Feature request on UT3 : a smart autoteambalance

[Fixxer.Linux.T32]

Hi,

I'm playing very often to UT2004 (on LINUX CLIENT version, PLEASE, take a good note of that as I'm expecting UT3 on Linux as well) on TitanInternet ONS-32 players server (ut2004.titaninternet.co.uk/forums/servers.php).

What appears to be sad on UT2004 is the filling of the teams.
On each map change, you are logged in a team depending on your time of arrival on the server, the last logged in the server being changed of team if it should balance the teams.

What could be more smart is to balance the teams depending on the players skills (only on map start).
This means that the skill of the player should be recorded on the server, in a small database containing probably only the game key and the skill level. Only the last x days (x could be changed by the server admins) to be saved on the server containing 2 numbers : the number of maps played and the overall score of the player.
These 2 numbers would give an average "skill by map", to use to sort players. An easy way of sorting players would be the following : odd "skill by map" go in red team and even "skill by map" players go in blue team.
A rule should be found out to clean the server database, as player skills can change and to avoid keeping scores of players that don't play anymore (and I hate wasting disk space with obsolete/non used information of any kind)

This way, you can guarantee that the team balance has better chances to spread good players between the 2 opponents and, by the way, the fun to play the game.
There is no more frustrating games as playing against un-consistents opponents, especially when you are in the "bad" team !

I hope you'll be interested in implementing this idea in UT3.

[Jrubzjeknf]

Expressing skill in numbers is flawed, therefore team balancing by software is flawed.

[Piglet]

Shambler's working on a "balance by skill" addin to his TitanTeamFix balancer.

We currently have it llinked directly to a MySQL database where it's reading and writing - so I expect it'll be up and running soon.

There are some stats that are useful for team balancing - over time win/lose ratio and PPH do mean something. If you have someone who nearly always loses and scores few points then they're either not that good or (rather more unlikely) are an excellent but unlucky defender.

[awaw]

Expressing skill in numbers is flawed, therefore team balancing by software is flawed.

Expressing skill in loading time is better?

[fuegerstef]

Expressing skill in numbers is flawed, therefore team balancing by software is flawed.

Trying is better than nothing. See post above mine.

A feature like this (although being far from perfect) would help a lot.

[Unreal Havoc]

I can't say I disagree.

[Fixxer.Linux.T32]

Shambler's working on a "balance by skill" addin to his TitanTeamFix balancer.

We currently have it llinked directly to a MySQL database where it's reading and writing - so I expect it'll be up and running soon.

There are some stats that are useful for team balancing - over time win/lose ratio and PPH do mean something. If you have someone who nearly always loses and scores few points then they're either not that good or (rather more unlikely) are an excellent but unlucky defender.

Thanks for the information Piglet.
The rounds will be longer to win as soon as this addin will be live on the server, but at least, we'll have fun !

[Hedge-o-Matic]

I wholeheartedly agree with this idea. Team balance can take into account as much complexity as desired; the players don't need to do a thing except play the game.

I'm always the first to jump onto a losing team; I love playing with the underdogs in difficult situations (which is why I love a good entrenched camper who rules a swath of map, forcing me to get creative to dislodge them).

Maybe an optional checkbox to be one of the first considered for switching to the weaker team would be possible, although presumably the system would just try to balance things out using skills, which is the way it should be. Still, a checkbox would allow the computer to switch you if there was a tie in skills.

[Kronos]

I'm not sure how it would judge your skill unless it judged you by your stats, and stats certainly dont mean **** in how actually skilled you are. I could see teams becoming even more unbalanced if it judged you by stats.

[fuegerstef]

I'm not sure how it would judge your skill unless it judged you by your stats, and stats certainly dont mean **** in how actually skilled you are. I could see teams becoming even more unbalanced if it judged you by stats.

I'd say that stats are bad, but how you did in the last few matches, could work, if the point-sharing in each gametype is improved.

[Kronos]

Wouldnt that make it unbalanced for the first few matches though until it could recognized how well you did in the last 2 other matches?

[fuegerstef]

Wouldnt that make it unbalanced for the first few matches though until it could recognized how well you did in the last 2 other matches?

It should only record players that are there for one full match. The rest is distributed like now.

[MonsOlympus]

I dont think its so much that stats arnt an accurate guide to skill but which stats give the best results (closest to true skill), or even the lack of stats for given skills when autobalancing.

[Jrubzjeknf]

And what if multiple players use one PC? Then your algorithm is already failing, because it's based on past results which don't have to apply on a certain player. This has potential to actually unbalance teams, because the algorithm matches that player to another that's much better.

[fuegerstef]

And what if multiple players use one PC?

Only if they change inbetween 2 matches during one playsession.

This has potential to actually unbalance teams, because the algorithm matches that player to another that's much better.

How does this unbalance more than now ? Can you enlighten us on this, please. Would be really great.