Naggy, i'm not going to quote all the different things you said, but you basically started by saying its impossible to balance skill on a server. I have no clue how that led to the rest of your 'points', but this is what i'll argue against.
Perhaps you didn't read the first post or any of the thread. But what i was referring to with 'balancing skill on a server' is using the skill-ranking tool that the devs said they may create. Instead of using it to choose which server would best suit a player, I thought it could be applied actually in-game, in the readyroom, to give an up-to-date estimate on the balance of the skill on both teams.
This of course assumes that this skill ranking system exists. To create such a system would be difficult, but thats up to the devs. Here we assume it will exist.
If it does exist, then it could tally each team's skill in numerical form and compare it to the other team. If this was done before the teams were set, you could actually balance them out at that time too. People could see an estimated stack, and try to fix it. Its not impossible at all if you assume the devs actually make the skill ranking system they mentioned.
But naggy brings up a good point against your idea, Stix. There are two <b>completely different teams</b>. It wouldn't work. You can't really compare skill as an alien to skill as a marine.
And also what about the players that 'earned their skill' playing on marines or aliens, but are forced to play the other team - one they're not as proficient at? Obviously your team balancing goes to ###### then.
Of course if you had two identical teams, such as in a marines vs marine or aliens vs aliens mode, switching players between teams based on skill might work.
Switching players between teams might not be necessary, NS has the capability to fine tune balance in a much subtler way, by scaling the income of resources to counter any difference in skill (this does reley on the low skill team knowing enough about NS to be able to spend the extra res).
About matchmaking: (i will shortly go back on-topic later)
Why not just some sort of "newbie flag", heh. Sort of like in Ultima Online, only it would obliviously not work the same way and disappear after a set number of playtime hours... It could provide some sort of short-time spawn protection to new players, tips (like tell him how to kill a fade, just after he died by a weapon_swipe) etc... Things like that, just throwing ideas there. It could even have some social impact if an icon is publicly displayed in the scoreboard, for example. "Oh, a new player, let's help him learn the game!"
Sort of like the defunct guide program, but inverted! <img src="style_emoticons/<#EMO_DIR#>/tounge.gif" style="vertical-align:middle" emoid=":p" border="0" alt="tounge.gif" />
The main deal here is not to drive away new players, right? Well i think it could help. Ideally there would be no need for a matchmaking system anymore... It would sucks anyway cause the only way to improve is to play against better players, in my opinion, and like it's been mentioned already in this thread (a bit harshly maybe... well... )
About balancing skill on a server between two completely different teams: I believe it's just simply impossible or... (pick one)
- overly-complicated - not so accurate - unnecessary - too much work - not worth it...
It could even be frustrating, including for new players...
Just my opinion <img src="style_emoticons/<#EMO_DIR#>/smile-fix.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile-fix.gif" />
lol as far as being impossible to skill rate a marine player and a alien player, it would have to be seperate.... for instance
A person plays on aliens all the time, has god like skills on aliens, so hes ranked as elite in aliens. But as far as marine goes he has no KNOWN skill to the server(s) so hes ranked as a newbie marine. It is a little complicated but still thats how i picture that in my mind.
"- not worth it... " It is worth it, but you shouldent force players to fit into their rank, there should be a flag raised like when a player joins the game giving you the approximate skill level of the player.
"- not so accurate" Id have to agree on you with that only slightly, but for my thoughts on how the balanceing skill would protain into allowing everyone to go into any server regardless of level, just a flag coming with them showing their "Approximate" skill level on seperate sides
which digs into your other argument
"- overly-complicated"
Your not doing the work lol! But... your right, it is complicated, but are you saying Unknownworlds cant whip up a skill notifyer? I still would love to argue the fact that some players need to know what they are up against, a pack of newbs, or a pack of newbs with a pub god (really good player) Its not THAT hard if you want to be pinpointing accuracy, but then again i have no idea how to program anything outside of byond.
<!--quoteo(post=1674341:date=Mar 26 2008, 12:48 PM:name=locallyunscene)--><div class='quotetop'>QUOTE(locallyunscene @ Mar 26 2008, 12:48 PM) <a href="index.php?act=findpost&pid=1674341"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->That's the point, this thread has <i>nothing</i> to do with making this game more like TF2. That's why I quoted that paragraph.<!--QuoteEnd--></div><!--QuoteEEnd--> Guess my hate for TF2 got the best of me.
<!--quoteo(post=1674341:date=Mar 26 2008, 12:48 PM:name=locallyunscene)--><div class='quotetop'>QUOTE(locallyunscene @ Mar 26 2008, 12:48 PM) <a href="index.php?act=findpost&pid=1674341"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Just because you haven't played it doesn't mean it doesn't exist and it doesn't work. For example there is an indie teamplay RTS with the unfortunate name "Trash" that had an auto-balance feature based on W-L and games played. This wouldn't be much different than K-D and minutes played. It worked well enough and not every server had it on, but most did because it was convenient and useful.<!--QuoteEnd--></div><!--QuoteEEnd--> As some people have said before me, (I'm assuming) NS2 will have totally different teams with two distinct play styles like its predecessor. By introducing auto team balancing your doing nothing but hurting people because as the players become better they lose more and more of their priority. For instance, someone new to NS2 might prefer playing on the Kharaa and would like to spend most of their playtime perfecting their skills and learning new things. This would be impossible with your proposal due to the fact that depending on everyone else in the server they are placed into teams depending on how good they are. Basically it's punishing the people who invest their time into the game and rewards the people who are new and have no idea what they are doing, and then when they do know what they are doing they will become the ones who are punished.
Also: Stix, I'm not sure that making a skill ranking system would be all that hard (atleast I think).. More time spent playing = better rank/icon. Just giving the player a little icon besides their name on the scoreboard would be enough (look at the WoW rankings/icons for example: <a href="http://www.worldofwarcraft.com/pvp/ranks.html" target="_blank">http://www.worldofwarcraft.com/pvp/ranks.html</a>.) Also, players could have separate rankings/icons for each team, for example; good ranking on aliens, crap ranking on marines. People would easily be able to distinguish the newer players from the better ones and it would help newer players get the information/help they need because they can just ask someone with a higher ranking instead of spamming "WATS THIS DEW?"
locallyunsceneFeeder of TrollsJoin Date: 2002-12-25Member: 11528Members, Constellation
edited March 2008
<!--quoteo(post=1674410:date=Mar 27 2008, 08:34 AM:name=naggy)--><div class='quotetop'>QUOTE(naggy @ Mar 27 2008, 08:34 AM) <a href="index.php?act=findpost&pid=1674410"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->As some people have said before me, (I'm assuming) NS2 will have totally different teams with two distinct play styles like its predecessor. By introducing auto team balancing your doing nothing but hurting people because as the players become better they lose more and more of their priority. For instance, someone new to NS2 might prefer playing on the Kharaa and would like to spend most of their playtime perfecting their skills and learning new things. This would be impossible with your proposal due to the fact that depending on everyone else in the server they are placed into teams depending on how good they are. Basically it's punishing the people who invest their time into the game and rewards the people who are new and have no idea what they are doing, and then when they do know what they are doing they will become the ones who are punished.<!--QuoteEnd--></div><!--QuoteEEnd--> Ah I see you point about two different teams now. I personally learn best by playing both teams at about the same skill level. I find it much easier to counter other players' movements then. However, I think you're right that some people could just learn one team and that skill wouldn't translate when they joined the other team. The two different rankings as you suggested seems the most obvious solution.
<!--quoteo(post=1674389:date=Mar 27 2008, 02:48 AM:name=Harimau)--><div class='quotetop'>QUOTE(Harimau @ Mar 27 2008, 02:48 AM) <a href="index.php?act=findpost&pid=1674389"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->But naggy brings up a good point against your idea, Stix. There are two <b>completely different teams</b>. It wouldn't work. You can't really compare skill as an alien to skill as a marine.
And also what about the players that 'earned their skill' playing on marines or aliens, but are forced to play the other team - one they're not as proficient at? Obviously your team balancing goes to ###### then.<!--QuoteEnd--></div><!--QuoteEEnd-->
Well, there are two ways it could work given the two completely different teams (which I already assumed for NS2).
1) The skill ranking system is global and just results in a number. Its not quite as accurate, but it makes the whole ordeal a bit easier. You can be the perfect marine and the worst alien, and your rank will be directly 50% of the max. Of course, that would be an almost impossible situation. Anyone who has played that much marine has definitely played at least a good amount of alien, to know what they are doing and to do well enough on the team. I'd guess that the extremes of skill would be at most 75/25, but even that would be very rare.
2) Or, you could have different numbers for each team. This would complicate the process a bit, but not by much. If a player is great on marines and the marines have a 70-30 advantage on the skill ranking scoreboard in the Ready Room, the player would leave the marines and you'd see the advantage drop. When he joins the aliens, the aliens' advantage wouldn't pop up as much as it would for the marines, since it uses the alien skillrank number instead.
So no, it doesn't ruin the whole idea, simply because the two teams are different. Also consider that should such a system be implemented, on the in-game TAB scoreboard, perhaps you could put your cursor over a player to see his skill rank for each team. That way, in the RR, players could see each other's ranks and decide who to ask to switch.
Building on Stix' idea but going in a slightly different direction, Imagine this:
Each player has a "threat rating" which is calculated separately for each team. It starts at 0% for a new player and rises towards 100% as you become more skilled and rack up more wins. Method for data collection is open to improvement but would probably be primarily based on win %, for example threat rating = number of games won of your last 100 games, with players under 100 games considered to have lost the rest of them so they correctly start at 0.
Even though the game (either your own NS2 install, or a global stat tracker) knows this rating as a 0 to 100 number, players never see that -- what players see when looking at you on the scoreboard or when you join the server is a small graphic with a vertical color bar that is blue at the bottom and red at the top, and a small skulk icon on the left side and a small marine icon on the right side positioned higher or lower depending on what your threat rating is. So if you have a high alien threat rating but a low marine threat rating, the skulk will be near the top of the red area while the marine icon will be lower into the blue area. This gives players an approximate idea of your skill level while preventing anyone from obsessing over 1% movements in rating.
When a new game is starting, during the time between the first player joining a team and the actual start of the game all players can see a similar graphic giving the total average threat rating of each team. This will not stop you from starting an unbalanced game, but you'll at least get a heads up on it if you see the skulk icon hovering in the blue while the marine icon is high in the red 10 seconds before the game starts. If the unbalance is bad enough you can ask players to switch.
To illustrate, I have attached a simple mockup of my threat rating:
I don't know if this has been said (again) yet, but it's really easy to quantify skill on marine vs skill on alien. You playtest until you have a good understanding of the difficulties of playing each class (based on an objective scale with objective "anchor" reference points) and then quantify the difficulty of the class based on experience and apply that as a multiplier (from 0-1) against the kills made as that class.
Yeah... from general knowledge, you'd think that 1.3 marine kills equals roughly 1 skulk kill (as in, its easier to kill as a marine).
But even that isn't really right. Any alien player who has any amount of patience and map knowledge, and will actually ambush, can easily get a bunch of kills.
That doesn't change the fact that it's stupidly <b>tedious</b>. I'm sure most people would prefer to jump straight into the game. "Balancing" is the job of the developers, not the players. And it's also restrictive. What if you wanna play with your friends on the same team, or against them, on separate teams? Or you completely suck at aliens, and want to play marines, but can't because your marine rank is too high? Essentially you have less control over which team you want to play - the later you make a choice.
Cxwf's idea is better and more practical, because it doesn't force players to balance the game (though it may still cause arguments when players absolutely want the game to be balanced, though it'll be more of an approximation), it just gives visual indicators of the threat level (I like that term, threat level - ###### "skill", we've got "threat") of individual players. I'm not sure however, if there's a "start button" in this idea, because if so.. that would be tedious; or if it starts when there are a certain number of players that have chosen a team in the game. (Random teamers wouldn't really be random, they'd be placed in order to best balance the teams.) However, I'm still all for rank-restricted, no, threat-restricted servers; so we could incorporate this idea into that one.
But there's also, well.. Synergy - the whole is greater than the sum of the parts. I'm averse to using the 'teamwork rules all' argument, so just to make sure you know, this isn't -that- argument. This is more about numbers. Three marines is three guns and three targets to choose from.
1. It is the developer's job to create the game, as well as listen to any reasonable balance suggestions. However, as far as the players go, the devs can only analyze trends of balance issues, not balance the players for the individual servers. How players play the game on individual servers are up to the individual admins to take care of.
2. The developers can only balance the games, and not the players. If a couple of pros are going to join games, then they are going to join games and most likely stack teams. This is prevalent in every other FPS in existence. Individual admins can and SHOULD regulate this, but in all honesty, stacking should not be a concern for the developers, as there is little they can do about it.
Any system involving skill points or threat levels or such would be extremely difficult to implement, with a very little return of actual results.
After reading the three threads about this, and fantasizing about some perfect balancing system that would solve all my woes and puree my tomatoes, I have come to agree with Firewater's conclusion.
locallyunsceneFeeder of TrollsJoin Date: 2002-12-25Member: 11528Members, Constellation
<!--quoteo(post=1674632:date=Mar 29 2008, 02:23 PM:name=Firewater)--><div class='quotetop'>QUOTE(Firewater @ Mar 29 2008, 02:23 PM) <a href="index.php?act=findpost&pid=1674632"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Hey all, I am back from the grave.
I have a few points on this.
1. It is the developer's job to create the game, as well as listen to any reasonable balance suggestions. However, as far as the players go, the devs can only analyze trends of balance issues, not balance the players for the individual servers. How players play the game on individual servers are up to the individual admins to take care of.
2. The developers can only balance the games, and not the players. If a couple of pros are going to join games, then they are going to join games and most likely stack teams. This is prevalent in every other FPS in existence. Individual admins can and SHOULD regulate this, but in all honesty, stacking should not be a concern for the developers, as there is little they can do about it.
Any system involving skill points or threat levels or such would be extremely difficult to implement, with a very little return of actual results.<!--QuoteEnd--></div><!--QuoteEEnd--> I agree this is not what FPS games do currently, but this is an area that NS2 could borrow from the RTS world to improve its gameplay. Autobalancing is never mandatory and is a very useful feature even if it doesn't work perfectly. The hard part of autobalance is getting an accurate skill measurement to begin with, but it is easily leveraged off of a match making system once one is in place. It would be easily leveraged off of any global skill ranking system.
Sarisel.::' ( O ) ';:-. .-.:;' ( O ) '::.Join Date: 2003-07-30Member: 18557Members, Constellation
edited March 2008
Having an indication of threat level + a few active admins to ensure that things aren't lopsided should be enough without a ready button or autoassigning woes. What might seem fair isn't necessarily fun and vice versa. Sometimes you have to sacrifice one for the other so that the game keeps its flavor. Be careful not to turn a game into a job, unless you're getting paid for playing it.
locallyunsceneFeeder of TrollsJoin Date: 2002-12-25Member: 11528Members, Constellation
<!--quoteo(post=1674793:date=Mar 31 2008, 10:43 PM:name=Sarisel)--><div class='quotetop'>QUOTE(Sarisel @ Mar 31 2008, 10:43 PM) <a href="index.php?act=findpost&pid=1674793"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Having an indication of threat level + a few active admins to ensure that things aren't lopsided should be enough without a ready button or autoassigning woes. What might seem fair isn't necessarily fun and vice versa. Sometimes you have to sacrifice one for the other so that the game keeps its flavor. Be careful not to turn a game into a job, unless you're getting paid for playing it.<!--QuoteEnd--></div><!--QuoteEEnd--> I don't understand what you mean. How would having an autoassign button for admins make the game feel like a job for me? If anything it would make the game less of a job for admins. It's a smarter "random-all" button.
If you mean that having a skill ranking would break immersion, I don't think it would. It doesn't have to be In-Your-Face regardless, it could be mostly behind the scenes. I haven't seen any "autoassigning woes" when it's not mandatory.
Sarisel.::' ( O ) ';:-. .-.:;' ( O ) '::.Join Date: 2003-07-30Member: 18557Members, Constellation
<!--quoteo(post=1674815:date=Apr 1 2008, 12:28 PM:name=locallyunscene)--><div class='quotetop'>QUOTE(locallyunscene @ Apr 1 2008, 12:28 PM) <a href="index.php?act=findpost&pid=1674815"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I don't understand what you mean. How would having an autoassign button for admins make the game feel like a job for me? If anything it would make the game less of a job for admins. It's a smarter "random-all" button.
If you mean that having a skill ranking would break immersion, I don't think it would. It doesn't have to be In-Your-Face regardless, it could be mostly behind the scenes. I haven't seen any "autoassigning woes" when it's not mandatory.<!--QuoteEnd--></div><!--QuoteEEnd-->
With respect to making the game feel like a job, I was referring more to players than admins. If you force balanced games, you are forcing people to play a certain way - which is very much what happens when you work, where you may be forced to do things a certain way. Personally, even if I wasn't playing with a group of friends, I wouldn't appreciate being sent to play with a group of people who I didn't choose to play with. It's not my idea of fun. If I was playing with a group of friends and we were all above the average skill level of the server, I'd appreciate it more if an admin acknowledged this politely and asked us to avoid stacking the teams (if we were). This would be better than a bunch of people whining about the game not starting due to unbalanced teams or being automatically segregated just because some skill-weighing system said this is the "fair" thing to do.
locallyunsceneFeeder of TrollsJoin Date: 2002-12-25Member: 11528Members, Constellation
You wouldn't want to play on servers that use it, I get that. I would however because I think games are much more fun when the teams are roughly equal.
<!--quoteo(post=1673928:date=Mar 22 2008, 11:24 AM:name=Radix)--><div class='quotetop'>QUOTE(Radix @ Mar 22 2008, 11:24 AM) <a href="index.php?act=findpost&pid=1673928"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Stix, locking would be optional, so a server admin could say "restrict this server to appropriate (meaning roughly similar) player skill" or they could just leave it off and let anyone join. The purpose of that is that at least some servers have "difficulty settings" that new players can filter through. That way if they do join an unfiltered server, they're going in with the mindset of "anything can happen".
Also, forming an algorithm to parse skill isn't as hard as you think. It's fairly accurate to just use K:D ratio, but you <b>do</b> have to contrast that K:D ratio with the amount of time in lifeform states.
For example, if a player spends 100% of his time with an LMG he should have a score multiplier in order to balance the 50% of the time his friend spent with a shotgun. Similarly, gorge kills should be worth much more than lerk or fade kills. This isn't to demean the player's ability to score with a fade, but to show the originally intended element - <!--coloro:lime--><span style="color:lime"><!--/coloro-->how good is this player, <i>despite</i> their current strengths and weaknesses?<!--colorc--></span><!--/colorc--> This could exclude the time spent building and killing structures so that it's a more pure statistic if you wanted to track it with a timescale as well.
Another idea that's been proposed presupposes that a lerk's goal is, rather than to get kills, to win games. Along that line of reasoning a commander, or a lerk (being support roles primarily) could be graded on either the outcome of the game (win or lose) or on how well their teammates do, with a multiplier for the skill of each of the teammates (where bad is good for the lerk or comm's score, and good is bad for it, in order to get an accurate result for the quality of the supporter).
You could add as many levels of detail to the algorithm as you wanted, but I still hold that it shouldn't be based on how much the player helps the team, but rather how skilled the player is at FPS techniques like movement and aim (which is quantified best by K:D ratio) because those are the elements that cause the game to become not-fun when they are paired against prohibitively more (or less) difficult opponents.<!--QuoteEnd--></div><!--QuoteEEnd-->
I disagree with this entire system because of the nature of NS. In NS the players decide when the game ends, not the game itsself. Allow me to explain.
In most other FPS there is a time limit to which the game abides by. In CS there is a timer in which the round will end when it reaches 0, and the defensive team will earn a round win. Same for CoD DoD etc...
In NS however, the players choose when the game ends. Marines and aliens can spawn camp until the cow comes home, and if they do not destroy the hive(s) or IPs respectively, members on both sides can rack up a very high K:D ratio without much effort at all. Even with a weighting system, the nature of the game prevents a true statistical way to balance.
Also, if someone gets a high K:D ratio in one server, it does not neccesarily mean that those same results will generalize to other servers. Back when NS was in its prime, it was popular to have very competitive pub servers as well as more casual ones. A casual player can do well on their servers, and then use that same K:D ratio score over in a pro server. The odds are that individual will be a statistical anomaly for whichever side he or she chooses, providing a negative bias to his or her own team. Conversely, an average player on those same pro servers, may connect to a casual server and be rated lower yet perform better given the amount of competition available.
Imagine a situation like this for 16 different people.
A system like this is really ineffective. I personally believe its up to the server admins to determine whether or not team stacking is an issue on his or her own individual server. I do not see a logical way that this system could work given the seemingly infinite amount of variables to consider.
locallyunsceneFeeder of TrollsJoin Date: 2002-12-25Member: 11528Members, Constellation
<!--quoteo(post=1674834:date=Apr 1 2008, 03:13 PM:name=Firewater)--><div class='quotetop'>QUOTE(Firewater @ Apr 1 2008, 03:13 PM) <a href="index.php?act=findpost&pid=1674834"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->In NS however, the players choose when the game ends. Marines and aliens can spawn camp until the cow comes home, and if they do not destroy the hive(s) or IPs respectively, members on both sides can rack up a very high K:D ratio without much effort at all. Even with a weighting system, the nature of the game prevents a true statistical way to balance.<!--QuoteEnd--></div><!--QuoteEEnd--> Again any sufficiently determined player can BS/workaround achievements, requirements, etc. in any game. This doesn't make stat systems invalid. Those players that spawn camp, whatever will get destroyed when they join a real game. Autobalance and matchmaking aren't substitutes for admins, they're just useful tools.
<!--quoteo(post=1674834:date=Apr 1 2008, 03:13 PM:name=Firewater)--><div class='quotetop'>QUOTE(Firewater @ Apr 1 2008, 03:13 PM) <a href="index.php?act=findpost&pid=1674834"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->In NS however, the players choose when the game ends. Marines and aliens can spawn camp until the cow comes home, and if they do not destroy the hive(s) or IPs respectively, members on both sides can rack up a very high K:D ratio without much effort at all. Even with a weighting system, the nature of the game prevents a true statistical way to balance.<!--QuoteEnd--></div><!--QuoteEEnd-->
In my posts on matchmaking I've referred to a system of weighting whereby the player's advantages are factored into the game. It's very simple logic to query whether or not the marines have a functioning tech tree or if the aliens are being spawncamped (ie, how many players are dead, how many upgrades the team has, and even how quickly players on a given team are dying vs the opposing team's death rate). When you weight the game effectively it solves this problem. It's possible (probable) that the weights will be broken at times by players determined to break them, but I think it's much better to implement a cohesive system than to ignore the problem and argue that it's a waste of time to even try, merely because it has a subset of shortcomings that would have to be dealt with by server admins manually.
<!--quoteo(post=1674834:date=Apr 1 2008, 03:13 PM:name=Firewater)--><div class='quotetop'>QUOTE(Firewater @ Apr 1 2008, 03:13 PM) <a href="index.php?act=findpost&pid=1674834"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Also, if someone gets a high K:D ratio in one server, it does not neccesarily mean that those same results will generalize to other servers. ... an average player on those same pro servers, may connect to a casual server and be rated lower yet perform better given the amount of competition available.<!--QuoteEnd--></div><!--QuoteEEnd-->
The point of matchmaking is to keep prohibitively different skill levels from playing against one another *at all*. This is not to say, restrict all new players from playing on intermediate servers, but certainly to keep totally-green fresh-buy-fresh-install players out of the advanced games, while keeping high level players out of starter games. As I've noticed, the vast majority of advanced players would rather leave the newbies alone and play with people who are interesting to fight with. There's no reason for either one of these groups to interact with each other until they get closer to each others' level. It isn't fun for anyone except griefers, who in my opinion don't even make up a large portion of the community.
Once you choose to lock the servers gauged to be prohibitive in skill level (too high or too low) the problem should work itself out, and if (for some reason) it doesn't you still have admins to fall back on, and if the devs decided to take it a step further, you could add a piece of math in the algorithm to look at the skill level of the server and adjust *all* kills and deaths that occur based on how tough it is. But to me that seems excessive because a precondition of reasonable skill is assumed when a player is playing on a server that they've already unlocked.
<!--quoteo(post=1674853:date=Apr 1 2008, 05:15 PM:name=locallyunscene)--><div class='quotetop'>QUOTE(locallyunscene @ Apr 1 2008, 05:15 PM) <a href="index.php?act=findpost&pid=1674853"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Again any sufficiently determined player can BS/workaround achievements, requirements, etc. in any game. This doesn't make stat systems invalid. Those players that spawn camp, whatever will get destroyed when they join a real game. Autobalance and matchmaking aren't substitutes for admins, they're just useful tools.<!--QuoteEnd--></div><!--QuoteEEnd-->
I have no problem with stats. If the developers want to create a leaderboard thats just fine. However, to use those stats given the different amount of variables given a WAY too simplistic system. I'm not suggesting that all players would work around. What I am suggesting is that the stats will be skewed simply based on when a player joins the game, the relative skill level at the time of joining a given server, as well as the momentum that the joining team has.
The stats system would have no statistical validity whatsoever.
a good scoring system will naturaly balance the teams. for instance if you have it be every person gets a number. When you score a kill you steal 2-10% of the persons you killed number adjusted for what class you are (more % for killing a HA/HMG then a LA or more for killing the onos then the skulks). The really good players will want to be on diffrent teams, because for every kill the guy with 10 points ges agasint the guy with 1000, he gets 20 extra points, while the 1000 only gets like 2 per kill. Clearly the guys with alot of points will not want to be all on the same team. Have it recorded and stored by steamID. No farming possible as just going to kill a bunch of bots doesnt do anything, and letting someone kill you over and over doestn help eather (they cant take what you dont have).
thsi is a fairly simple example of what could balnace the teams, a great scoring system will include these elements along with alot of other types of scores, such as acomplishmens and time played.
<!--quoteo(post=1674870:date=Apr 1 2008, 06:39 PM:name=obsid)--><div class='quotetop'>QUOTE(obsid @ Apr 1 2008, 06:39 PM) <a href="index.php?act=findpost&pid=1674870"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->a good scoring system will naturaly balance the teams. for instance if you have it be every person gets a number. When you score a kill you steal 2-10% of the persons you killed number adjusted for what class you are (more % for killing a HA/HMG then a LA or more for killing the onos then the skulks). The really good players will want to be on diffrent teams, because for every kill the guy with 10 points ges agasint the guy with 1000, he gets 20 extra points, while the 1000 only gets like 2 per kill. Clearly the guys with alot of points will not want to be all on the same team. Have it recorded and stored by steamID. No farming possible as just going to kill a bunch of bots doesnt do anything, and letting someone kill you over and over doestn help eather (they cant take what you dont have).
thsi is a fairly simple example of what could balnace the teams, a great scoring system will include these elements along with alot of other types of scores, such as acomplishmens and time played.<!--QuoteEnd--></div><!--QuoteEEnd--> As stated earlier, even with a weighting system, balancing the skill base from server to server is impossible. A skilled player on one server does not neccesarily mean a skilled player on another server.
Sarisel.::' ( O ) ';:-. .-.:;' ( O ) '::.Join Date: 2003-07-30Member: 18557Members, Constellation
edited April 2008
I posted a topic that supported stat collection a long time ago, although it was just a thought exercise. Now, if this is actually being considered, in order to make it *work* (ignoring the likelihood of statistical validity and assuming you can actually implement ways to measure environmental factors that play into determining the "value" of kills/deaths) exactly how much data are we looking at here to collect per player, per game, per server, per day - and how practical would it be for this amount of data, after processing in a particular server, to be distributed between all the servers?
I was also making another post in the other topic but lost most of it - so I might as well just add the bit I'm interested in here: "I suspect that NS2 is going to be made in a way that makes individual excellence much more limited than in NS1, in favour of teamwork (even though I might not agree with this). In this case, you'll have a bunch of average K:D stats that won't be of very much use to you - whereas individual contributions to teamwork, which will make and break victories, will not be measurable."
Yes, but you seem to be missing the point here. This (ranked servers) isn't as much about giving teams a roughly equal chance to win, but to give players a roughly equal chance to kill (other players) or die (to other players). You must understand that the majority of players in a server will be playing a First Person Shooter, not a Real Time Strategy. Obviously, assuming every player on the server has a similar twitch skill level, the smarter and more organised team will win. That's not only to be expected, but completely fair.
Sarisel.::' ( O ) ';:-. .-.:;' ( O ) '::.Join Date: 2003-07-30Member: 18557Members, Constellation
Ok, so it's completely fair if one team is stacked so that it has good teamwork while the other is completely inept? Unless the learning curve in NS2 remains as lopsided as NS1's, you're going to have teamwork trumping individual skill a lot more often than the reverse.
Also, balancing for skill on either team isn't going to accomplish your objective of interest (fairness in ability to kill others or die to others) - this can't be achieved via a simple average, because the skilled players on either team will still dominate unskilled opponents, unless your model assumes the game is played in a tight, short, L-shaped corridor with teams starting on either end...
Comments
Perhaps you didn't read the first post or any of the thread. But what i was referring to with 'balancing skill on a server' is using the skill-ranking tool that the devs said they may create. Instead of using it to choose which server would best suit a player, I thought it could be applied actually in-game, in the readyroom, to give an up-to-date estimate on the balance of the skill on both teams.
This of course assumes that this skill ranking system exists. To create such a system would be difficult, but thats up to the devs. Here we assume it will exist.
If it does exist, then it could tally each team's skill in numerical form and compare it to the other team. If this was done before the teams were set, you could actually balance them out at that time too. People could see an estimated stack, and try to fix it. Its not impossible at all if you assume the devs actually make the skill ranking system they mentioned.
And also what about the players that 'earned their skill' playing on marines or aliens, but are forced to play the other team - one they're not as proficient at? Obviously your team balancing goes to ###### then.
Switching players between teams might not be necessary, NS has the capability to fine tune balance in a much subtler way, by scaling the income of resources to counter any difference in skill (this does reley on the low skill team knowing enough about NS to be able to spend the extra res).
Why not just some sort of "newbie flag", heh. Sort of like in Ultima Online, only it would obliviously not work the same way and disappear after a set number of playtime hours... It could provide some sort of short-time spawn protection to new players, tips (like tell him how to kill a fade, just after he died by a weapon_swipe) etc... Things like that, just throwing ideas there. It could even have some social impact if an icon is publicly displayed in the scoreboard, for example. "Oh, a new player, let's help him learn the game!"
Sort of like the defunct guide program, but inverted! <img src="style_emoticons/<#EMO_DIR#>/tounge.gif" style="vertical-align:middle" emoid=":p" border="0" alt="tounge.gif" />
The main deal here is not to drive away new players, right? Well i think it could help. Ideally there would be no need for a matchmaking system anymore... It would sucks anyway cause the only way to improve is to play against better players, in my opinion, and like it's been mentioned already in this thread (a bit harshly maybe... well... )
About balancing skill on a server between two completely different teams: I believe it's just simply impossible or... (pick one)
- overly-complicated
- not so accurate
- unnecessary
- too much work
- not worth it...
It could even be frustrating, including for new players...
Just my opinion <img src="style_emoticons/<#EMO_DIR#>/smile-fix.gif" style="vertical-align:middle" emoid=":)" border="0" alt="smile-fix.gif" />
A person plays on aliens all the time, has god like skills on aliens, so hes ranked as elite in aliens. But as far as marine goes he has no KNOWN skill to the server(s) so hes ranked as a newbie marine. It is a little complicated but still thats how i picture that in my mind.
"- not worth it... "
It is worth it, but you shouldent force players to fit into their rank, there should be a flag raised like when a player joins the game giving you the approximate skill level of the player.
"- not so accurate"
Id have to agree on you with that only slightly, but for my thoughts on how the balanceing skill would protain into allowing everyone to go into any server regardless of level, just a flag coming with them showing their "Approximate" skill level on seperate sides
which digs into your other argument
"- overly-complicated"
Your not doing the work lol! But... your right, it is complicated, but are you saying Unknownworlds cant whip up a skill notifyer? I still would love to argue the fact that some players need to know what they are up against, a pack of newbs, or a pack of newbs with a pub god (really good player) Its not THAT hard if you want to be pinpointing accuracy, but then again i have no idea how to program anything outside of byond.
Guess my hate for TF2 got the best of me.
<!--quoteo(post=1674341:date=Mar 26 2008, 12:48 PM:name=locallyunscene)--><div class='quotetop'>QUOTE(locallyunscene @ Mar 26 2008, 12:48 PM) <a href="index.php?act=findpost&pid=1674341"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Just because you haven't played it doesn't mean it doesn't exist and it doesn't work. For example there is an indie teamplay RTS with the unfortunate name "Trash" that had an auto-balance feature based on W-L and games played. This wouldn't be much different than K-D and minutes played. It worked well enough and not every server had it on, but most did because it was convenient and useful.<!--QuoteEnd--></div><!--QuoteEEnd-->
As some people have said before me, (I'm assuming) NS2 will have totally different teams with two distinct play styles like its predecessor. By introducing auto team balancing your doing nothing but hurting people because as the players become better they lose more and more of their priority. For instance, someone new to NS2 might prefer playing on the Kharaa and would like to spend most of their playtime perfecting their skills and learning new things. This would be impossible with your proposal due to the fact that depending on everyone else in the server they are placed into teams depending on how good they are. Basically it's punishing the people who invest their time into the game and rewards the people who are new and have no idea what they are doing, and then when they do know what they are doing they will become the ones who are punished.
Also: Stix, I'm not sure that making a skill ranking system would be all that hard (atleast I think).. More time spent playing = better rank/icon. Just giving the player a little icon besides their name on the scoreboard would be enough (look at the WoW rankings/icons for example: <a href="http://www.worldofwarcraft.com/pvp/ranks.html" target="_blank">http://www.worldofwarcraft.com/pvp/ranks.html</a>.) Also, players could have separate rankings/icons for each team, for example; good ranking on aliens, crap ranking on marines. People would easily be able to distinguish the newer players from the better ones and it would help newer players get the information/help they need because they can just ask someone with a higher ranking instead of spamming "WATS THIS DEW?"
Ah I see you point about two different teams now. I personally learn best by playing both teams at about the same skill level. I find it much easier to counter other players' movements then. However, I think you're right that some people could just learn one team and that skill wouldn't translate when they joined the other team. The two different rankings as you suggested seems the most obvious solution.
And also what about the players that 'earned their skill' playing on marines or aliens, but are forced to play the other team - one they're not as proficient at? Obviously your team balancing goes to ###### then.<!--QuoteEnd--></div><!--QuoteEEnd-->
Well, there are two ways it could work given the two completely different teams (which I already assumed for NS2).
1) The skill ranking system is global and just results in a number. Its not quite as accurate, but it makes the whole ordeal a bit easier. You can be the perfect marine and the worst alien, and your rank will be directly 50% of the max. Of course, that would be an almost impossible situation. Anyone who has played that much marine has definitely played at least a good amount of alien, to know what they are doing and to do well enough on the team. I'd guess that the extremes of skill would be at most 75/25, but even that would be very rare.
2) Or, you could have different numbers for each team. This would complicate the process a bit, but not by much. If a player is great on marines and the marines have a 70-30 advantage on the skill ranking scoreboard in the Ready Room, the player would leave the marines and you'd see the advantage drop. When he joins the aliens, the aliens' advantage wouldn't pop up as much as it would for the marines, since it uses the alien skillrank number instead.
So no, it doesn't ruin the whole idea, simply because the two teams are different. Also consider that should such a system be implemented, on the in-game TAB scoreboard, perhaps you could put your cursor over a player to see his skill rank for each team. That way, in the RR, players could see each other's ranks and decide who to ask to switch.
Each player has a "threat rating" which is calculated separately for each team. It starts at 0% for a new player and rises towards 100% as you become more skilled and rack up more wins. Method for data collection is open to improvement but would probably be primarily based on win %, for example threat rating = number of games won of your last 100 games, with players under 100 games considered to have lost the rest of them so they correctly start at 0.
Even though the game (either your own NS2 install, or a global stat tracker) knows this rating as a 0 to 100 number, players never see that -- what players see when looking at you on the scoreboard or when you join the server is a small graphic with a vertical color bar that is blue at the bottom and red at the top, and a small skulk icon on the left side and a small marine icon on the right side positioned higher or lower depending on what your threat rating is. So if you have a high alien threat rating but a low marine threat rating, the skulk will be near the top of the red area while the marine icon will be lower into the blue area. This gives players an approximate idea of your skill level while preventing anyone from obsessing over 1% movements in rating.
When a new game is starting, during the time between the first player joining a team and the actual start of the game all players can see a similar graphic giving the total average threat rating of each team. This will not stop you from starting an unbalanced game, but you'll at least get a heads up on it if you see the skulk icon hovering in the blue while the marine icon is high in the red 10 seconds before the game starts. If the unbalance is bad enough you can ask players to switch.
To illustrate, I have attached a simple mockup of my threat rating:
But even that isn't really right. Any alien player who has any amount of patience and map knowledge, and will actually ambush, can easily get a bunch of kills.
That doesn't change the fact that it's stupidly <b>tedious</b>. I'm sure most people would prefer to jump straight into the game. "Balancing" is the job of the developers, not the players.
And it's also restrictive. What if you wanna play with your friends on the same team, or against them, on separate teams? Or you completely suck at aliens, and want to play marines, but can't because your marine rank is too high? Essentially you have less control over which team you want to play - the later you make a choice.
Cxwf's idea is better and more practical, because it doesn't force players to balance the game (though it may still cause arguments when players absolutely want the game to be balanced, though it'll be more of an approximation), it just gives visual indicators of the threat level (I like that term, threat level - ###### "skill", we've got "threat") of individual players. I'm not sure however, if there's a "start button" in this idea, because if so.. that would be tedious; or if it starts when there are a certain number of players that have chosen a team in the game. (Random teamers wouldn't really be random, they'd be placed in order to best balance the teams.)
However, I'm still all for rank-restricted, no, threat-restricted servers; so we could incorporate this idea into that one.
But there's also, well.. Synergy - the whole is greater than the sum of the parts.
I'm averse to using the 'teamwork rules all' argument, so just to make sure you know, this isn't -that- argument.
This is more about numbers. Three marines is three guns and three targets to choose from.
I have a few points on this.
1. It is the developer's job to create the game, as well as listen to any reasonable balance suggestions. However, as far as the players go, the devs can only analyze trends of balance issues, not balance the players for the individual servers. How players play the game on individual servers are up to the individual admins to take care of.
2. The developers can only balance the games, and not the players. If a couple of pros are going to join games, then they are going to join games and most likely stack teams. This is prevalent in every other FPS in existence. Individual admins can and SHOULD regulate this, but in all honesty, stacking should not be a concern for the developers, as there is little they can do about it.
Any system involving skill points or threat levels or such would be extremely difficult to implement, with a very little return of actual results.
I have a few points on this.
1. It is the developer's job to create the game, as well as listen to any reasonable balance suggestions. However, as far as the players go, the devs can only analyze trends of balance issues, not balance the players for the individual servers. How players play the game on individual servers are up to the individual admins to take care of.
2. The developers can only balance the games, and not the players. If a couple of pros are going to join games, then they are going to join games and most likely stack teams. This is prevalent in every other FPS in existence. Individual admins can and SHOULD regulate this, but in all honesty, stacking should not be a concern for the developers, as there is little they can do about it.
Any system involving skill points or threat levels or such would be extremely difficult to implement, with a very little return of actual results.<!--QuoteEnd--></div><!--QuoteEEnd-->
I agree this is not what FPS games do currently, but this is an area that NS2 could borrow from the RTS world to improve its gameplay. Autobalancing is never mandatory and is a very useful feature even if it doesn't work perfectly. The hard part of autobalance is getting an accurate skill measurement to begin with, but it is easily leveraged off of a match making system once one is in place. It would be easily leveraged off of any global skill ranking system.
I don't understand what you mean. How would having an autoassign button for admins make the game feel like a job for me? If anything it would make the game less of a job for admins. It's a smarter "random-all" button.
If you mean that having a skill ranking would break immersion, I don't think it would. It doesn't have to be In-Your-Face regardless, it could be mostly behind the scenes. I haven't seen any "autoassigning woes" when it's not mandatory.
If you mean that having a skill ranking would break immersion, I don't think it would. It doesn't have to be In-Your-Face regardless, it could be mostly behind the scenes. I haven't seen any "autoassigning woes" when it's not mandatory.<!--QuoteEnd--></div><!--QuoteEEnd-->
With respect to making the game feel like a job, I was referring more to players than admins. If you force balanced games, you are forcing people to play a certain way - which is very much what happens when you work, where you may be forced to do things a certain way. Personally, even if I wasn't playing with a group of friends, I wouldn't appreciate being sent to play with a group of people who I didn't choose to play with. It's not my idea of fun. If I was playing with a group of friends and we were all above the average skill level of the server, I'd appreciate it more if an admin acknowledged this politely and asked us to avoid stacking the teams (if we were). This would be better than a bunch of people whining about the game not starting due to unbalanced teams or being automatically segregated just because some skill-weighing system said this is the "fair" thing to do.
Also, forming an algorithm to parse skill isn't as hard as you think. It's fairly accurate to just use K:D ratio, but you <b>do</b> have to contrast that K:D ratio with the amount of time in lifeform states.
For example, if a player spends 100% of his time with an LMG he should have a score multiplier in order to balance the 50% of the time his friend spent with a shotgun. Similarly, gorge kills should be worth much more than lerk or fade kills. This isn't to demean the player's ability to score with a fade, but to show the originally intended element - <!--coloro:lime--><span style="color:lime"><!--/coloro-->how good is this player, <i>despite</i> their current strengths and weaknesses?<!--colorc--></span><!--/colorc--> This could exclude the time spent building and killing structures so that it's a more pure statistic if you wanted to track it with a timescale as well.
Another idea that's been proposed presupposes that a lerk's goal is, rather than to get kills, to win games. Along that line of reasoning a commander, or a lerk (being support roles primarily) could be graded on either the outcome of the game (win or lose) or on how well their teammates do, with a multiplier for the skill of each of the teammates (where bad is good for the lerk or comm's score, and good is bad for it, in order to get an accurate result for the quality of the supporter).
You could add as many levels of detail to the algorithm as you wanted, but I still hold that it shouldn't be based on how much the player helps the team, but rather how skilled the player is at FPS techniques like movement and aim (which is quantified best by K:D ratio) because those are the elements that cause the game to become not-fun when they are paired against prohibitively more (or less) difficult opponents.<!--QuoteEnd--></div><!--QuoteEEnd-->
I disagree with this entire system because of the nature of NS. In NS the players decide when the game ends, not the game itsself. Allow me to explain.
In most other FPS there is a time limit to which the game abides by. In CS there is a timer in which the round will end when it reaches 0, and the defensive team will earn a round win. Same for CoD DoD etc...
In NS however, the players choose when the game ends. Marines and aliens can spawn camp until the cow comes home, and if they do not destroy the hive(s) or IPs respectively, members on both sides can rack up a very high K:D ratio without much effort at all. Even with a weighting system, the nature of the game prevents a true statistical way to balance.
Also, if someone gets a high K:D ratio in one server, it does not neccesarily mean that those same results will generalize to other servers. Back when NS was in its prime, it was popular to have very competitive pub servers as well as more casual ones. A casual player can do well on their servers, and then use that same K:D ratio score over in a pro server. The odds are that individual will be a statistical anomaly for whichever side he or she chooses, providing a negative bias to his or her own team. Conversely, an average player on those same pro servers, may connect to a casual server and be rated lower yet perform better given the amount of competition available.
Imagine a situation like this for 16 different people.
A system like this is really ineffective. I personally believe its up to the server admins to determine whether or not team stacking is an issue on his or her own individual server. I do not see a logical way that this system could work given the seemingly infinite amount of variables to consider.
Again any sufficiently determined player can BS/workaround achievements, requirements, etc. in any game. This doesn't make stat systems invalid. Those players that spawn camp, whatever will get destroyed when they join a real game. Autobalance and matchmaking aren't substitutes for admins, they're just useful tools.
In my posts on matchmaking I've referred to a system of weighting whereby the player's advantages are factored into the game. It's very simple logic to query whether or not the marines have a functioning tech tree or if the aliens are being spawncamped (ie, how many players are dead, how many upgrades the team has, and even how quickly players on a given team are dying vs the opposing team's death rate). When you weight the game effectively it solves this problem. It's possible (probable) that the weights will be broken at times by players determined to break them, but I think it's much better to implement a cohesive system than to ignore the problem and argue that it's a waste of time to even try, merely because it has a subset of shortcomings that would have to be dealt with by server admins manually.
<!--quoteo(post=1674834:date=Apr 1 2008, 03:13 PM:name=Firewater)--><div class='quotetop'>QUOTE(Firewater @ Apr 1 2008, 03:13 PM) <a href="index.php?act=findpost&pid=1674834"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Also, if someone gets a high K:D ratio in one server, it does not neccesarily mean that those same results will generalize to other servers. ... an average player on those same pro servers, may connect to a casual server and be rated lower yet perform better given the amount of competition available.<!--QuoteEnd--></div><!--QuoteEEnd-->
The point of matchmaking is to keep prohibitively different skill levels from playing against one another *at all*. This is not to say, restrict all new players from playing on intermediate servers, but certainly to keep totally-green fresh-buy-fresh-install players out of the advanced games, while keeping high level players out of starter games. As I've noticed, the vast majority of advanced players would rather leave the newbies alone and play with people who are interesting to fight with. There's no reason for either one of these groups to interact with each other until they get closer to each others' level. It isn't fun for anyone except griefers, who in my opinion don't even make up a large portion of the community.
Once you choose to lock the servers gauged to be prohibitive in skill level (too high or too low) the problem should work itself out, and if (for some reason) it doesn't you still have admins to fall back on, and if the devs decided to take it a step further, you could add a piece of math in the algorithm to look at the skill level of the server and adjust *all* kills and deaths that occur based on how tough it is. But to me that seems excessive because a precondition of reasonable skill is assumed when a player is playing on a server that they've already unlocked.
I have no problem with stats. If the developers want to create a leaderboard thats just fine. However, to use those stats given the different amount of variables given a WAY too simplistic system. I'm not suggesting that all players would work around. What I am suggesting is that the stats will be skewed simply based on when a player joins the game, the relative skill level at the time of joining a given server, as well as the momentum that the joining team has.
The stats system would have no statistical validity whatsoever.
thsi is a fairly simple example of what could balnace the teams, a great scoring system will include these elements along with alot of other types of scores, such as acomplishmens and time played.
thsi is a fairly simple example of what could balnace the teams, a great scoring system will include these elements along with alot of other types of scores, such as acomplishmens and time played.<!--QuoteEnd--></div><!--QuoteEEnd-->
As stated earlier, even with a weighting system, balancing the skill base from server to server is impossible. A skilled player on one server does not neccesarily mean a skilled player on another server.
I was also making another post in the other topic but lost most of it - so I might as well just add the bit I'm interested in here:
"I suspect that NS2 is going to be made in a way that makes individual excellence much more limited than in NS1, in favour of teamwork (even though I might not agree with this). In this case, you'll have a bunch of average K:D stats that won't be of very much use to you - whereas individual contributions to teamwork, which will make and break victories, will not be measurable."
Also, balancing for skill on either team isn't going to accomplish your objective of interest (fairness in ability to kill others or die to others) - this can't be achieved via a simple average, because the skilled players on either team will still dominate unskilled opponents, unless your model assumes the game is played in a tight, short, L-shaped corridor with teams starting on either end...