How to Balance Skill on a server
StixNStonz
Join Date: 2006-11-06 Member: 58439Members, Reinforced - Shadow
Join Date: 2006-11-06 Member: 58439Members, Reinforced - Shadow
<div class="IPBDescription">a response to the 'less elitist gameplay' thread</div>Right about <a href="http://www.unknownworlds.com/forums/index.php?showtopic=103976&st=0&p=1673722&#entry1673722" target="_blank"><u>here</u></a> in the Less Elitist Gameplay thread, the dev's changed the topic from dumbing down the game, to balancing player skill against each other. They talked about matchmaking and such. My reply turned into this.
I can't imagine how you could possibly make matchmaking actually work to segregate servers.
A far more feasible plan, as far as I can envision it at least, is to have the skill ranking like you say but use it to keep the teams balanced within a server, rather than keeping better and worse players on separate servers.
Matchmaking only works if you don't allow players to manually choose their server to join. Which I think could only possibly work on consoles, and probably rarely at that.
How about this. A map loads, everyone is in the Readyroom. No one joins a team at first, they just go into the areas for each respective team (instead of the current norm of using doorways, you'd have something more akin to platforms). If you create this skill ranking system, it could be used to show two numbers on the wall, one for the % strength of each team compared to each other. Perfectly balanced would say '50-50' on the wall. This way, you could see the server's best guess at the balance. Then, players could switch sides (still in the RR, by just walking to the other side) to see the balance change.
To begin the game, perhaps there could be a button in the areas of each team, which is the 'go' button. It is not enabled until the balance is at least 40/60, and can only be pressed by the lower team. As such, the lower team could wait for better balance, or could choose to just play anyways (since a lot of people wouldn't actually believe the skill rankings, or at least see it as a challenge). You could also have a 5 second timer once the Go button is pressed, so other players could stop it if it was pressed by accident or by sabotage (by a dumba** on the lower team).
You could really do a lot with this idea. Perhaps you could add in triggers into the teambalance entities that can send an output for each 10% interval. As in, a level designer could make his own custom balance images to use instead of the standard percentages. Imagine seeing an infant vs an onos (20% vs 80%) or a Heavy versus an egg (80% vs 20%), or a Shotgun-Marine vs a Lerk (50-50), as an example. Or, for a much more basic image portrayal, you could have a bar chart. Etc.
This of course is just a preliminary idea, but I've tried to flesh it out as much as possible.
Thoughts?
I can't imagine how you could possibly make matchmaking actually work to segregate servers.
A far more feasible plan, as far as I can envision it at least, is to have the skill ranking like you say but use it to keep the teams balanced within a server, rather than keeping better and worse players on separate servers.
Matchmaking only works if you don't allow players to manually choose their server to join. Which I think could only possibly work on consoles, and probably rarely at that.
How about this. A map loads, everyone is in the Readyroom. No one joins a team at first, they just go into the areas for each respective team (instead of the current norm of using doorways, you'd have something more akin to platforms). If you create this skill ranking system, it could be used to show two numbers on the wall, one for the % strength of each team compared to each other. Perfectly balanced would say '50-50' on the wall. This way, you could see the server's best guess at the balance. Then, players could switch sides (still in the RR, by just walking to the other side) to see the balance change.
To begin the game, perhaps there could be a button in the areas of each team, which is the 'go' button. It is not enabled until the balance is at least 40/60, and can only be pressed by the lower team. As such, the lower team could wait for better balance, or could choose to just play anyways (since a lot of people wouldn't actually believe the skill rankings, or at least see it as a challenge). You could also have a 5 second timer once the Go button is pressed, so other players could stop it if it was pressed by accident or by sabotage (by a dumba** on the lower team).
You could really do a lot with this idea. Perhaps you could add in triggers into the teambalance entities that can send an output for each 10% interval. As in, a level designer could make his own custom balance images to use instead of the standard percentages. Imagine seeing an infant vs an onos (20% vs 80%) or a Heavy versus an egg (80% vs 20%), or a Shotgun-Marine vs a Lerk (50-50), as an example. Or, for a much more basic image portrayal, you could have a bar chart. Etc.
This of course is just a preliminary idea, but I've tried to flesh it out as much as possible.
Thoughts?
Comments
Rather than, "join marines", or "join aliens", they get assigned once they try to join a team. It should try to "prefer" the team they chose, but if they are unbalanced skill wise they get put on the other team.
Players should also maybe be penalized for dropping off of the team (ie, they get forced to alien, so they go back to the readyroom and wait to join marine), as well as F4'ing just before the round ends.
I see a situation with your idea where the teams will always be 'unbalanced', or jackasses always canceling the "go", or always making sure the teams are unbalanced so the game can't start.
Just my opinion anyway.
Also, ranked servers aren't that difficult. It would be required to log all stats a server is responsible for, if a server starts modifying gameplay, the stats that the server is responsible for can be discarded and removed from players who played there.
But preferably, ranking should be either global (all stock servers), or community wise (modified servers, or just servers that want to have their own stats data), so that a server community can run their own stats/rankings seperate from the main rankings.
If 'custom' servers are not allowed to rank (for obvious reasons), and have no way of running their own ranking system that's seperate from the main systen, it will kill any modding or customization.
Marines for 'newer' players
Aliens for the more 'expierienced'.
More often than not, frag ######s, for some reason or another like to gang up on marine team and ruin overall game expeirience.
Some days I can play well, i.e. 'In the zone', thus I would join/prefer alien team. Other days I feel a little 'off key', and would rather be a marine and just let the comm tell me what to do. Ofcourse, you could desighn the server to only allow accomplished' marines to join alien. Or visa versa if required.
- Persistent statistics tracking (achievable on Steam, requires your own hosting for all the data/data transfer, which UWE probably can't afford/manage just yet)
- An option for server admins to control what skill ranges they want to cater for
- A suitable range of statistics that cover all aspects of gameplay (i.e. not just skill)
This way any server admin could firstly choose to have matchmaking enabled or disabled on their server. Then they would select a 'difficulty' or a matchmaking 'profile'. A server's profile could be 'beginner' or 'advanced', with a slant towards general competancy across the range. Or it could be 'teamworker', with more emphasis on co-operative stats like 'average points welded when welder equipped', '% of Hives dropped when on 40 res or more' (if empty Hive available), '% of RTs dropped in the first 30 seconds of a round'.
The better the stat tracking, and the wider (yet pertinent) range of possible profiles available to the server admin for matchmaking, the better.
---
To begin with, though, I think UWE should forget about stats tracking and concentrate on how points are awarded to players on the scoreboard, and make sure that players understand which actions gain them points. When you can have a Gorge at the top of the scoreboard in NS, like you can have the Medic at the top of the scoreboard in TF2, co-operative gameplay will become more rewarding for a greater proportion of NS2 players.
Stix you have to accept a few bad things when you start talking about balancing and matchmaking. First, new players are bad and most of them don't realize how bad they are. A lot of them also don't want to admit it. Combining that with randomly finding new expert players to play against as though they were equals is like teaching calculus to a preschooler, and then rubbing their face in their failure.
Beyond that (which would be quite enough) I can personally superimpose that on Saraph's opinion of pubbing being primarily about ego (however true or false it might be) and see a great deal of angst created by mismatches.
Balancing has been done to death especially on TG, and it works to an extent but is a far more complex problem than you'd think. The primay issue is someone who latejoins in order to avoid the balance. How do you handle that person? You can shuffle the teams around dynamically but this creates a situation where someone might have been working for 20 minutes on a good game and ends up losing it due to their own efforts, and often. This reason alone makes balancing a bad answer.
Consoles are a great example of matchmaking by the way - a console (until recently) required that players be in the same general neighborhood, and to know each other socially, in order to play with one another. This worked excellently because it took players of similar age and maturity and put them in the same area without even realizing how brilliantly it was being implemented. Pubs don't do that, and it's an element which is excrutiatingly lacking and needs to be addressed.
Matchmaking, I think, will let a much longer influx of new players sustain the game through its lifecycle, while keeping the high level games truly high level, and just basically letting everyone have a better time overall, no matter what their skill level might be. If it's done correctly, it would, I think, turn the game away from an egofest and into more of an enjoyable outlet for uniting people through great experiences in an excellent game.
QFE
Could you explain how else it could be done? Am i really that far off?
One suggestion was to have servers set as 'beginner', 'advanced', etc. I can't even enumerate the problems that would arise from this, but I'll try for a few.
How would it work in the beginning release? Would the proportion of beginner vs advanced servers necessarily sway towards a majority of advanced servers as the playerbase plays for years?
NS1 is a very complex game that I got into on my own, just like some friends. But I also had to 'bring in' a few people, to hand-guide them through at least the first hour or so of playing, for them to be able to have their questions properly answered. Wouldn't this be impossible under the matchmaking example given?
Would we have server lists, and would they have special icons for servers that 'i cant join because im level x'?
Perhaps I'm missing something fundamental here. If i'm actually that far off, could someone explain to me what I'm missing?
There's also the question of learning. Having new players playing with experienced players seems like it would actually be a benefit rather than a bad thing in NS, as it is such a complicated game, and a quick tip from a pro could save a newb hours of frustration.
Ugh, its late, and this thread has already derailed so hard.
The original point of this thread was to be about this new balancing mechanism, using the skillranking concept. Instead of hopping into whatever team a player wants in a haphazard fashion, the teams would get fully sorted out before the round commences (akin to the current Freeze Time), but with the addition of the estimated skill balance algorithm.
As I think about it more, I feel that the whole idea of skillranking, matchmaking and whatnot is not worth the rather long amount of dev time it would take to implement it properly, in any fashion.
I thought we were talking about the following:
<!--quoteo(post=1673819:date=Mar 21 2008, 06:06 PM:name=StixNStonz)--><div class='quotetop'>QUOTE(StixNStonz @ Mar 21 2008, 06:06 PM) <a href="index.php?act=findpost&pid=1673819"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->One suggestion was to have servers set as 'beginner', 'advanced', etc. I can't even enumerate the problems that would arise from this, but I'll try for a few.<!--QuoteEnd--></div><!--QuoteEEnd-->
So I thought this was the suggestion, yeah. (In which case, the optional button would just place you in a suitable server with other players.)
<!--quoteo(post=1673819:date=Mar 21 2008, 06:06 PM:name=StixNStonz)--><div class='quotetop'>QUOTE(StixNStonz @ Mar 21 2008, 06:06 PM) <a href="index.php?act=findpost&pid=1673819"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->How would it work in the beginning release? Would the proportion of beginner vs advanced servers necessarily sway towards a majority of advanced servers as the playerbase plays for years?
NS1 is a very complex game that I got into on my own, just like some friends. But I also had to 'bring in' a few people, to hand-guide them through at least the first hour or so of playing, for them to be able to have their questions properly answered. Wouldn't this be impossible under the matchmaking example given?
Would we have server lists, and would they have special icons for servers that 'i cant join because im level x'?<!--QuoteEnd--></div><!--QuoteEEnd-->
Probably, but with Crispy's implementation:
<!--quoteo(post=1673801:date=Mar 21 2008, 11:45 AM:name=Crispy)--><div class='quotetop'>QUOTE(Crispy @ Mar 21 2008, 11:45 AM) <a href="index.php?act=findpost&pid=1673801"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->- Persistent statistics tracking (achievable on Steam, requires your own hosting for all the data/data transfer, which UWE probably can't afford/manage just yet)
- An option for server admins to control what skill ranges they want to cater for
- A suitable range of statistics that cover all aspects of gameplay (i.e. not just skill)
This way any server admin could firstly choose to have matchmaking enabled or disabled on their server. Then they would select a 'difficulty' or a matchmaking 'profile'. A server's profile could be 'beginner' or 'advanced', with a slant towards general competancy across the range. Or it could be 'teamworker', with more emphasis on co-operative stats like 'average points welded when welder equipped', '% of Hives dropped when on 40 res or more' (if empty Hive available), '% of RTs dropped in the first 30 seconds of a round'.<!--QuoteEnd--></div><!--QuoteEEnd-->
Everything's set up by server admins, and they <b>can</b> have un-enabled servers. For "hand-guiding" you'd play on unenabled servers.
I would guess that in the beginning there would be more beginning servers and unenabled servers. In the middle/prime there would be a mix. And then later in the game's life, when there's a lower number of servers, there would a larger proportion of unenabled servers and less of the others.
<!--quoteo(post=1673819:date=Mar 21 2008, 06:06 PM:name=StixNStonz)--><div class='quotetop'>QUOTE(StixNStonz @ Mar 21 2008, 06:06 PM) <a href="index.php?act=findpost&pid=1673819"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->There's also the question of learning. Having new players <b>playing <i>with</i> experienced players</b> seems like it would actually be a benefit rather than a bad thing in NS, as it is such a complicated game, and a quick tip from a pro could save a newb hours of frustration.<!--QuoteEnd--></div><!--QuoteEEnd-->
Sure, but you seem to be playing down the <b>playing <i>against</i> experienced players</b> aspect.
So I think the ranked servers implementation would be great.
However, the only thing missing is this consideration: what about playing from (public) netcafes (or even friends' PCs)? You can't exactly use the netcafe's profile.. it wouldn't be accurate. And I wouldn't wanna be restricted to the wrong server type and unenabled servers.
I think a few beginner only servers that block acess to players that have played over X amount of hours online may be a good idea. or that only allow access to the X number newest players. This may provide a good environment fo newer players where they dont get crushed constantly, but i dont see any need for intermedate or elite servers unless they are privately run.
1.) Who I killed/died to: NSPlayer, Mustang, a bot(their skill level at the time)
2.) What lifeform/upgrades/equipment they had
3.) My lifeform/equipment/upgrades
4.) The weapons used
5.) assists with the above information, maybe dmg based instead of kill based
Should structure kills be included?
maybe not kills but damage dealt? If 5 guys are shooting an onos, why should the one who actually killed it have more points? He could only spray and pray hitting the onos once, which eventually killed it.
Back to the structures -> yea, why not add structure damage into it. I hate it when after the marines are dead none of the skulk stays to destroy the rt, because it won't add up to their beloved and worshiped K:D ratio.
Also, not adding structure damage would discourage skulks to run into weakly defended base to destroy SC/Proto/ArmsLab in mid-game which is rather skillful thing to do etc., because they would be treated on the same level as others who simply run around the place and die (assuming that it takes a one or two suicide rushes into base to destroy these structures).
The algorithm would be a function that's like four lines long and would likely blow up your head from reading it.
We still seem to all have a few assumptions in this topic that we need to agree upon. When we're talking about this segregation of servers, would the player's 'upgrading' from Beginner to Advanced be one-way, or could a player actually get a bad K-D once he's in Advanced and be bumped down again? I'd think not.
I still think the idea of server segregation, as well as matchmaking, are bad. I can't fathom how it could work in a game and a server environment like NS.
I was going to suggest that instead of having classed servers, have a few dedicated 'newbie' servers, where players can only join for their first 10 hours of play or so (those 10 hours would be identical to what has already been classed as a beginner server).
But there are fatal problems already with this. There could be no community on such a server whatsoever, and there would also be no experience. If you took 20 people who are completely new to NS1 and put them into a server and just sat back and watched, what do you think would happen? The rounds would be over in 2 minutes with the IP-less marines wandering around until they somehow run into the clueless aliens and die. No RTs, no chambers, no comm. Sure, you could try to have dedicated experienced players on these servers, such as the NSGuides type of thing, but that would be insanely boring and draining for them.
As far as I can see, servers should stay the way they are now; anarchic and free. Do the tutorial videos, and perhaps create a special Helper Icon and only enlist those who are willing to really help (and to be professional). NS has the best gameplay out there because of its depth, but its depth is what creates all these other issues, especially the learning curve.
New players need to play with experienced players. It sucks to play against experienced players, but playing WITH an experienced comm, a marine to follow, or a fade to bark out orders and defend against the siege, is absolutely worth it. And necessary.
The algorithm would be a function that's like four lines long and would likely blow up your head from reading it.<!--QuoteEnd--></div><!--QuoteEEnd-->
It's all a matter of how accurate the devs try and make it. A reasonably good system could be worked out that wouldn't take a NASA to compute, I've seen them in indie RTS's all ready. No, it wouldn't be easy to measure the skill accurately and coming up with those calculations would be hard. Once you have your calculations they're not so bad though. for example mySkillRate + (yourSkillRate - mySkillRate)*(yourWeaponFudgeFactor - myWeaponFudgeFactor)*(mySpecialEquip - yourSpecialEquip)
will not take very long for a CPU to compute. It's all just trial and error.
<!--quoteo(post=1673868:date=Mar 21 2008, 03:54 PM:name=StixNStonz)--><div class='quotetop'>QUOTE(StixNStonz @ Mar 21 2008, 03:54 PM) <a href="index.php?act=findpost&pid=1673868"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->We still seem to all have a few assumptions in this topic that we need to agree upon. When we're talking about this segregation of servers, would the player's 'upgrading' from Beginner to Advanced be one-way, or could a player actually get a bad K-D once he's in Advanced and be bumped down again? I'd think not.<!--QuoteEnd--></div><!--QuoteEEnd-->
Good point. I'd imagine your skill can only go up not down. On most ranked servers that's how it works now.
<!--quoteo(post=1673868:date=Mar 21 2008, 03:54 PM:name=StixNStonz)--><div class='quotetop'>QUOTE(StixNStonz @ Mar 21 2008, 03:54 PM) <a href="index.php?act=findpost&pid=1673868"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->I still think the idea of server segregation, as well as matchmaking, are bad .... New players need to play with experienced players. It sucks to play against experienced players, but playing WITH an experienced comm, a marine to follow, or a fade to bark out orders and defend against the siege, is absolutely worth it. And necessary.<!--QuoteEnd--></div><!--QuoteEEnd-->
NSPlayer doesn't want to play with a 50-0 fade though. I agree with you that goal isn't to prevent NSPlayer from joining an experienced level server if he/she <i>wants</i> to, but to prevent the 50-0 fade from joining a bunch on newbie games. You're right about an experienced commander in some respects, but this would also be a good place for a new commander to learn.
Personally, I think the ideas in this thread assume that NS2 will eventually level off into a player base much like the one in NS today (ie, few servers few players). In other games, like CS:S, if you get banned or don't like a server, you can simply join another and continue there. Shouldn't that be the goal for NS2? If the game is fun, deep, and competitive with a large player base, then why is this even an issue? The skilled players can go buy their own servers or stop playing.
I would argue that a good indication of a booming game is one where skilled players will only join pubs to warm up or for purely comedy purposes, because the rest of the time they can, and are, scrimming or practicing with other good players/teams to get better.
Thats horrible, man. So once you've reached the point that all players strive for, where you're basically expert skill level... you are forced to either quit or invest thousands of dollars?
<!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec-->I would argue that a good indication of a booming game is one where skilled players will only join pubs to warm up or for purely comedy purposes, because the rest of the time they can, and are, scrimming or practicing with other good players/teams to get better.<!--QuoteEnd--></div><!--QuoteEEnd-->
Not at all. Thats one of the most biased opinions on gaming I've ever heard. I played CS for 4 years and only scrimmed for a few months. Although I was always top or near top on the servers, I enjoyed the relaxed atmosphere far more than the crazy seriousness of competitive. Such a choice, to venture to either side of the fence even temporarily, is always a personal choice and is often decided by the mood one is in.
If you join a regular pub of NS these days, there's almost always enough skilled players to offer more than enough of a challenge, assuming you actually do try to counter a stack. If you stack, then yes, its 'pure comedy', in a very bad way. If you balance the matches against people of equal skill level, you end up with amazing games. And you simply can't get games of those sizes or 'epicness' in 6v6 scrims.
For example, the requisites for being able to progress from Beginner to Intermediate servers might be hitting milestones something along the lines of:
- Killed 50 (human player) skulks
- Killed 5 Fades
- Welded 2000 armour/structure points
- Spotted Alien structures for 5 Siege volleys
- (As Marine) Built 10000 structure points
- (As Alien) Dropped 15 Resource Towers
- (As Alien) Dropped 5 of each upgrade chamber
- (As Gorge) Healsprayed 1000 points of health
- (As Marine) Dealt 1000 points of Hive damage
- etc.
The same could then be transferred to Commander(s) wishing to command on an intermediate server. They may have had to fulfil certain requirements on a Beginner server in order to progress, such as 25 scans, 5 round victories, 10 beacons, etc.
<!--coloro:#FF0000--><span style="color:#FF0000"><!--/coloro-->The point about this system is that you would always be moving towards hitting meta-game objectives (i.e. goals outside of the game). It would promote players to gain experience in all areas of NS's gameplay, from weapons, to classes, to using support items and special structures. In a way, these meta-game objectives could be considered an interactive tutorial that the player goes through at their own pace, and in whichever order they prefer.<!--colorc--></span><!--/colorc-->
Alternatively, or even additionally, if a reputation system were attributed to Commanders, at the end of a round a Marine vote could be held to attribute reputation points to the player last in the Commander chair before the round ended. This could then be used to help give an indication of how qualified a player was to command a game.
(- But griefer Commanders could go on private servers to boost their reputation so they could grief!
-Yes, but they would have to go on private, not public servers to guarantee reputation boosts, which means they spend less time griefing!)
---
I think in terms of intuitiveness, the 'Beginner', 'Intermediate' and 'Advanced' matchmaking server profiles should be determined by the devs, with these servers using a custom icon in the server listing to show they were using default/official matchmaking conditions. This way if you knew you were an Intermediate player according to the official dev definition of 'Intermediate' you would always have access to 'Intermediate' servers.
However, there should also be the option for server admins to set their own conditions to allow for custom profiles, such as 'Teamwork' or 'Hotshots'. Furthermore, server admins could then toggle these settings on and off. So an uber-friendly, well-adminned server might want to allow Advanced, Intermediates and Beginners to play together. A server which didn't have full admin coverage might restrict their server to Beginner-only to prevent very good player from dominating. Similarly an Advanced server would always have the option to prevent absolute rank Beginners from joining their servers and being unable to cut it.
---
Having the icons in the server browser is utterly dependant on Valve making the Steam matchmaking service what it needs to be, so I wouldn't hold my breath. But with Steamworks, as a developer it may be possible to customise the server browser somehow to suit your needs. I hope UWE would look to work closely with Valve on this because both parties could benefit from developing matchmaking together: one providing the need and the desired schematics, the other implementing features into a system that meets that need.
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.
Why is that horrible? This might be before your time, but the HAMPTONS server is a great example of a server run by admins who were top competitive players and knew what other competitive players wanted. New players avoided it because they were quickly kicked off due to not having a reserve slot (all competitive players on teams got reserve slots). That's how skill should be balanced on servers. And as for investing thousands of dollars, well, competitive players are willing to donate for that sort of thing. If you're not willing to pony up and pay, you better be willing to accept whatever an admin throws at you, no matter how good your argument for staying on their server is. Even the most rational arguments made in ban appeals threads today by competitive players are largely ignored on purpose. Today, these competitive players can either stop playing, buy a new half-life account and change their IP and hope they don't get banned again (they will), or buy their own server and create their own community (doesn't work). The last one might work with a popular game, but I digress.
<!--quoteo(post=1673883:date=Mar 21 2008, 04:23 PM:name=StixNStonz)--><div class='quotetop'>QUOTE(StixNStonz @ Mar 21 2008, 04:23 PM) <a href="index.php?act=findpost&pid=1673883"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Not at all. Thats one of the most biased opinions on gaming I've ever heard. I played CS for 4 years and only scrimmed for a few months. Although I was always top or near top on the servers, I enjoyed the relaxed atmosphere far more than the crazy seriousness of competitive. Such a choice, to venture to either side of the fence even temporarily, is always a personal choice and is often decided by the mood one is in.
If you join a regular pub of NS these days, there's almost always enough skilled players to offer more than enough of a challenge, assuming you actually do try to counter a stack. If you stack, then yes, its 'pure comedy', in a very bad way.<!--QuoteEnd--></div><!--QuoteEEnd-->
Okay, maybe I should have said "extremely skilled" instead of just "skilled" players, and by extreme I mean being the determining factor in winning a game on their own. You were top or near top on those servers because the best players had other (scrimming, pugging, matching), more fulfilling outlets to spend their time in.
<!--quoteo(post=1673883:date=Mar 21 2008, 04:23 PM:name=StixNStonz)--><div class='quotetop'>QUOTE(StixNStonz @ Mar 21 2008, 04:23 PM) <a href="index.php?act=findpost&pid=1673883"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->If you balance the matches against people of equal skill level, you end up with amazing games. And you simply can't get games of those sizes or 'epicness' in 6v6 scrims.<!--QuoteEnd--></div><!--QuoteEEnd-->
<!--quoteo(post=1673883:date=Mar 21 2008, 04:23 PM:name=StixNStonz)--><div class='quotetop'>QUOTE(StixNStonz @ Mar 21 2008, 04:23 PM) <a href="index.php?act=findpost&pid=1673883"><{POST_SNAPBACK}></a></div><div class='quotemain'><!--quotec-->Thats one of the most biased opinions on gaming I've ever heard.<!--QuoteEnd--></div><!--QuoteEEnd-->
Again, my point is that it'd be nice if there was an easy way to filter out particular players to more easily allow admins to weigh the pros and cons of which class of player is playing in their server, but that'd be as far as you'd need to go. Let admins sort out stacking, and let admins move players around themselves to achieve balance. If you get banned or don't like admins messing around with your team, go somewhere else that's closer to what you want. If you can't go anywhere else, make a place that caters to what you want. If that doesn't work, move on (which most of the best players have done)
I'm not saying everyone with twitch skills or high K:D ratios display this sort of behaviour, but you definitely cannot say that a high K:D rating shows a maturity of conduct and a deep understanding of the gameplay.
K:D has so many problems it would take too long to properly go through them, but an itemised list might look something like:<ol type='1'><li>K:D is massively relative, it promotes unsportsmanlike techniques like spawncamping (imagine what will happen when Marines are asked to shoot down the Hive faced with the glorious opportunity of free kills), which while valid tactics in certain situations, are detrimental to the fun aspect</li><li>This also makes it totally prone to exploitation by griefers, who will just find 'open' noobie servers where they can boost their rating for bragging rights that they really didn't earn and don't deserve</li><li>K:D doesn't begin to cover the importance of non-combatant roles. When building an RT, one person has to build and one must cover, but with K:D determining which servers a player can and cannot go on, neither player will want to build anything.</li><li>Furthermore, K:D is dynamic and playing against a stacked team will reduce your K:D rating. This promotes team-stacking and creates an additional problem to solve.</li></ol>So, yeah. Your proposed K:D ranking effectively promotes selfish behaviour which undermines the team-oriented gameplay of NS. It is only effective as determining rank in deathmatch games. It might work for Combat, but it would totally screw over Classic mode. Since Combat doesn't exist for NS2 until a third party makes it, K:D matchmaking isn't really an option.
Actually, every rational ban appeal I've ever seen has either gone through to unbanning, or forced one or many admins to slug through massive logs to prove that the person banned is actually lying. Which happened about 75% of the time it had to be done.
It sounds like you or many of your friends have been banned from a lot of servers. Why did this happen in the first place? If its newb admins banning for hacking when they have no idea what they're talking about, then that just sucks <img src="style_emoticons/<#EMO_DIR#>/sad-fix.gif" style="vertical-align:middle" emoid=":(" border="0" alt="sad-fix.gif" />. But I find the majority of bans still held up these days are for blatant racism, complete obnoxiousness, or attitudes and actions that go against the general enjoyment of all.
Even in my time as a Mr| Admin, while we had some bad admins, we always made everyone justify any bans. NOONE could be banned for hacking without sufficient demo proof, which was available for all admins to review and discuss. Often we'd have had multiple admins review a demo within the hour of the ban occuring, and maybe 25% of the time the ban was removed immediately.
But there were PLENTY of cases where someone would turn up on the forums, 'Hello, I was banned a month ago for something, I don't really know what, but I'm here to request an unban.' Rational and collected. But then we go through the logs, find him to be using the worst of the racial slurs the instant the admin leaves, or actively ruining games by recycle, etc, and we laugh and him and then ban him from the forums.
Of course, he'd come onto the other servers spouting how horrible Mr| is.
What I'm saying, is that you shouldnt be getting banned in the first place. I've had thousands of games on BAD or Guns or any of the still active pub servers where i'm annihilating the other team, but the idea of banning doesn't even go through an admins head, because I'm friendly, helpful, always fighting the stack, etc. If you are having all these issues getting banned from every server, it *might* be your own problem, not a problem with the general admin community.
aswell as ns has always been about playing with my friends (for me anyway) and we are all roughly the same skill level, so i'd hate for the damn sever always put us on opposite teams.
Matchmaking in its most desirable form is about putting players at or above a dev-defined level of <i>experience</i> (not skill) on the same servers. If everyone on the same server has the same level of competancy, there shouldn't be much stopping you and your friends from playing on the same team, other than an autobalance system that shifts the last person to join onto the opposite team for balance reasons. If you make sure you and your friends all join servers with enough space to allow one more person in after you've joined, it's less likely that you will be split up by autoteam, since someone will spot an almost full server and join, so you and your friends won't be eligible for autobalance.
I'm not saying everyone with twitch skills or high K:D ratios display this sort of behaviour, but you definitely cannot say that a high K:D rating shows a maturity of conduct and a deep understanding of the gameplay.
K:D has so many problems it would take too long to properly go through them, but an itemised list might look something like:<ol type='1'><li>K:D is massively relative, it promotes unsportsmanlike techniques like spawncamping (imagine what will happen when Marines are asked to shoot down the Hive faced with the glorious opportunity of free kills), which while valid tactics in certain situations, are detrimental to the fun aspect</li><li>This also makes it totally prone to exploitation by griefers, who will just find 'open' noobie servers where they can boost their rating for bragging rights that they really didn't earn and don't deserve</li><li>K:D doesn't begin to cover the importance of non-combatant roles. When building an RT, one person has to build and one must cover, but with K:D determining which servers a player can and cannot go on, neither player will want to build anything.</li><li>Furthermore, K:D is dynamic and playing against a stacked team will reduce your K:D rating. This promotes team-stacking and creates an additional problem to solve.</li></ol>So, yeah. Your proposed K:D ranking effectively promotes selfish behaviour which undermines the team-oriented gameplay of NS. It is only effective as determining rank in deathmatch games. It might work for Combat, but it would totally screw over Classic mode. Since Combat doesn't exist for NS2 until a third party makes it, K:D matchmaking isn't really an option.<!--QuoteEnd--></div><!--QuoteEEnd-->
I see what you mean, but I can't agree that other things than pure aim and movement should affect match making. While I do agree that mechanical proficiency alone is not enough to be a great player, you need hardly be a great player to make the game less fun on a public server.
What I think is the big issue is the skill gap that enables a single marine to hold off alot of skulks on his own and make his way into a hive to for example initiate a spawn camp. Now if the skulks who need to stop him aren't good enough, purely mechanically, to stop him from getting there - to actually bite him down and overcome his aim and capacity to predict and move away - the team who are on the recieving end will not be happy. They are beaten and they can't do anything about it.
I think we need to separate two different issues connected with a match making system and the ranking of players associated with that function. The first is: what we want to achieve by the system; the second is: how will the ranking system be recieved and appreciated by the players.
On the first issue I believe the main thing is what I described above. To reduce the differences in pure skill between players so that "lame" things will be more rare since that marine can be killed by the mechanically equally skilled skulks before he ruins the game. A more enjoyable game for all. Yay.
The complexity of the second issue on the other hand is displayed well by your last response. Players will not see and appreciate the reason behind the rankings and the match making system, they will merely see the effects and the tangible results of it.
My idea of the reasoning of a random player would be something like this: I have rank x, and you have rank x+10, thus you are considered better than I even though I understand the game better and do more for my team.
Or of another player: I have rank x and you have rank x+10, even though I can kill twice as many enemies than you, how is that?
So in summary, we have two separate issues. The system itself and the way we envision it works and helps the community, and the way the system is recieved and actually affects the psyche of the community. What we <b>want</b> to do is make the game more enjoyable for everyone despite large skill gaps (1on1 LT ladder games, 2v2 BGH no rush gogogo). What will actually happen is more ######ing about whose horse is higher.
In what I envisage as an effective matchmaking system, the jump from Beginner to Intermediate would require a new NS player to hit certain milestones that cover the basics of NS's gameplay. One example of this could be parasiting 10 different enemy players or structures. This would be an achievement target which would be viewable on the SteamID page, just like the TF2 stats and achievements. Now, it is clear that a player who wanted to progress to Intermediate could rudimentarily parasite everything he sees until he hits this particular milestone, without caring too much about why he's doing it other than to fulfil a statistic. But- during this process it may become clear that parasiting an enemy makes them MUCH easier to track and take down.
Whether the particular player in question realises this or not isn't that important, the major benefit this gives is that it will at least provide information about how to perform the action and what its basic effect is. What's more, and probably best of all, is that ALL the Alien players will benefit from having parasited Marines on their HUD. A total newbie may wonder what this blob is and go to investigate, before realising this is a Marine. They may even ask why some Marines are displayed on the HUD and some aren't. Other, more experienced players will simply have better reconaissance of their enemy's movements. All of this enhances the collaborative aspect of the gameplay without forcing players to consciously work together.
<!--quoteo--><div class='quotetop'>QUOTE</div><div class='quotemain'><!--quotec-->I think we need to separate two different issues connected with a match making system and the ranking of players associated with that function. The first is: what we want to achieve by the system; the second is: how will the ranking system be recieved and appreciated by the players.<!--QuoteEnd--></div><!--QuoteEEnd-->Firstly, I don't agree with having a ranking system. A basic milestone system of progression from Beginner to Intermediate to Advanced is enough to segregate certain player types from others for the benefit of everyone playing. Open servers cater for anyone who doesn't want to play above or below their experience/skill level for whatever reason.
Server-side plugins cater for local ranking. Global ranking isn't a basic requirement for NS. Competitively-natured players would prefer a ranking system, but it does little for a new player, because they are measured against other players on a dynamic basis and to answer your second issue: there is no clear indication of what a player has to do to increase in rank, since it is all relative and dependant on how well other players are playing. E.g. there are no clear static milestones that a player must hit to improve. A player might go up in rank one week just because less players are playing the game in that week, or because a lot of people coincidentally have a lot of off-days. Then, when they have what they think is a great game, it just happens that a lot of other people are playing well, and their rank doesn't improve as much as they thought it would, or it could even fall.
The advantage of a static milestones system is that players have clear goals communicated to them via their SteamID page that guide them through the fundamental elements of the gameplay. Kills:deaths focuses purely on combat, nothing else. Ranking focuses on how well a player did in a day versus potentially millions of other players.
Finally, one other problem with ranking is that it requires data to be communicated at the end of every round for every single player regardless of the conditions (transmitting it live would create too much server traffic). If a connection was interrupted this information would either be lost, or would need to be stored locally and then transmitted on the next reconnect. This creates a problem, because local storage of data brings in the problem of players hacking their accounts to boost their stats (sad, pathetic stat######s WILL attempt to do this). Because milestones won't be achieved at such a high rate, there is the potential for them to be transmitted live to the stats server via Steam. Even if they are held locally to be transmitted at the end of a round, since there is no competitive element involved, it is purely a measure of personal achievement, so anyone attempting to hack their own milestones are only defeating themselves.
As for how the 'segregation' aspect will be received by players, I strongly believe that the protection it provides will be much appreciated, whether from Advanced players using 'pr0' tactics to demolish Beginner players, or from Beginner players being unable to grasp the finer-yet-crucial details of NS's complex gameplay.