I'm not against statistics, but for a game like STW I think there are too many variables to have a unit rating system that works. I find that it is actually easier to go by my gut instinct about how to use my units. I usually pick a balanced army and micro-manage as much as I can.

Also, different units have different purposes. For example, I don't expect my Yari Samurai to get many kills. I expect them to hold their formation, soaking up monk attacks while my No-Dachis flank the monks. So my Yaris are going to take casualties without killing very much, and the No-Dachi will get the majority of the kills whilst (hopefully) not losing very much. To compare the two, it would appear that the No-Dachi are incredibly effective whilst the Yaris are not. However, this is not true. It might be possible to compare similar units - e.g. Yari Samurai vs Naginata, No-Dachi vs Monks, but the differences between these units are quite straightforward and well-known already.

------------------
MizuRob Hojo
Honour to clan Takiyama