Short Version that you can'r really see a damn thing: http://www.youtube.com/watch?v=K5Y5S72YAnQ Long Version (Just more explaination and going through each slide in detail, can see clearly): http://www.soundspawn.com/bsid/balancer_vid.zip Powerpoint File: http://www.soundspawn.com/bsid/balancer_ppt.zip
wow will we always get so detailed descriptions now. i feel insecure, is this really empires? :D if you somehow manage to get it right im sure it might become a good system - will players be able to see their points btw?
Yes, sm_ratings in console (or !ratings or /ratings in chat) will print all the ratings of the players currently connected to console. Splits it out by team and shows team totals.
Ow...18 min video takes me hours to watch due to this madhouse. I skipped through the last 5. Still, good stuff
One problem you have is: what happens when the smaller team starts to lose. They may even have been the overwhelmingly better team, but as soon as the system creates a 9 vs 12 and the 9 lose, they will blame the system. The best idea i've seen so far is the time played = ability one.
The smaller team should have a challenge, meaning they shouldn't always win. That's not a problem, that's a feature. Also, if you don't want to lose 9v12, then maybe you shouldn't have put the 9 best players on the same team. See, stack != win with this, stack = harder game. If you want a fair game, don't stack and it will be even team size and team skill. Using just the time played is way to simplistic. Let's say I implement that right now... we'd all be at zero hours, so skilled players could stack the shit out of games, and as long as they didn't log as many hours as average players (say stack nonstop but only for 4 hours a week), they would always be "under skilled" and not only be allowed but downright encouraged (forced even) to stack because the system would say "not many hours, not much ability". That would backfire so hard...
It's true, hours played = ability wouldn't cover the 1000's of hours put into empires or other FPS' before it was implemented, so wouldn't neccessarily work. The point I was making was that if the 9 players lose, they'll feel like the system ripped them off by making them fight with 9 players. They won't think "we were the stacked team, we just lost this one guys", they'll think "The damn system got it wrong. As evidenced by the last game, we were not the superior team... empires is dumb". If you don't calculate ability by points per unit time (which favours people who do points heavy roles (engineers and some riflemen) and doesn't do justice to others (commanders and grenadiers)), and you don't do it by time played, as we don't have enough information for that, then I don't see how you can judge it. Is it worth the coding if in the end people might just say "the 9 vs 12 system sucks". Actually, there's only one place where the 9 vs 12 situation is good. That's when the 9 is winning. Perhaps something not to difficult to code but useful to implement would be that when one team has +2 refs compared to the other team (would need to factor in those refs that give out 2x), then players can only join the team with less refs. So even if the match is even teamed, 10 v 10, a player that joins is forced to join the losing team, which would balance things out. If another player joins, he also joins the losing team. Even if players leave the winning team, as long as the winning team has +2 refs, new players would join the losing team. Hopefully the stack would work itself out. That has the added benefit that players that join half way have to join the losing team. as the stacked team is almost always going to win the opening battle, that means that people that would LIKE to stack can't, because the stacked team will already have +2 refs.
I'm tired of defending the system I made from speculation, I'm ready to put it on more servers and prove it works (it can monitor ratings without forcing teams for proof of concept). I can however quite easially explain why these other ideas won't work, so I will continue to do so The ref idea, two things will happen. 1) the winning team will recycle a ref or two whenever they want someone from spec. "Hey bro, hang on, lemme recycle two refs and you can join" And 2) players will sit in spec and wait for the match to end because they don't want to join the doomed team. Put these factors together, and the winning team will just get more and more stacked, but the numbers will be about even. This also doesn't address stacking early in the game, when both teams have equal refs but the skill is visibly lopsided.
this sounds lovely soundspawn, if its mechanic works like i think it works form your video it will work great
I only watched the short version, but it seemed very generic. It could have been about any system. I'll watch the long version later.
i has solution. put the mighty [R]eX on losing team, stax is reversed. oh wait, deadmod must go play HoN bye.
after watching the vid, i like it a lot, adding it in 2.25 would be a good test provided that the devs are willing to address it (remove/tweak) if it is obviously flawed. +1 though
I walked Trickster through it over steam and I think he likes it a bit too now. I think I made the common mistake of not explaining how "not a big deal" it really is in terms of forcing teams. So it's a plugin right now and it sounds like it should be going on G4TC and ELC as well as already being on House of Rage... at this time I believe G4TC and ELC will just be monitoring where it ranks every one, but it's just a matter of changing a cvar to enforce teams so we should be able to get some testing in when the players connected are in the mood for some number stacked games. Once it's tweaked right, we can talk about getting it hard coded (as one balancing option, an alternative to mp_autoteambalance but not the only mechanism, remember, not a tyrant).
Ounin, that was one of the things I argued out with Spawwn and he's more than likely going to add it. There were a few "features" I requested which are being considered for addition, as well as building on it later. I won't lie, I still think the videos suck a bit, none of them explained half the stuff you explained to me, but either way, because we'll have it stat track for a while before it goes live, we can see the issues there anyway. Above all, I asked for tracking once people have left, so they can't just bail 5 seconds before the game ends to save their points. I think Spawwn is looking into that to prevent it. Likewise, hopefully it can at some point track turning points in games (no refs + winning?) so that vets will still have a reason to join that losing side and bring it back from the brink.
