Spring RTS Engine

Harbinger wrote:JJ I like the idea of springmark as a way to avoid the problem of having a really slow player ruin a big game.
However, you need to improve it so that it takes RAM into account and gives more consistant scores.

you cant take ram into account, because you never know how many other programs are running at that moment...
happely most ppl have enough ram!

jj the same can be said about cpu usage.

You would hope people arent running high cpu usage programs while playing spring...

spyware, malware, antivirus, firewalls, cpu temperature, Operating system actions, background processes, hardware events, and the fact most users arent tech saavy.

Wait, what does the autokick do? What's wrong with autokick, we've been looking for options like that for a while...

It doesnt autokick ppl you dont liek ro who keep rejoining thats already there.

What this autokick does is boot you if your spring mark scores not higher than what the host wants.

jj, you realize everyone hates you now because youve brought this monster of a program into spring?
Just so you know.

LOrDo wrote:jj, you realize everyone hates you now because youve brought this monster of a program into spring?
Just so you know.

Not every one : all players who haven't enough spring mark score :D

drolito wrote:

LOrDo wrote:jj, you realize everyone hates you now because youve brought this monster of a program into spring?
Just so you know.

Not every one : all players who haven't enough spring mark score :D

Not true... I would assume that I have enough Spring Mark points (Intel Dual Core 2 2.67Ghz), but I still refuse to use Spring Mark... though I wouldn't say that I dislike jj... just his program

I like the option to only allow people in your game that won't lag. It's the hosts game and he should be able to decide which players he wants. He doesn't need a reason to kick specs or newbies.. It's his game..

But I don't think the SpringMark results are good enough to decide if someone will lag or not.. When I ran it yesterday I got <300 points. JJ would kick me from all his games with that number.. The reason was that my CPU ran at half it's speed.. Having enough RAM and a good gfx card, I still wouldn't lag the game even if my CPU runs at half it's speed..

Now I've been thinking. The only good source for determining lagging is the lagging itself. I could include a statistic into IceUI that compares ping values and CPU usage (always relative to the average or minimum in the game) and save it into a file. With a simple batch (and shell script) file it could be syncronized with a server to build a lag database. This database could be only displayed ingame or integrated into the Lobby like SpringMark.

If you say that only few people use IceUI, I agree.. But one IceUI user in a game of 16 players would be enough to report the lag in that particular game so I don't think that's an issue. This could also be done as a standalone widget but I would prefer the results to be displayed in the Player Info window. The performance penalty would be minimal because it could be calculated each second or every 5 frames and it's just arithmetic with some small numbers.

Except for the lobby integration this could be setup in a few hours and I'm willing to do that. The question is: Do we want that? Does it discriminate against low end machines? Does it violate privacy (because you don't report your own values.. others report it for you)?

Something similar could be done to identify rage-quitters or people who attack allies, etc.. It wouldn't be much work but the questions above also apply..

jj wrote:

Harbinger wrote:JJ I like the idea of springmark as a way to avoid the problem of having a really slow player ruin a big game.
However, you need to improve it so that it takes RAM into account and gives more consistant scores.

you cant take ram into account, because you never know how many other programs are running at that moment...
happely most ppl have enough ram!

Pfft. I've bee playing games while having my whole load of dev tools open. That's no hurdle, that's what swap space is for.

AF wrote:If satirik adds autokick to his client I hope he's prepared to suffer the consequences. This is exacly why people objected to spring mark in the first place and it is the worst route you could go down jj.

Your nuking your own project JJ.

its so funny, a couple of months ago YOU was that guy with the idea of intergrate it into your lobby client!

Integrating spring mark != autokick feature for ppl with a low springmark result

Theres a very big difference there, one is immoral the other is not. I have explained why I wont be including springmark in AFLobby, especially since you never provided enough information for it to be done to begin with.



Meltrax, if you provided me with a lua widget that did that I would bundle it with aflobby and integrate within hours. A cross platform reliable setup like that is far more valuable. I'd say that calculating average ping and fps ingame and then averaging that with those already recorded would do the trick.

AF,

you can grab info for a lobby system info feature like this:
http://www.jobjol.com/lobby.php?players ... yers[1]=DC

AF, I will implement that if the people flaming jj in this thread think it's a good idea.. I think it will be a better indicator than jjs CPU test but the question if we should do it in the first place is the same..

meltrax,

does every user need to use that lua script or only one per game?

/edit: what about ppl only playing small games, do they get a higher score? and ppl that play laggy speedmetal games, do they get a low score? i hope you have a solution for that, if so, i think it will get really usefull

record also the # of players, the # of units & the map....

Only one per game would have to use it (and submit it if it should be accessable to all)..

I would substract the average lag from the game.. So the player lagging the least (except the host ofc) would be the basis of the calculation.. For example:

Team0: 000ms -> 000ms (Host)
Team1: 100ms -> 000ms (least lagging player)
Team2: 400ms -> 300ms (difference to above)
Team3: 250ms -> 150ms (same)

The same can be done for the CPU usage and both could be combined to some kind of Lag factor.. Or we can use both..

BrainDamage: That could also be done.. I don't know if it's enough to just do the above.. Some more votes?

jj wrote:AF,

you can grab info for a lobby system info feature like this:
http://www.jobjol.com/lobby.php?players ... yers[1]=DC

Hmm... isn't that kind of fubar... ?
Shouldn't the fields be separated by tabs and the different players by newline?

exactly, also the numerous fields dont make sense save the most obvious ones