Complete Warrior

Definitely an improvement over Sword and Fist, but unfortunately it's not without its own misses.

Complete Warrior aims to expand on what it is to be one of the warrior classes in D&D, with the term “warrior” being loosely defined to include monk and a bit of rogue. It introduces the hexblade, the samurai, and the swashbuckler as new core classes. I've always felt that the latter two don't deserve to be a core classes because they're mostly flavour, but that's just me. The hexblade I've heard to be an underpowered class. There are also non-spellcasting variants of the ranger and the paladin.

The bulk of the sourcebook is filled with prestige classes - which was a big letdown in two ways. The first is that it seemed to me that WotC thinks warrioirs are defined by their PrC, which they really shouldn't be. The second is that most of the prestige classes are simply updated versions of 3.0 prestige classes and some reprints from Dragon magazines. They even chucked in Forgotten Realms specific ones.

Some new feats were present, the most interesting of which are tactical feats. This was wasted opportunity I think. It could have been expanded more and given better treatment. Another weak point is equipment - you'd think that warriors are a lot about tactics and choice of weapons - there's very little in the way of new equipment and weapon options. Along the magical front, we have some new domains and spells aimed for more combat usage, plus a brief 9-deity warrior-oriented pantheon.

Bringing up the rear are some ideas for warrior-focused and low magic campaigns, and a small section of warfare ideas, which ended up being somewhat pointless with its brevity.

I was hovering between two or three stars, but I went with latter. It's not a poor sourcebook at all, it's just that it tried to cover everything, but made the wrong decision to cover one part too heavily, and missed out on other parts.