Unreal Tourney guru's ... I need help with ngStats problem.

[merlocka]

I've been hosting some UT CTF LP's. The server is Win9x, running dedicated UT version 4.13 (I've also experimented with others, not including anything newer than 4.25). The game machines have various versions of UT on them, mostly Win9x boxes, but there are also 1 Mac and 1 Linux machines

For the last 4 lan parties, stats have been all wrong. ngStats is "skipping" over alot of events, and it contradicts itself in the data. Also, mods like CTF+ and CTF4 show up as Tournament Deathmatch games, but in the "High Scores" category in the stats it lists "Most Flag captures" and such so I know the Mods are recording events properly.

Anyone else see this? Am I going MAD!?! I've done fresh installs of both UT and Win9x. I've run different versions of UT. The messed up thing is that I cannot reproduce it at home with bots. Is it possible that a specific client is screwing things up and I don't see it unless they are connected?

Thx

 

[Becks2k]

not sure what exactly is wrong, but ng stats are never right. I just play CTF and whenever you look at ngstats the points people have are always different than what that had in a screenshot.

screenshots are the only way to go, demos also if its important.

 

[merlocka]

Has it always been like that?

 

[Poof]

I've been hearing about alot of problems with the ngstats being updated correctly. You might want to opt for running local stats (although it adds some overhead to your server/lan) to see if that is more accurate.

 

[Soccer55]

I think the reason that you can't reproduce the results with bots and why the scores are different between the game and ngStats is that when there is a bot in a multiplayer game, I believe that you get no points whatsoever for a kill. You get the stat for the kill with the weapon you use, but you don't get the point for the kill on the bot. This is my understanding of how the stats work, so I might be wrong. Regardless, multiplayer UT is tons of fun (especially on an OC-12)

-Tom

 

[merlocka]

Thanks for all the replies. Just to clarify, I am referring to local ngstats for CTF games. I've always heard that ngworldstats are truly hosed up, I just assumed that local ngstats would be accurate.

I tend to believe that Becks2k has hit it on the head. I suspected this awhile ago. I've hosted 8 lanparties and we were all new to the game early on so I doubt we would have noticed any differences between the screen scoring (at the end of a CTF match). Now that we have become more competitive we are very aware of what our numbers "shoulda" been in each game. The ngstats numbers don't add up. Here are some things I've noticed.

Flag kills, flag pickups, and flag drops seldom add up correctly. In theory, one teams total flag drops should be nearly equal the other teams flag saves. In our ngstats these numbers are hugely different. In addition, It appears that ngstats *sometimes* has flag kills and flag saves reversed.

I'm gonna have a mini-lan and run some tests to see if I can reproduce anything consistantly. I'll post any finding.

thx

 

[Becks2k]

I think ngstats get a little messed when someone picks up the flag, is killed, a team mmeber picks it up before it's returned. Or what if the guy telaports, thats a flag drop but not a flag kill? And if at the end of the game some guy is holding the flag, there will be a pickup but no flag kill/drop/cap for it.