For me, the defining characteristic of a truly great game is replayability. Despite being a story-driven RPG, Fallout 3 - like its predecessors - has plenty of replayability. For instance, when you first begin to venture out from the comfortable surroundings of Megaton - which for most players is their home base - you will probably encounter a nearby grocery store. I say probably because it totally depends on which direction you decide to wander. Anyway, within and around that store is a band of raiders, and they guard a fair-sized pile of resources. Now, you can skip this encounter if you like. There are a lot of them and only one of you. You can sneak in and take only what you need, then get the hell out, which is the approach I took my first time through. I returned later on to confront the raiders when I was stronger. But there are many more ways to play this. I can see someone deciding to challenge the raiders early on, to take on this store full of maniacs, picking them off one at a time, heading back to Megaton to heal and rearm, until they are all wiped out. Cool.

Rarely have I had such mixed feelings about a game as with Far Cry 2. The first Far Cry was loads of fun. Sure, the ending was stupid, but 95% of the time I was playing it I thoroughly enjoyed myself. Great sound and graphics, scary and humourous in turn, nice weapon choices, a variety of transport choices, tough encounters and innovative gameplay made it one of those rare commodities: a mostly-scripted shooter that is fun to re-play.

Far Cry 2 has a lot of those same qualities. The graphics and sound are so utterly immersive, I often feel that I am alive inside the landscape of the game. Escapism at its finest. I haven't reached the end yet, so I can't comment on that, and so far there has been very little humour or anything particularly scary, which is a shame. Weapon choices are diverse, although quite limited at first. Transport choices are also abundant. Encounters are so far relatively straightforward, but I have heard that I should expect more challenge later on. And of course FC2 includes some gameplay innovations. But does it all work?

I've re-enabled the old stats system, because it shows a few things that the new system doesn't have, like player rankings per map and server stats. It's still at the original location: oneirozoo.com.

I also reset the old stats system and reprocessed the data to match the new stats system, which started operation early on February 10, 2009.

You may notice that the stats on the two sites don't always match perfectly. I won't be able to find or fix any bugs in the old stats system since I don't have the source code, but I'm still looking into possible bugs in the new stats system and will fix any I find.

Stats reprocessing was completed as planned early this morning. While the main player and map summary stats haven't changed, most of the detailed stats are now much more accurate and useful.

Three major problems were fixed: 1) kills of bots by human players were not being recorded except in the main summary; 2) character type (kit) usage numbers were not being counted correctly; and 3) vehicle usage numbers were not being counted correctly.

These three issues have been resolved. Areas of the stats site that should now be much more accurate include the player details for character type, weapon and vehicle use; the character type list on the character type usage page; the vehicle usage page and the weapons page. The weapons page now shows kills by weapon category.

You may notice that some of the detailed numbers still don't match up perfectly with the summary numbers. This is being looked into as well, but it may be simply the result of missing information in the game logs.

Tonight I will clear the stats in preparation for reprocessing all the data since about February 10, using the new stats processing software. Several major bugs have been fixed and the resulting stats should be much more accurate and useful.

The server search function at the bottom of the main TPU stats page was not working and the associated year dropdown list was not up to date. Both of these issues have been resolved and server search now works correctly.

There was a bug in the player details on the TPU stats site whereby the 'last seen' date/time for every player was the same as 'last seen' for player 1. This bug has been corrected. 'Last seen' now shows when the player last played on the server.

The Tanks 'n Planes Unlimited rules have been posted in the forums.

The Tanks 'n Planes Unlimited server/stats FAQ has been posted in the forums.

The points formula on the TPU stats site was revised again, to add average heals per round and average repairs per round.