a persistent TF2 world server

gamemaster1996

L13: Stunning Member
Sep 30, 2009
1,064
134
Hmm i'd have to have a look at what this place looks like.

EDIT:
2489557238_757f84a87a.jpg


If thats what the buildings look like there i might just pass.
 

Trotim

aa
Jul 14, 2009
1,195
1,045
A Schwarzwald map would probably be a lot like Lumberyard. I can't really imagine a big A/D CP map since everything looks the same in a big forest.

And yes, some buildings look like that, but that's just some cottages right in the middle of the forest. There are bigger industrial complexes too, though these are kinda generic and I'd suggest using that architecture if you really wanted to make a map.
 

SPHinx

L2: Junior Member
Aug 29, 2009
81
24
Since this thread has been resurrected a bit, I've muddled over some of my initial considerations about a server rotation script. Curious what others might think of this. Totally speculative here, but fun to think about.

I'm assuming most people did not see this on the TF2 World project site, so I will summarize: Given a 36 map rotation, I came up with a way to cycle them without too much repetition, represented graphically like this:

36pointring.gif


In case it isn't obvious to you, the red nodes would be RED defend maps, the blue nodes BLU defend maps, and the gray nodes are neutral maps (like 5CP, PLR, KOTH). Each node has two entrances and two exits, and the shortest repeating cycle is 8 maps long, requiring one team to win 7 times in a row.

Perhaps interesting in itself, there are some obvious problems with a simple script like this. First, do people stay on a server long enough to care about their team's victories? Second, what's to prevent people from just switching to the winning team or team stacking? Third, what about maps that swap the teams at round's end? Those are all legitimate concerns, but some of them, I think, can be answered, perhaps bordering on the need for a mod.

Using a tool like HLstats, it would be possible to track the performance of players on the server, generating some aggregate rating of their skill to balance their team association to RED or BLU. The team selection could then take place only once each time a player connected to the server (or automatically switched by the server) based on their skill. This way, as maps changed, teams stay constant and, hopefully, balanced.

This could also be taken a step further, locking players to a particular team on that server for some amount of time, perhaps until RED or BLU wins the super-round and controls the entire world map (or the teams become too unbalanced and a player is moved automatically). In this way, even if a player only wishes to play for a map or two, they can return later to pick up the fight for their side. This should be possible by querying the player's server-side team association (after their initial choice on first connect) and then passing this variable immediately during team selection thereafter. If adjustments for balance are not made too frequently, I think this could be fun.

In fact, as I write this, it occurs to me that you could represent a player's performance skill in $. Rather than switching players automatically, If RED or BLU starts to get too far ahead in skill points (which is then also to say in $), there could be a screen pop-up where the administrator or Ms. Pauling shows up to offer players with high skill ratings a $ bonus (or an unlock?) to switch sides.

Disabling team swaps (and having to repeat defensive victories) ought to be a pretty simple fix. This will shorten the time on some maps; however, I don't think most players would mind this. It may even make the defensive victory more satisfying, or could lend itself to an attack/defend/counter-attack kind of map design. For example, say RED defends its base in the first round (of a three round A/D), there could be a kind of bonus round where they turn around and attempt to sabotage the outpost BLU launched its original attack from.

Putting all of this together into a mod, with things like a loading screen to show the status of world territories controlled and a leader-board for each team, would be quite the awesome package, I think. Not a replacement for the excellence that Team Fortress 2 normally is, but an alternative way to deepen the experience a bit.

Whew! Sorry for the long post. Thanks for bearing with me.
 

Trotim

aa
Jul 14, 2009
1,195
1,045
...I'm not sure if even half of this is reasonably doable. Sure it sounds nice, but wow. That's a lot of changes.
 

Jindo

L3: Member
Aug 6, 2009
121
123
I'm still very interested in this idea and I will most likely try and come up with the rotation script in the near future when I get the chance!
 
Last edited:

SPHinx

L2: Junior Member
Aug 29, 2009
81
24
I'm still very interested in this idea and I will most likely try and come up with the rotation script in the near future when I get the chance!

Cheers mate. But as luck would have it, FireSlash has pretty much written the thing already. You can read his posting on our forums here. You could certainly still work with us on customizing/tweaking the implementation in some way though.
 

Jindo

L3: Member
Aug 6, 2009
121
123
Well that's excellent, he certainly seems to have taken it much further than I would have (to begin with).

If there is anyway I can help, I'd be glad to!
 

SPHinx

L2: Junior Member
Aug 29, 2009
81
24
If there is anyway I can help, I'd be glad to!

Of course! In addition to the FireSlash's WAR! script, we've done some brainstorming about how to best link up the maps, how or whether to restrict teams, collecting and reporting stats along side the world domination web interface. The last few are elements are outside the bounds of his script, so we may yet need more people scripters (like yourself) to implement those kinds of things.

Then there's always a need for more mappers, testers, and feedback-givers. If you like, you can also just join our forums or steam group to keep an eye on how things go.
 

Jindo

L3: Member
Aug 6, 2009
121
123
I'll sign up to the forums and see what I can come up with, I've got a fairly good grasp of SourcePawn and might be able to resolve some of those issues (provided we first find the right solution of course!)