First off, as you can see I'm new to the forum. However, I'm far from being new to Natural selection. So, I hope you won't dismiss my idea too lightly.
Also, I am what you would call a pro-gamer. I take a lot of time to play games, much less time to comment them. If I'm making this effort to share my opinion, it is solely because I like Natural selection (a lot) and would like its sequel to be as perfect as possible.



I've read as many posts as I can and a lot of them have interesting suggestions. Nonetheless, there is one particular point that troubles me: the ranking. I've seen the many discussions this has aroused and by scanning the rather lengthy topics I came to 2 conclusions:
1. There's a lot of arguments, suggestions and ideas...and as many counter-arguments, critics and questions.
2. In spite of this first observation, no real viable solution has yet been offered.

I'm not going to discuss further the points already brought up (cf. 1.). Suffise to say that I am for implementing a global ranking system because of the numerous reasons which have also already been given (better gaming experience, better matchmaking, better regulation of troublemakers, etc.)
My suggestion (to cf. 2.) is simple: find out what ranking systems have worked in the past, copy them and eventually try to improve them. It's as simple as that.

I've seen some people arguing between a frag-based and a TF2-based scoring system, but in my humble opinion they aren't seeing the big picture. A global ranking system means you don't need to link any points to a particular session, but can instate an overall rating. Consecutively, the ranking doesn't need to take into account any vain and minor accomplishments such as frags, structures built/destroyed or any individual accomplishments. It simply needs to record the number of games in which you have been in the winning or losing team. This would solve the problem of individualist players since the scoring would only be concentrated on how well your team fares.
In other words, you need to forget about the FPS-styled ranking systems (such as CS or TF2), but try to prospect RTS-stylish ranking systems. Some of them work great and could work very well with Natural selection as well.

My favourite ranking systems are those of Myth I, Myth II and Supreme Commander.
The Myth series (created by Bungie, who is the developer of Halo and other very successful games) in particular had an unique way of delivering the "whole" package in multiplayer. You logged on a global server and could discuss in lobbies with other players while always having your rank portrayed as a nice symbol next to your name (comet for number 1, sun for number 2...axe and shields for number 500-1000, axe for 1000-2000, etc.). The scoring system was quite complex (involving points attributed to several categories, which could also be done in Natural selection by having a commander category, a marine category and an alien category), but you could sum its essence to this: each team's strength is evaluated by calculating the average number of points its players have and in case of victory/defeat the winning/losing team would win/lose a number of points accordingly to their initial strength. In other words, a team with less good players (with less points/rank) would lose less points in case of a loss and win more points in case of a win...and vice versa.
Supreme Commander, in the other hand, has a very functional matchmaking system...I just don't like the fact that your rank is displayed with multiple images (related to different accomplishments) instead of just one ladder picture.




Ultimately, a ranking system could and should be implemented in the game; that is what I feel the most strongly about. I just hope it will be set in a proper way, as it affects deeply the entire gaming experience.




This thread having grown to tl;dr proportions, I have decided to edit the original post (above) and add a condensed and up-to-date version of this ranking system with all the pertinent remarks and issues brought up by the repliers (underneath).



The detailed description of this ranking system:

Harimau sums the system up quite well:


The three main categories (Commander, Marines and Aliens) could also be divided into several subcategories. Different alien-species and marine-weapons could represent different subcategories. The subdivision could also be made between "support/assist" and "offensive" play. Although subcategories could be related to elements such as kills or assists, these statistics would never be apparent to players.
Also, the important thing to keep in mind is that you still would get points only if your team wins. The main goal remains to contribute to a teamwin, but your points could be awarded/substracted to/from you according to the role you played in the win/loss (that is to say as an offensive player, as a support, as a Fade, etc.).

Here is an additional step-by-step description of the system works:
1. You have 210 points in your overall ranking. You have 100 points in Comm category and 100 points in Alien category, but only 10 points in your Marine category.
2. You play with players having an average rank of 210 points. The average number of points they have in the specific categories they play is 100.
3. The scoring weight can either be based entirely or partially on players' overall ranking (210 points) or on the category players are playing (100 points). This should be decided by the developers (possibly after a beta-testing?).
4. The scoring is based entirely on the overall ranking (210 points).
4.a. You play the commander.
4.a.i. You win the game. You get 15 points, which are added to your commander category (making it 115 points) and your overall score (making it 225 points).
4.a.ii. You lose the game. You lose 15 points, which are substracted from your commander category (making it 85 points) and your overall score (making it 195 points).
4.b. You play a marine.
4.b.i. You win the game. You get You get 15 points, which are added to your marine category (making it 25 points) and your overall score (making it 225 points).
4.b.ii. You lose the game. You lose 10 points (and not 15 points since you only have 10 points in your marine category), which are substracted from your marine category (making it 0 points) and your overall score (making it 200 points).
5. The scoring is based entirely on the category each player is playing (100 points).
5.a. You play the commander. Note that the weight of your rank in this specific category remains the same as in the overall rank.
5.a.i. You win the game. You get 15 points, which are added to your commander category (making it 115 points) and your overall score (making it 225 points).
5.a.ii. You lose the game. You lose 15 points, which are substracted from your commander category (making it 85 points) and your overall score (making it 195 points).
5.b. You play a marine. Note that the weight of your rank in this specific category is no the same as in the overall rank. Hence, your team's average rank will be slightly lower than the one of the other team. This basically means you will win more points in case of a win and lose less points in case of a loss.
5.b.i. You win the game. You get You get 21 points, which are added to your marine category (making it 31 points) and your overall score (making it 231 points).
5.b.ii. You lose the game. You lose 9 points, which are substracted from your marine category (making it 1 point) and your overall score (making it 201 points).

The algorithms involved might be a little bit more complex than what I've portrayed here, but I guess the developers could simply contact Bungie in order to get a closer look at their scoring system. The key element with this ranking is that once you start getting points in one category (or subcategory), it will get increasingly harder to get more points in that specific category.
If you have for example over 500 points in the Aliens category, you may only get 1-5 points for a win and risk losing 20-25 points for a loss. And if you have 0-100 points in the other categories, you would win 20-25 points for a win and only risk losing 1-5 points for a loss. So to get the best overall rank the easiest way is to have roughly the same amount of points in each category.

The points awarded for all players in a teamwin or the points substracted in case of a loss are not exactly the same to all the players in a team. The weight of a rank is indeed diluted in the average rank of a team and each player of a team gets or loses points accordingly to the average rank of the team versus the average rank of the opposing team. However, there is some differencies in the distribution of points within the team depending on the rank each player has.
So say that team 1 wins and all players should get about 15 points. A much lower-ranked player of that team may even get 20 while a much higher-ranked player of that team may only get 10 points. The same goes for the losing team. Players from team 2 would lose about 15 points, but a low-ranked might only lose 10 points, while a high-ranked might lose 20.

A category or a subcategory could periodically be reset to 0 points (for all players simultaneously). This would prevent players from getting infinite amounts of points in any number of categories, reset the rank of accounts which have been inactive for a long period of time, and allow newcomers to have access to higher ranks as well.

You don't distinguish individual contributions (such as K:D), otherwise you end up with statwhoring. Despite some players may contibute more than others, the sole objective of the game remains to make a teamwin. In any case, in the long run, players contributing the most will have more wins than the ones that aren't. It doesn't take too many sessions for the statistics to even out. This might be a bit troublesome at beginner levels, but a noob will never get miraculous wins one after the other and just ride the ranks by winging winning teams.

The working elements:

- This system worked on the Myth series which has some very similar elements to Natural selection: it's team-based, it's got a multiplayer support for 16 players, it's an RTS and yet heavily oriented on fast-paced action, and the duration of sessions are roughly the same.
- Each session would have a fixed starting and ending point. Thus, all players would play an entire session together from the beginning till the end. This would add a lot of teamplay, teamspirit and knowledge of other players' behaviour/tactic/skill, etc. It would also make the gaming environment seem much more organized and clean, even with a lot of players around. This is absolutely not the feeling I get in FPS's in which there are extremely many players, especially when players just can go in and out of the server as pleases them. Often I get a pretty chaotic impression of any kind of active teamwork when new players arrive in middle of a battle and others quit or simply stay afk. I'd be overjoyful if I'd see a proper 32-player game in which each participant would play his part from the beginning till the end and would stick to the tactics laid out at the start.
In other words, random pubs offer too random of a gameplay and chaotic gaming environments while competitive play requires too much involvement, time and skill. A global ranking system would combine the benefits of both, without their drawbacks; it would permit easy access to somewhat competitive play and controlled environment, without demanding as much involvement, time and skill.
- Players quitting a server in middle of a session would be punished by substracting their points as if it were a loss (like in RTS games). A suggestion has been made that the penalty for dropping from a game should increase the closer to the end of the game you are. So leaving a game with 30 minutes left gives you, say 1% of the penalty, while leaving the game with 1 minute left gives you 100% of the penalty. Although this variant would give a bit of flexibility to the fixed starting and ending point issue, it induced a number of new problems (which are displayed in detail on page 8).
- Unranked servers would allow for players to play with friends which haven't got the same rank or to be able to leave a game whenever they want without sacrificing their rank.
- Servers do not need any admins to regulate troublemakers. Since players who would be prone to act mischievously (i.e. TK'ing) would more often lose than win, they would stay low in ranks. So this kind of disruptive behaviour would only afflict beginner levels. Also, cheaters always get detected eventually, especially in higher levels. This would naturally mean that their accounts would get deleted and as such would have to restart at the beginner levels.
- Match-making would have to rely entirely or partially on each team having roughly the same average "overall rank" (this is to say the sum of all points in all categories) or roughly the same average "category rank" (this is to say the points in the Commander, Marines or Aliens category which each player has chosen to play).
- Match-making could either be automatic (giving each team as many bad and good players as the other team) or manual. I'd be in favour of the latter. Let people choose with whom to partner with, knowing that if they have only good players in their team against a bunch of bad players they risk much more points than they could possibly win (i.e. in the Myth series, if you had too unbalanced teams, the good team would win only 1 point and risk losing 32 points). I'd say players would partner up either as aliens or marines in some sort of in-game chat/lobby (their overall and category points being visible to everyone). Then, the marine players should have to vote for their commander (which would need to accept the post). If subcategories would exist, those points could also be visible and as such they could be an indication for players of what weapons they're good at as marines or what creatures they're good at as aliens. So, eventually, players could already even decide what specific roles they would play.
Also, either the server should allow only a certain range of ranks to enter, either the players should be able to kick other players out.

The issues to be tackled with:

- Deadlock situations would be particularly troublesome for any ranked game, since quitters are punished. This isn't really a problem related exclusively to the ranking and should be coped with in any case.
- Competitive players who are focusing on pure clan versus clan play are going to ranked oddly due to the low number of games actually played. However, these kind of players are a minority and don't play a lot on public servers. In other words, their skills wouldn't interfere often with the match-making system and, in any case, only at beginner levels.
- A bad commander could screw up the whole marine team. This isn't, however, a problem per se, because the commander does represent a pivotal point in teamplay and teamwin; if he does do a mistake, the team will and should suffer from it and possibly lose the game.
- As the number of players in a game increases, the impact a single player makes on the game decreases. It does decrease, but every player still has an impact. Sure you may need to play a few more games in order for it to show on your ranking, but the statistics eventually even out. Also, since the match-making should create pretty even teams, the outcome would depend on how every player finally surpassed themselves individually. The scoring system proved to be functional in the Myth series, which supported up to 16 players. It would probably still be working for sessions gathering up to 32 players, but the major issue would concern servers with a more important number of players than that. One way to solve the problem -but rather crudely- is to simply limit any ranked games to 32 players.
- Match-making that is manual would need to either let servers put a (low and high) limit on ranks that are allowed to enter or let players somehow select with whom they play with. This latter option would, in other words, necessitate for servers to let players kick other players for any number of reasons (such as inadequate rank). The problem is that there is no host or any other particular player who could be endowed with special powers such a kicking other players. One solution, although I don't know if it's technically possible, would be to let a single player actually host the chat/lobby environment in which teams and preliminary strategies are decided. The host would control player entries and exits as he would see fit. Once all players check "ready" the host loses these provisional privileges and the game takes place on the real server.
- Match-making that relies entirely or partially on each team having roughly the same average "overall rank" would force players to play all sides (as a Commander, Marine and Alien) in order to get the highest rank possible.
- Match-making that relies entirely or partially on each team having roughly the same average "category rank" would confer the overall rank a simple symbolic status and could possibly push players to stack the alien-side since marines can get penalized by a lousy commander. This problem could be solved or at least palliated by giving the marines and commander an easier scoring (i.e. 10% more points in case of a win and 10% less points lost in case of a loss).
- The "average" casual FPS player does not want a long time-commitment so much as a time-sink (unlike the "average" casual RTS or RPG player). Nonetheless, I'd like to believe that a Natural Selection player is not that "average" casual FPS player. This hybrid game focuses so much on RTS elements that it requires, in my opinion, much more depth in gameplay (and therefore, in time-commitment) than most of the other FPS games.
I, for one, would like to have an FPS game with something more than "shooting virtual monsters ASAP" without the need to join a clan and having to plan beforehand when I can have a challenging and enjoyable session, such as a clan-match.

Thanks for the post, and you make some great points/suggestions.

Its definitely a good read, however I would be opposed to that kind of system, where...


If only the winning/loosing ration would be taken into account, then players will get around this by quitting the server or typing 'retry' in the console when the last hive/cc is almost destroyed and wait in the ready room for another game. That kind of system would also put good and bad players on the same level, as long as they were in the same team. Even worse: lets take a guy who shoots at and throws nades under the feet of his team mates, the legendary archetype of bad player and TK'er , if he will not get kicked/banned (e.g. admin is not on the server) the guy can even get a point, as long as he was causing trouble in the winning team.

Won/lost ration is good for RTS when you have to quantify the "goodness" of single commander vs other commander (or commanders). In NS not only the commander should be rated, also the men/alien on the ground who apart from killing off the opposite team should be also encouraged to help each other / work together... therefore I believe that NS should have a rating system based on team-linked things, which sums up together. Unfortunately, I also realize that balancing this in terms of which aspect of game play take into account and which is more important would be as hard as hell... Personally I hope that the alien commander and the unified resource model will shed more light to that problematic issue.

As long as it isn't like In Company of Heroes, when you win a game you get a point. But the losing player will quit 9/10 times if he is starting to lose. Now I think they've fixed it so that the winning player will still get his point, by killing the player, now replaced by AI. However it does negate the negative "lost games" points for the quiting players, so quiting the game has no impact on your points. And finishing of AI isn't quite as fun as destroying a player

Not that I am interested at all in levels/ranks or other nonsense (I think ranking systems are silly), I'm just interested that it will not promote none teamplay behavior...

I'm against any kind of "ranking" system that uses individual points, frags/death ratio if you will, for its messurements. Statwhoring/don'twannadiecrap comes to mind. So yeah even if this "RTS ranking system" is kind of limited, it does have a few interesting ideas to it.

But they do need to look into players quiting/reconnecting/getting disconnected. I mean what if a player leaves in the middle of a match or quits when his team starts losing, how are his points calcuated. This system can work (none CoH system) for RTS, but you do have to realise that this is and RTS/FPS hybride, meaning there are more players on a team then just a commander like in a RTS.

Only a win/lose ratio per game is quite limited and simple things like:
- structures built
- amount healed (gorge)
- welding (marines/structures/weldingpoints)

Do need to be in there as these will nurture teamplay, which is the most important part of this game...


But most of the time when a game comes with a ranking system, players start to ingnore teamplay. You say TF2 has it all wrong or isn't seeing the big picture. then explain to me why people like to play medic/pyro spy defence Now I'm not saying the whole system is perfect, but the individual classes do need an individual point system and it does promote some teamplay behavior as well as statwhoring.

Perhaps UWE will be the first to implement a perfect system, we shall see. But I'm not holding my breath (ranking systems are for the weak anyways )

Pro Gamer huh? What professional teams have you played for?

Also, as a Pro Gamer, I am surprised that you are in favor for a stats system for balancing NS. I'm sure that you are aware that most professional players arent the ones who are high up on the leaderboard, they are the ones that are high up in skill. I'll take the top 8 players on the leaderboard for BF2 and put them against 20ID my former team, any day of the week. Most 20ID players aren't even ranked, and if they are, they aren't very high up the ladder. http://20id.org/?page=matchlist&gameid=8 for reference.

Most pro players do not do well on the leaderboard because it would interfere with their practice and match time. These are the players that would have the low score when they actually do pub, and would completely destroy any ranking system that was in place.

Definitely shocking that a pro would favor a ranking system for balance, rather than an alternative system. Unless of course we have different definitions of pro.

I would agree with copying existing, proven systems which could be used for NS2, except that no such systems exist. NS is a "large" team game, so RTS ranking systems simply do not apply. Rankings based on team wins would not represent anything except how pissy you are about getting on the "good" team. And in the process, the system would suck the life out of most public games.

The reason the situation is different is because the teams are sufficiently small in the previous systems, so you would have to carry your weight if your rank was over-loaded. However, in NS, stackers would be able to free-load and a significant ranking advantage would be conferred for doing so.

In NS, if ranking were purely win-based, it would be a stack-fest. All day long. Stacking occured in NS1 to some extent, but it would totally destroy public games if win-ranking was implemented. And don't even get me started on the drop factor and server emptying. Is ranking based on wins an ultimate solution? No way.

Really? I thought Firewater's Reserve Slot System would work fairly well, although I would definitely want to include a weighted ranking system as a secondary test for the skill of a given server, as I mentioned in that thread.

Keeping this short so as not to retread things discussed in other threads.

First, if 5 good gorge players all join aliens, their points total would make them look top tier but they might be incapable of fading. This would not be an enjoyable game.

Second, as Firewater stated (perhaps a bit too pointedly), people who are focusing on pure clan vs clan play are going to ranked oddly due to the number of games actually played. A casual player can drift from server to server, while a clan-style player is going to be practicing strategy in-house or spawning bots and practicing their aim and reactions. Not things that will give them enough "points" to indicate their fading ability.

As others have stated, NS is a large team game and thus makes it very hard, if not impossible, to rank players individually in a manner that would give a realistic and tangible benefit.

In your opinion. I think K:D that accounts for player skill and weapons used would provide a meaningful, even though not perfect, system.

First of all, my ###### is way bigger than any of yours.

Alright, now that that's out of the way.
There is no perfect system. This system is as good as any other, but at least it is a system that is not based in being infowlable. A bad player will of course get a few points that may be undeserved, but the idea is that the really good players will always get more.

Ok, flatrick, you draw attention to the sections in green - but the most important one is in RED. That's what is a big problem with the ranking system idea - it's difficult to come up with something that even remotely resembles a just assessment of player overall skill for NS2 (right now most arguments are being based on conceptions of twitch-skill's importance from NS1) so that it can be used to restrict access to certain servers and provide access to others.

K:D unfortunately does not account for support Lerks, base attack Onos, or healing Gorges. It doesn't account for the Fades or Skulks that play diversionary tactics. Secondly, you have not defined "player skill". This is a nebulous topic, and is the real meat-and-bones of a ranking system. You cannot simply say "rank on skill" and then walk away. Skill has to be compartmentalised, defined, measured. In addition, pure K:D still does not support the CAL type player effectively.


So, with respect, it is VERY HARD if not impossible to create a ranking system from this data. You WILL get players who prefer the support role. They will be very good players, very effective players, very cooperative players, but they will not have awesome K:D. NS is not purely a game of K:D. It is a team game based on cooperative effort. Under K:D, you are saying that a hit-n-hide "I don't want to jeopardise my K:D" Fade contributes more than a healspamming gorge making sensory checkpoints and OC ambushes. One hand washes the other.

Sarisel has highlighted the appropriate quote in the original post.

What juice said. As the number of players in a game increases, the impact a single player makes on the game decreases.
And I disagree with the notion that K:D ratio is necessarily a good measure of player skill - cap team tends to die more, LMGs building the sieges, permagorge usually doesn't get many kills, etc.

Thanks, that's encouraging. I wasn't exactly sure if these points had already been made, so I'm glad I had something to contribute.





I don't really see this as a problem. You would simply need to have some kind of system to punish players who quit servers, as RTS games have in their ranking systems.
This would of course restrict somewhat the possibility of just jumping into an ongoing game and leaving it when needed. The solution to this accompanying problem would be to have separate servers/rooms: those dedicated to unranked games and those dedicated to ranked games. You could join and quit the first ones whenever you'd want, but you couldn't join nor quit the other ones after the beginning of a match.



Yes, I've experienced this in other games and it often proves to be a good way to proceed with things. Bad players would learn faster by sometimes playing with better players, while good players would have to cope with some bad players (which adds some challenge and some gratification by having other players look up to you).
Both the Myth series and Supreme commander had a way of preventing too big differences between the ranks of persons playing together: in the Myth series, players/servers could manually restrict a certain category of players from entering a game and, in Supreme commander, the matchmaking system would automically search for an opponent sufficiently near your rank.



The beauty of a ranking system is that servers do not need any admins anymore and that's specifically what I meant by a "better regulation of troublemakers". Since players who would be prone to act mischievously (i.e. TK'ing) would more often lose than win, they would stay low in ranks. So this kind of disruptive behaviour would only afflict beginner levels.
Also, I'd like to point out that cheaters always get detected eventually, especially in higher levels. This would naturally mean that their accounts would get deleted and as such would have to restart at the beginner levels. In other words, cheating (as any other stupid behaviour) is a rather rare affliction in ranked RTS games.



I don't really see why you're saying that a "rating system based on team-linked things, which sums up together" would be "as hard as hell", especially when I previously explained this part. Natural selection could have a category for commanders, another for marines and a third for aliens...maybe even subcategories? Each category would start at 0 points and would gain/lose points only by playing that specific category. Since your rank would be determined by the sum of all points, this system would push players to play all categories (in more or less equal proportions).
The Myth series had an excellent system in which you would easily gain points at the beginning (from about 0 to 50 points), but have a much harder time winning any once you had a lot of points in a certain category (over 300 points). You would also lose much more points if you had a lot of them. This pressed people even more to play all categories -if they were eager to have as many points as possible- as it was easy to get the first 50 points in each category but increasingly harder to get more of them.

Same response as for La Chupacabra.



I think ranked games would seriously support a good teamplay behaviour since only teamwins/teamplay would give you points. It's true that ranking would add some maybe unnecessary "seriousness" to the game, but in the other hand wouldn't you be glad to know you would playing with 31 other players of somewhat equal skill, all playing at least a bit seriously?
Moreover, you would play with the same players from the beginning till the end. This would add a lot of teamspirit and knowledge of other players' behaviour/tactic/skill, etc.



I do like ranking systems, but much like you I don't like the ones which use individual points: the spirit of counter-strike was much ruined by this and while the TF2 system had some endeavouring ideas, I think it's far from being a compelling or good system.



In the Myth series, you could have up to 32 players in a game, if I remember correctly. A game would usually take up to 15-30 mins, but could even go on as long as 1 hour. The point I'm making is that a game with a fixed starting and ending period, with a lot of players involved, is very much feasible.



I still think the win/lose ratio is the way to go. Otherwise you would have stat######s going for other points than the ones needed for the team to win...but maybe some kind of sub-categories could be made, I'm not sure of what to think of that.



Could you give an example of some specific game?
Because talking from my own experience, I have seen far better teamplay in ranked games than in unranked games. This is the case in all of the RTS games I've played, but also in other genres as RPG's (i.e. Diablo 2) or FPS's (i.e. clan-based matches in Counter-strike).



I'd like to know in which servers you play Because in almost all the servers I've been in there is never enough medics, if any. People in TF2 rarely care about the points (since it's based on a simple session) and just play whatever classes they want. They somewhat try to get a teamwin, but often enough I see people just happy fragging other people and having no consideration whatsoever about capturing or defending points.
Imagine you could have unranked and ranked games in TF2...now you could have some ludicruous fun in unranked games in the same way you do now, but also you could try to play more seriously in a ranked game. Once people would have something at stake (winning/losing points), they would most certainly make a real teameffort to win the game.



I very much hope that he does, because for my part, the best games I've played always had excellent ranking systems. I like the challenges they offer and the sporting communities they generate. Also, these kind of games are usually the ones that last (and have successfull sequels, since they already have large fan bases). Take for instance Myth II, which was published in 1998; it still has 3rd party game servers up and running and a large community backing the game up.

I was in the Counter-strike clan "Killer Penguins" and we won a few tournaments back in the Half-Life 1 days, the French Cup of 2002 being the most important one. But, all in all, it's rather hard to earn a good living as a pro PC player in Europe, so I switched to poker games. Been playing professionnally for five years and am part of the Poker United team.



For one, the fact that I'm for a ranking system in Natural Selection doesn't necessarily mean I intend to be high up on the leaderboard (nor that I think I will be high up in skill). I just want a leaderboard in order to have an enjoyable gaming experience.

As for what you said about professionnal players, your statements are quite wrong.
First of all, skills usually match leaderboard scores very well. I hate to brag -ummm, no actually I don't - I was ranked second on 1v1 and first on 2v2 on the European server of Warcraft III and considered myself quite skilled. I also was the player with the longest consecutive number of days as number one in the Myth series, and was undeniably the most skilled player in that game.
Second of all, pro gamers usually play games which do not even have any decent ranking systems, that is to say mostly FPS games (the same goes for poker games). The few RTS games which have a pro community usually have a leaderboard with the best players in the top 10 (i.e. Elky in Starcraft II).



I have never played Battlefield 2 and am unaware of its ranking system, but if what you say is true then its ranking system is simply uneffective. If teamplay was encouraged as it should, the top players should be skilled enough to be able to produce excellent improvised tactics for teamplay.

I'll give you two counter-example: me and the second top player on the 2v2 leaderboard of Warcraft III could probably have beaten any clan-based team. In Myth II, I and four high-ranked friends once accepted a challenge for a 5v5 game coming from the best clan there was and with no kind of preparation in advance we were able to beat them.



I assume you are once again referring to games which do not have any decent scoring systems. FPS clans are playing matches outside any kind of leaderboard exactly because there isn't one that is working effectively.



What's your idea of an alternative system? At least I can't figure out a better system for balancing the game correctly and to enhance other aspects as well (i.e. the regulation of mischievours).



My definition is simple: a person who earns enough money for a living by playing games.
Most so called "pro gamers" are adolescants who earn a few bucks but could never make a living out of it...that's why most of the real pro gamers went for easier money in poker a few years ago (i.e. Elky who plays in Pokerstars nowadays).

I don't doubt though that the pro community will keep growing in PC games but the pro gamers still will remain very few in my opinion.





You're wrong, such systems do exist. There are RTS games which allow large team games as well. Supreme Commander allows up to only 8 players, but as I said before, the Myth series had up to 32 players.



Not if the match-making is good enough. You could have an automatic match-making (giving each team as many bad and good players as the other team) or some kind of manual selection. I'd be in favour of the latter. Let people choose with whom to partner with, knowing that if they have only good players in their team against a bunch of bad players they risk much more points than they could possibly win (i.e. in the Myth series, if you had too unbalanced teams, the good team would win only 1 point and risk losing 32 points).



Hence, it's more like an FPS where individual skills are measured.



Same response as for La Chupacabra concerning unranked games and prohibiting free-loading and droping.

There's a bunch of problems with his suggestion, three of them being major:
1. In order to have a functional server, you need admins.
2. In order to have enough players on a specific server (not even to mention a full server), you need hundreds (probably even more) of reserved slots and you need to continously create new ones.
3. In order for a player to play on a (popular) competitive server, he needs to get a reserved slot, which could prove difficult for many reasons.


Same response as for La Chupacabra concerning point categories. In order for players to look top tier, they need to have points in all categories.





I don't see the problem with this though?
As Firewater said, these kind of players don't care about their rank.
Also, as he said, these kind of players don't play a lot on public servers. In other words, their skills wouldn't interfere often with the match-making system and, in any case, only at beginner levels.





I guess you haven't seen my ###### yet



Thanks, finally someone who agrees wholeheartedly!



OK, that might be the case, but I thought that as something rather obvious.





I really don't see how it can be so difficult to come up with a functional system: just look for what I said about distributing points to different categories in response to La Chupacabra.





It does decrease, but every player still has an impact. Sure you may need to play a few more games in order for it to show on your ranking, but the scoring is at least functional.
And I speak out of experience (especially with the Myth series), the system does work. The best players have a huge impact, by distributing orders and having some sense of teamplay, and usually players of lesser ranks listens/follows them. Also, since the match-making should create pretty even teams, each game should be very interesting and come as a challenge for all players. The outcome would never be sure, as it would depend on how every player finally surpassed themselves individually.



PS. Sorry for the time it took me to reply to any of you, I've had a really busy month!

Well, I've gotta say, he's pretty damn good.

I'm quite convinced, but I'll reserve judgment for after a few more replies.

Ok - how much is a good fade worth versus a good gorge or a good lerk or a good commander? How do you go and quantify these things in-game to make up the grand total for "skill"? What about the worth of a person who welds versus a person who likes to rambo and solo skulks? A person who drops chambers and plays as a skulk versus a person who saves up res for onos?

I'm sure there's a simpler way to give a rough semi-quantitative measure of skill than this.



Great, now how would we go about defining the rank difference in NS that is "prohibitive"? When players join a server, will they sign a virtual agreement to only play as a commander (or not to play as a commander), only as a fade (or not as a fade), only as a marine (or not as a marine, or only using certain weapons)? If we are going to split the individual skill into subsections for commander, alien, and marine - then when we combine them back together again (or even if we don't) how do we know who is going to play as what? The only saving grace I can think of for this is that really skilled players are just better overall at everything - then a value could be assigned. However, what level of "better" warrants restriction of entry to servers? I'm not really comfortable with this idea and never have been.



The TK player won't be the only one suffering from this. If stats are somehow based on wins, then the entire team's stats would be affected. Now that I think about it, that would be a very useful way of decreasing one's rank to get access to lower ranked servers.




You only need reserved slots once the server becomes popular. I don't see how these points that you have presented are problems.

In short -

Players quitting IS a problem, because you can't just arbitrarily "punish" quitters. Out on the pub circuit, you will find that there are times where players simply MUST quit because the game has become deadlocked at endgame. I'm not talking the midgame struggle for supremacy, I'm talking Alien players who cannot get into the Marine base to cement the win. On the flip side, there are people who would F4 when the HA Train pulls into the last hive - especially if marines start nobbing around.

Yes, this is a issue related to skill, tactics, and common courtesy - but you can't just slap a punishment on it. Not at a global level.


With regards to your points system - Its all very well saying 3 main categories and possible subcategories...... but have you actually looked at it in depth? "Team" is very nebulous. As for subcategories..... aren't you kind of crippled by the choice of your Commander? I mean, you might be hot stuff with the shotty, but if your Comm is HMG mad then you're going to have fun trying to pad that stat. Always keep in mind that this isn't Halo, this isn't SupCom, this is a team based game where offensively weak players can play a phenomenal support role. You don't get to dictate upgrades, they're made for the team. If this means you can't pad your personal K:D, unfortunate for you.

Further, you cannot just slap on an arbitrary "top tier" requirement of, say, kill 300 marines, have 300 build points, have 300 assists, and have dropped 100 hives. NS doesn't work that way. I concede that top tier players are expected to have a bit of flex in their play, but noone has the time to chalk up those sorts of points. Not in a way that cannot be exploited for statpadding, thats for damn sure.

Firewater said that certain players wouldn't care about their rank, but I can assure that players will very much care about their rank if it is being used to debar them from entering certain servers. They may not often play on pub servers, but that is entirely different from being incapable of playing on pub servers.


I for one would not appreciate a system that thoroughly encourages statpadding as a "golden ticket" into any server you choose. We already have enough problems with people ramboing, with people playing very poorly at the team level and then justifying any loss with a cry of "Well I've 10:1 K:D so I did great". We really don't need a Battlefield situation where people are running Knife Only servers, or related madness, purely to earn badges or look 1337.

Whether or not the ranking system is implemented, the deadlock situation should be taken care of for Natural Selection 2. It's up to the game developers to make it nigh impossible to drag on a game forever.
So the players won't have a reason to quit anymore, which means they can be punished...



I'm not sure what you mean by "nebulous" when the system is pretty clear and simple. In a nutshell:
Step 1: you start with 0 points in total and 0 points in each category (let's say Marines, Aliens, Commanders, to keep it simple)
Step 2: you win a game in which you played marines and get 14 points. This means you get 14 points in the Marines category and 14 points in total, the remaining categories remaining at 0.
Step 3: you play a few games as a commander and accumulate 61 points for the day. This means you still have 14 points in the Marines category, but also 61 points in the Commander category. Thus, the total is 75 points, which represents your rank.




Well the reason there would be subcategories would exactly be to push you to play all weapons and not just the one you excel in. And I never said, subcategories needed to be related to weapons. It could be subcategories like "damage dealt", "structures destroyed", "assist", etc.
And, in any case, subcategories wouldn't even be necessary (maybe even troublesome), as I've already hinted before.



Pretty much all games have a need for support. I've never played Halo, but I'd guess snipers/cover is needed there also. And you obviously haven't played a lot of Supreme Commander to suggest that there is no support role...



K:D? This ranking system is no way related to that ratio, but only on team wins:losses.



I think you've completely missunderstood what I've tried to explain. There is no statistics based on marine kills, buildings or assists, except if there should be some kind of subcategories related to these elements.

top tier requirement is to have points --> points are gained by winning games as a marine, an alien or a commander --> points could eventually be distributed in subcategories such as kills, buildings, assists, etc.
But you still would get points only if your team wins...in other words, the main goal is to make your team win, but your points could be awarded/substracted to/from you according to the role you played in the win/loss (that is to say as an offensive player, as a support, etc.)



This bit I don't understand...I mean why not? Did work on the Myth series...besides, this system doesn't involve countless of hours of play unless the game developers decide otherwise.



How come? I mean explain me how this ranking system could be exploited?



Now you're confusing me even more. Who said anything about restricting players from playing on public servers?
Unless you are referring to Radix's Reserve Slot system, the only suggestion even remotely close to what you said was that teams could be matched to each other depending on the ranks of their players. So are you thinking that beginners can't play against high ranked players? Although it would make sort of sense that players of both teams had similar ranks, the teams could -and probably often would- be mixed with higher and lower ranks.



You finally lost me...the ranking system I proposed is supposed to encourage teamplay. So who are you replying to? Me or someone else?

All this said, keep in mind that NS could simply have two kind of games: unranked and ranked. This way everyone should be content and able to join the type of game he wants.

If you're basing this all on wins, that's even worse than what I thought it would be.

Problems that come to mind:
1. For the commander: how do you tell if it is the commander's fault that the team lost or if he just got a bad marine team.
2. For the aliens: if stats are based on wins, there is nothing that will tell you what role a particular alien player is going to take in any particular game. For example, there are many players who are piss-poor at skulking but can lerk very well or gorge very well.
3. For the marines: a mirror image of the problem with the commander
4. Incentive - not everybody plays to win, nor should they be forced to in order to maintain some kind of rank
5. How to distinguish between the contributions of players in each game? Who actually carries the team versus who gets carried?

Fair enough. My argument however is that those players who are truly high up in skill will be misrepresented in the stats system, and will completely break it if a few players join a game.



*buzzzzzzz* incorrect. In team games such as ET:QW as well as the battlefield series. If you took the top 8 or 10 of those players on the leader boards they would NOT be the same as players from the top teams in that game. If you took those same leaderboard players and put them against teams that are not on leaderboard, they would get destroyed by the competitive teams. Imagine that same concept applied to your stats system and server balancing based on that.





The ranking system in battlefield 2 is just fine and dandy. Want to know why? The system does not attempt to balance teams based on leaderboard ranks. The leaderboard simply provides data on who accomplishes what. The top players in a pub system WILL NOT be able to compete against the top players in a competitive situation. I would feel slightly embarassed if I could not understand that concept.



"Could have probably" means didn't correct? I have no comment for the Myth competition as I have no idea what you define as "high ranking team" and how the leaderboard was run. For all I know you could have played a bunch of people that were posing as a top team, but really a bunch of scrubs. Was this top team sanctioned by any league? Or were they just self proclaimed.

We're a bit off topic though, no? Because those you depict as high skilled and pro gamers represent a very small percentage of players, which don't play on public servers anyway, right?



Pretty much same as above.



Well I hope the Natural Selection developers take a look at it then.



But does that mean players are virtually forced to play whatever they're good at?



Please don't be cocky, I understand very well what you're trying to explain. I know how important strategies can be and especially how effective they are if well honed in advance, but I still think you are being a bit presomptuous. Plus, you weren't talking about the "top" players in a competitive situation before, but only about a random team.
I don't know how many players there are in Battlefield, but from what I gather a lot more than in Counter-strike. This said, in the latter at least, one skilled player in a 5-person team can make a big difference individually and eventually compensate a lack of teamplay. I'm pretty sure that, if you took the 5 best players of Counter-strike in the world, they could beat any team out there with little, if any, preparation.



No.
"Could have" is the past tense of can used here to indicate ability in the past.
"Probably" means most likely.



The Myth series had its incorporated system for individual and clan rankings. They also had regular championships, the most important one being the Myth World Cup. The team I beat was called BMF, which was number one in the rankings for some time and came 2nd in the Myth World Cup of 1999.

Flatrick, the devs will never be able to engineer a legitimate undeadlockable game scenario (no, I don't think a fixed time limit counts). Let me just clarify that deadlocks are not strictly speaking "deadlocked" but they do drag a game out longer than it should. The problem is that "just one more attack" takes more than just one more attack. You can't really predict if the attack is going to succeed - sometimes you get lucky and they're all armory humping, or they decide to charge out and get ripped apart.

Unfortunately this is a skill issue and not a balance issue - most bases can be cracked open by some dedicated Xeno skulks and a few well supported Onos. Hell, even a well managed Onos train can win by simple careful picking of their digestable targets. But I digress.

Player will also always need to quit. Either to end a stale game, or just to go off for dinner. Or, in rare cases, because they've joined a very green server and realise they're outclassing everyone. You can't just arbitrarily punish players for leaving.



By "nebulous" I mean very very very very expansive. How does one define "team"?? Is it by going Fade and killing all the Marines? Is it by dropping Shotties or HMGs? Where did I get those 14 points?

Are you suggesting that an excellent Commander (that can't shoot for crap) is equal to a welder monkey? Is equal to a flex marine who followed orders? Compares to the Rambo who shot skulks in some back alley regardless of the mission objective?

How do we weight the points? A Commander might decide to turret up for easy kills, so does he get points for that? Does it only count if he wins? What if he sends his men into the meatgrinder and doesnt med or cat them? What if he never spends a penny on upgrades because he's making points on his turrets?

Why should I Gorge and drop a Hive when I can go Fade and let someone else worry about it? What is stopping me from dropping a strategically poor chamber first because it benefits me? How are you even going to detect what the most strategic chamber is?

Does NSPlayer get a giant killer bonus because he managed to knife Firewater? What if he only got the last hit in after FW ran into a pack of HMGs?

Yes, anyone can understand that "you do stuff, you get points", but the nebulous part is deciding exactly what that stuff is and exactly how much it is worth.



To continue, lets examine your subcategories such as "damage dealt", "structures destroyed", and "assist".

"Damage dealt" - To what, with what? Kind of unfair if you're packing a Shotty when your buddies have HMGs or two dudes with grenade launchers. Let alone the poor Commander in his chair. Again, what happens if the Comm drops the "wrong" weapons?

"Structures destroyed" - How much is each worth? Do you go by their usefulness or their ability to end a game? Obs? Chair? IP? Which is more vital? Do I get massive points for taking down a Wall of Lame? Doesn't this somewhat penalise the Alien team? Do you get a point for attacking it, or do you need to do X damage to get a point? Do the points go to whoever deals the killing blow? At what point does it become an assist??

"Assist" - What, exactly, is an assist? Healing? Building? So that means Gorges top the leaderboard?? I can guarantee you that in most pub games there'll be more healing than killing. Is it Lerkspam? A weldermonkey with zero situational awareness can get more points than the teammates he endangers? A Commander who medspams a hopeless situation?


These are all things that need to be rationed out. Classic isn't just quite as simple as Combat, and if ranking is to matter than the system must flawlessly reward the best teamplayers (which Combat arguably fails to do).

As for Halo and SupCom, barring what appears to be your deliberately offensive ignorance -

Halo is primarily not an objective based asymmetric team game. Barring hitbox oddities, the players are all the same, the weapons are all the same, and the player CHOOSES what to equip. Not the case in NS. Not by a long shot. You can't TrueSkill cooperative objective-based teamplay. Not with any great ease, I can assure you. As for the "team" aspect of Halo... its more team deathmatch than team objectives, sniper cover is only of certain benefit on certain maps, no medics, bugger all in terms of support other than "Cover me while I try and hit that tank with the Splaser".

SupCom - Ok then tooltip, lets do a quick clarification. I did not say there was no support game. Second, since you're playing at "gaming expert", let us hear how SupCom rewards green players who team with skilled ones? No? Are you remotely aware of how an ELO rating system operates and what it requires in order to work? While we're on the subject, how EXACTLY does SupCom reward pure support players? If you wanted to pick a good example, you should have tried World in Conflict, which at least rewards supportive players ingame in real time.

Back to the subject in hand, here's the kicker - Neither Halo nor SupCom actively reward pure support players. Thats not to say that support doesn't exist, rather it is optional, you gain no benefit - wholly different to NS, which demands support play.




Here's the worrying part - you can't just rank NS on Wins:Losses. It's not that simple. A bad Comm screws his team, you don't penalise the team for his screwup. Likewise a good Comm should not be held accountable if his team is incompetent. Now, if you can't penalise the Marine team for their W:L, then FOR BALANCE you cannot penalise the Alien team (despite the Alien players all having an equal share in their win or loss). Otherwise noone plays Marines, for fear that a bad Comm snafus their rank and gets them kicked off the server.

To continue, a Win is a TEAM WIN, and reliant on the other players. This goes against your holy cow ranking system, because ELO would rely on everyone on each team having roughly the same skill - which just isnt how NS works. As we have tried to point out to you, NS rewards supportive players, players that complement their team. I'm not sure how clued up you are, but in NS it's very hard to round up 12 players of equal skill. If you can, you're in league play, and in that case you don't NEED a game-based ranking system.


"Completely missunderstood" - afraid not, old bean. See, subcategories are going to have to cover this stuff, and that means stat padding. If you only get points when the team wins...... then why play Marines? All that reliance on a Commander, you could play well all round and end up losing for no gain. Whereas with Aliens, there's no Commander and no problem. You further state that points are lost or gained on performance, so again we go back to having to decide what counts and in what situation.

Certainly, L4D seems to reward on a situational basis, but then it only has 2 sets of 4 players doing roughly the same job. Not quite the same as 2 different sets of 6 players pursuing different goals.



Let's just reflect for a second. In the space of one post, you say that subcategories don't count, but that they do count. You've said K:D doesn't count, but that it will count. You've said that its all about W:L, but that the points will be handed out even if you lose. Forgive me, but your idea seems to be rapidly losing coherence.



You don't understand why NS teamplay doesn't work the same as others, by your own admission. This is where the problem lies, chief. Not everything ranks like Chess does. NS allows a player to act defensively for the entire length of the game, and receive the same benefits as the most aggressive player on the team. It allows for a bunch of skilled Marines to save the day despite their middling (and sometimes meddling) Comm. RTS and FPS games reward cooperation, but there are very few that reward PURE support.


Taking an ELO system as an example - if pubs play league players and win, their ranking shoots up exponentially. If the league guys win, their ranking barely moves. The reverse is true when considering losses (assuming other players are playing equals). Now break that down to an individual level - why would a high tier player want to jeopardise their ranking by potentially losing to a green recruit? Further, a league player plays only a fraction of the number of "proper" (and by that I mean gaming outside of training) games than a pubber will. This only further stacks the ranks in the pubbers favour, since he has more games against a wide variety of opponents - everything to gain, and nothing to lose. Conversely, the high tier player has fewer games, and against players of equal skill - giving them a fairly stationary position but one prone to catastrophic failure if they are unlucky on a pub server.


I mean, in one fell stroke we manage to turn a system of segregation for enjoyable gameplay into one where the wedge is driven more deeply than ever before. Its just completely counter to its intended purpose!




Exploiting the ranking system - picking your games so that the loss is minimal, the gain is maximal. Spending the game racking up individual kills rather than supporting your team. BY YOUR OWN ADMISSION this is taken into account in "subcategories" and counts for, or against, you when the round is over. I should think the result is pretty clear.

As for the restricting effects of rank - the whole principle of a ranking system is to allow balanced and enjoyable matches. Second, this forum is discussing rank restriction on servers, so one must highlight the flaws in THIS system if it were adopted for generating rank.

If ranking serves no purpose in server balance, then there is no point to having ranking. People have leagues to tell them how good their team is compared to others. If ranking is there "just because", then its a wang competition, and NS doesnt need people that think their rank is more important than their team.

Here is the take home message - NS should allow all players to play against each other. Individual ranking is at best an indicator that allows Admins to balance their servers in terms of skill. Ranking should allow for enjoyable games across a certain range. Adopting an ELO style system effectively DEMANDS that players stick with their own ranks or higher, and play fairly predictable games for fear of losing a GLOBAL rank.

Please compare and contrast with other systems that adopt a local, server-based approach and do not globally penalise a player for choosing to play with lower skilled buddies.

As for ranked and unranked...... Last I was aware, players tend to prefer the Ranked servers to Unranked ones. Probably because there is much less mucking about, and certainly because Ranked servers tend to give skill/exp/item perks over time.




In summary and conclusion - the ranking system you propose has not been fully fleshed out, you have made contradictory statements within the same post, and on examination it appears that the system would in fact drive a deeper wedge between dedicated and casual play. While your intentions are noble, the end result cannot be countenanced.



EDIT - Little more work needed.

This is escalating into a drawn-out essay writing contest with no clear description of what is being discussed.

Wrong, NS was a smaller game than CS and there were not scrims available 24/7, and even if there were teammates were not always on as well. These competitive players would often pub if they could not get a scrim or match going.



Here is the full quote "The ranking system in battlefield 2 is just fine and dandy. Want to know why? The system does not attempt to balance teams based on leaderboard ranks. The leaderboard simply provides data on who accomplishes what. The top players in a pub system WILL NOT be able to compete against the top players in a competitive situation. I would feel slightly embarassed if I could not understand that concept."

The ranking system that ranks players is great. As far as it being used for balance, it would be an AWFUL idea.




The 5 best players in counter strike would not be on any developer created leaderboard, your point falls apart based on that fact.



The fact that they don't care about their rank jeopardizes the whole system. Competitive players would be "under-rated" according to this system, but would shake up the system enough to disrupt the balance in the game, thus creating more false statistics. This is part of the reason why a leader-board style of balancing fails. See Sarasiel's post for more reasons why its flawed.