Days of Unreal are over, poppycock I say.
Epic's way forward: take current 3.x engine with its kilocore-powered lighting and soft-body whoknowswhats and Scaleform-powered UI that doesn't suck and all other manners of fanciness ... rebake the original and bonus pack maps ... add top MSU/community maps released to date, rebaked to taste ... add 2-3 top gametypes ... add 12-20 additional top-flight all-new Epic maps ... and why not, buck up and complete that thought that was once "conquest" or whatever, with the mining resources and whatnot ... and give a nod back to the node-heavy, fog-laden, truly expansive maps from the days of 32-player Onslaught servers that did so much to conjure many vibrant ut2004 communities....
... then tell publishers to slag off until it's done™ ...
... and release as a full-on SKU, named UTidontcarewhat, to be hailed as the uber mega value game of 2010, à la 2004.
All that's got to cost less (way less tbh) than producing a whole new full game, and must be the only near-term solution to turn a failed launch back into the proud winner that it deserves to be.
this-
As long as there are enough people playing that there is always a server of some sort that you can join to get a game, it's fine. And there is. UT3 doesn't have the exposure that Valve's multiplayer games do, but that's just fine. the smaller UT3 community is closer-knit, and has a smaller percentage of douchebags in my opinion.
Negative. Classical historians term this fallacy as "The Wartourist Defense": as long as there's one server playing a game you like, the minuscule size of the overall game's player base doesn't matter.
In an ideal world this might make some paradoxical sense, but the reality is that having an extensive selection of several servers substantially raises the quality of play for every player and within every server. For example, with several servers available that offer a particular game type, players can select one based on not just game type but also the maps that are being played. Tracking the healthier flow in and out of servers as maps change over can help admins determine which maps are most popular and worth keeping. Same with mods, etc. Same also with the quality and character of the communities that form within any given server. Same with many many nuanced aspects that contribute to a healthy global community, far beyond what is possible with "one or two" servers per game type.
Also, UT3 is in fact filled with douchebags. Many of whom happen to be admins of those one or two servers you managed to find, which turns the point right back to the one above—with several choices available to the players, popular servers have more incentive not to employ douchey admins, lest they lose their popularity to someplace more congenial.
And the circle of reason is complete.