sc2 match making style


(Apples) #161

Matchmaking wont change much I guess, as usual the people accustomed to stack together will do, and even a slight différence in range (say there are 8 “70” player and 8 “50” player) will result in all the “70” players stacking against the “50”, making an unbalanced game.

You cant go against this, except with good admins, judging “skills” is too subjective to be a simple calculus made by a bot IMO.

Well it’s still a nice try and will reduce the deltas beetween teams, but rollover and “full” hold (not really full as obj changes during the map if not completed) will happen, as in every video game since the creation of “team” shooter.

Peace


(MrX) #162

They should use a mechanism to balance teams according to a rating system similar to starcraft.Most pub games are ruined by experienced ppl who join one side and if the opposite side has more newbies they just dont change team and continue to pwn them.That is why i dont like pub games in general.There is no balance mechanism in public games and brink should implement one.
Now regarding the matchmaking system a good idea would be to have ladders only for clans and seperate divisions just like clanbase works.You and your opponent team choose server join the game and the match starts when they are ready.I insist on choosing a dedicated server cause if matchmaking chooses for u then its a big failure for me.The server streams directly to the ladder it adds the win or loss according to the result.


(Bullveyr) #163

I wasn’t describing an exact system, just some ideas to solve your mentioned problems.
As I said I dunno how well current systems works for that specific issue but I’m sure it’s solveable, it’s also mentioned in the MS TrueSkill FAQ

Number of games to determine your skill

Q: I have been playing a lot of unranked training games and I think I am now a much more skilled player. Will the TrueSkill ranking system be able to identify my new, higher skills? If so, how many games do I have to play before the TrueSkill ranking system knows my new skill?

A: The TrueSkill ranking is assuming a small skill change between any two consecutive games in a game mode so it is able to identify your new, higher skill. But, if your skill has completely changed (you became the best player in the world from previously being the worst player in the world), then you would need to play a large number of games. We designed the system such that it would need between 50 - 100 games before the system would be able to track a substantial skill increase/decrease.

Q: If I understand the TrueSkill update formula correctly then the change in μ is largest for the first few games and decreases over time. Thus, my first few games are most important; if I lose these games, it will take the TrueSkill much longer to converge to my skill. Right?

A: Not exactly right. It is correct, that the change in μ is getting smaller and smaller with every game played, but regardless if you win or lose them. However, TrueSkill always takes more recent game outcomes more into account than older game outcomes. Hence, when playing against a set of players of same skill multiple times, a late win counts more than an early win. As an example, try the following in the interactive rank calculator (we will choose Alice for the analysis and assume a draw probability of 10%)

I don’t see where this is more abusable than your idea of ranked servers.

@Apples

Sure, it will not make all games perfectly equal but at least more equal than most current matches.


(Auzner) #164

The fact this thread is now 9 pages means there is a lot to discuss on this matter. My posts on the first page pointed out this is exactly why there won’t be a system from the start. There is a lot of time required to consider what would work and even more to implement it.

I get the feeling the greater concern held now is if they did implement a system, would it be agreeable? Another point I made on the first page is that it is very uncertain since Brink will be quite untraditional.

So now it’s been back and forth with “do what this game does, but change…” and “no, I can think of factors that make that less ideal.” SD has the game picked apart the most and knows every factor in there. Right now with the game unreleased, they’re the only ones who can conceive something suitable. If not Coolaguy himself, they likely have someone just as discerning for this matter.


(MrX) #165

[QUOTE=Auzner;265968]The fact this thread is now 9 pages means there is a lot to discuss on this matter. My posts on the first page pointed out this is exactly why there won’t be a system from the start. There is a lot of time required to consider what would work and even more to implement it.

I get the feeling the greater concern held now is if they did implement a system, would it be agreeable? Another point I made on the first page is that it is very uncertain since Brink will be quite untraditional.

So now it’s been back and forth with “do what this game does, but change…” and “no, I can think of factors that make that less ideal.” SD has the game picked apart the most and knows every factor in there. Right now with the game unreleased, they’re the only ones who can conceive something suitable. If not Coolaguy himself, they likely have someone just as discerning for this matter.[/QUOTE]

The main factor i dont know if its stated before is that there is not going to be a similar system ,cause the game is multiplatfom and requires huge amount of time to be implemented on 3 systems and beta tested


(tokamak) #166

[QUOTE=Auzner;265968]ere is a lot of time required to consider what would work and even more to implement it.
[/QUOTE]

Then again, there’s a thread on this almost as old as the brink section itself.


(Atavax) #167

[QUOTE=Bullveyr;265951]I wasn’t describing an exact system, just some ideas to solve your mentioned problems.
As I said I dunno how well current systems works for that specific issue but I’m sure it’s solveable, it’s also mentioned in the MS TrueSkill FAQ

Number of games to determine your skill

I don’t see where this is more abusable than your idea of ranked servers.
[/QUOTE]

first I’d like to point out that it appears that this system is for 1v1, in a teambased shooter you cannot evaluate each player by if they won or not, the system would have to be much more complex and would probably take even more games to figure out a large skill increase, much more then the 50-100 your source listed, which is pretty unreasonable to start with.

lets think of some abuses… ok, there’s good player intentionally losing matches to be able to play against worse players. and instead of only bugging the worse players, he’s going to be bugging everyone in the matches he intentionally loses to get to play against the worse players.

there could be people in a server intentionally trying to lower a single guy’s ranking.

there would also likely be people buying other people to either increase or lower their skill lvl to save themselves the trouble.

If you make this huge effort for a central matchmaking system, it means you can’t have good tools on each server… So unlike a system where you leave it up to the servers to police themselves, you wouldn’t be able to deal with the people playing way too high or way too low for their skill set. In a ranked server system where all the power is in the server, you find a server with admins with similar values as your own, and as soon as someone goes against those values, and tries to grief, he gets kicked or banned.

If you try to make a universal formula that is going to work for every server and every group of people, just common sense should tell you its going to be way more subject to abuse then a system where each individual community develops their own rules and have the power to enforce them on their own servers. The idea that there is a perfect matchmaking formula that fits every person’s values for gaming in that game is idiotic… There are no perfect set of rules that will make everyone happy, you have to just let small communities form their own rules and methods for balancing teams.


(tokamak) #168

first I’d like to point out that it appears that this system is for 1v1, in a teambased shooter you cannot evaluate each player by if they won or not

In an objective game with different classes you can’t evaluate players period. The number of games they’ve been on the winning side (provided teams are locked) is still the best indicator you can have.

lets think of some abuses… ok, there’s good player intentionally losing matches to be able to play against worse players. and instead of only bugging the worse players, he’s going to be bugging everyone in the matches he intentionally loses to get to play against the worse players.

Happens in Starcraft. Found myself in one game that apparently was on a live stream, a big US player on an EU account showing off silly tactics.

I don’t see the problem though. This guy can’t win too many matches against lower players before being placed back where he belongs.

If you make this huge effort for a central matchmaking system, it means you can’t have good tools on each server… So unlike a system where you leave it up to the servers to police themselves, you wouldn’t be able to deal with the people playing way too high or way too low for their skill set. In a ranked server system where all the power is in the server, you find a server with admins with similar values as your own, and as soon as someone goes against those values, and tries to grief, he gets kicked or banned.

This is absolute nonsense, nobody here wants a rating system enforced on all servers. SC2 doesn’t have that either.


(Bullveyr) #169

It works allthough, as I allready wrote, it would be great if someone would find a way to take personell performnace into account.

ets think of some abuses… ok, there’s good player intentionally losing matches to be able to play against worse players. and instead of only bugging the worse players, he’s going to be bugging everyone in the matches he intentionally loses to get to play against the worse players.

Possible but shouldn’t be a real problem because simpllified for every game where he can noob-bash he must lose one intentionally.

there could be people in a server intentionally trying to lower a single guy’s ranking.

Paranoia?
Besides it’s not that you need an ELO system to ruin the fun on the server for everyone, we allready have them and we would deal with them same way wel allready do.

there would also likely be people buying other people to either increase or lower their skill lvl to save themselves the trouble.

Allthough possible but do you really think that it will be done in such an extend that it causes a problem for the rest of the gamers?

If you make this huge effort for a central matchmaking system, it means you can’t have good tools on each server… So unlike a system where you leave it up to the servers to police themselves, you wouldn’t be able to deal with the people playing way too high or way too low for their skill set. In a ranked server system where all the power is in the server, you find a server with admins with similar values as your own, and as soon as someone goes against those values, and tries to grief, he gets kicked or banned.

If you try to make a universal formula that is going to work for every server and every group of people, just common sense should tell you its going to be way more subject to abuse then a system where each individual community develops their own rules and have the power to enforce them on their own servers. The idea that there is a perfect matchmaking formula that fits every person’s values for gaming in that game is idiotic… There are no perfect set of rules that will make everyone happy, you have to just let small communities form their own rules and methods for balancing teams.

I thought I made it clear that I’m not talking about a forced system.

You choose if you wan’t that system running on your system (or to what extend), the server can still police itself the same way it does allready.

Nobody says that such a system is perfect but the current system by far isn’t either.


(Atavax) #170

[QUOTE=Bullveyr;266322]
I thought I made it clear that I’m not talking about a forced system.

You choose if you wan’t that system running on your system (or to what extend), the server can still police itself the same way it does allready.

Nobody says that such a system is perfect but the current system by far isn’t either.[/QUOTE]

right because people choose in the CODs whether to buy the map packs in that game… its really simple in them, you either buy the map packs or you stop playing the game. If there is a volunteery ranking system that some servers require, you’re either going to use it or you’re not going to play the game.


(Atavax) #171

[QUOTE=tokamak;266297]In an objective game with different classes you can’t evaluate players period. The number of games they’ve been on the winning side (provided teams are locked) is still the best indicator you can have.
[/QUOTE]

no, the best indicator is peer review, but you and other people pushing formulas to judge performance are so afraid of peer judgement that you push for a clearly inferior system. Why don’t justice systems the world over just develop algorithms to determine what is and isn’t enough proof to be convicted of a crime? because its clearly inferior to peer review, and using algorithms in criminal justice to determine guilt is just as ludicrous as developing programs to determine player skills in shooters.


(Nail) #172

Why don’t justice systems the world over just develop algorithms to determine what is and isn’t enough proof to be convicted of a crime?

they’re not populated by teenagers


(tokamak) #173

Absolutely!

but you and other people pushing formulas to judge performance are so afraid of peer judgement that you push for a clearly inferior system.

Wait, what the hell?

Why don’t justice systems the world over just develop algorithms to determine what is and isn’t enough proof to be convicted of a crime?

because its clearly inferior to peer review, and using algorithms in criminal justice to determine guilt is just as ludicrous as developing programs to determine player skills in shooters.

… I may get this wrong but are you seriously putting a jury trail as a valid alternative to an ELO rating or whatever numerical system you use to determine someone’s rank within a game? It’s a bit convenient don’t you think? Proposing a system that would only work in an utopic dream world where everyone is concerned about keeping a system pure and willing to spend effort in order to achieve that.


(Atavax) #174

[QUOTE=tokamak;266347]Absolutely!
… I may get this wrong but are you seriously putting a jury trail as a valid alternative to an ELO rating or whatever numerical system you use to determine someone’s rank within a game? It’s a bit convenient don’t you think? Proposing a system that would only work in an utopic dream world where everyone is concerned about keeping a system pure and willing to spend effort in order to achieve that.[/QUOTE]

not a jury trial, tools in the system server that would empower the players to get rid of undesirable players from their server, because they are the best judge of who is detrimental to the server, not some computer program thats designed to define what is ideal behavior for 100% of the population. And tools so the server can tag itself to be easily findable by the type of players they want.


(Coolaguy) #175

This is false. Do some research. There are some great articles out there about how this is done… which leads me to the next point: [QUOTE=Atavax;266292]the system would have to be much more complex[/quote]


I like your acknowledgement of the fact that people will try to abuse and ‘game’ any system, regardless of its complexity or sophistication (see Addendum).

How does the rest of society deal with those who commit infractions against established systems? Detection and Enforcement. Splash Damage still has the discretion to wipe account statistics or ban user’s accounts who are suspected of foul play.

Developers Are Responsible For All Aspects of the Gamer’s Interaction with Their Product - Including the Gamer’s Interaction with the Gaming Community:

You discount the fact that the game developers have a real stake in what type of communities evolve in concert with their game’s lifespan. When the power to set matchmaking conditions, etc. resides with the server administrators then the developers can lose control over the type of experience that the player has.

Sometimes developers are willing to surrender that control to server administrators. Sometimes developers like to hand over the control to the community at large. Other times, developers place a premium on certain elements of the user experience and do their best to ensure that the player’s experience is strictly consistent with respect to those elements.

A significant reason that a default matchmaking system was implemented in Starcraft 2, is that Blizzard wanted to ensure a consistent experience for players when finding games. Their former Battle.Net system of searching for games was ‘broken’, as they put it. Did interesting communities arise under the old Battle.Net system? Absolutely. However, the system was less user-friendly and players had to be more savy when finding games. In the case of custom games, their former system was a disaster.

There are real problems with the current Starcraft 2 custom game implementation, but it is, nonetheless, much easier to find, join, start, and play custom games in Starcraft 2 than it was on the old Battle.Net.

One notorious original Starcraft and Warcraft 3 community arose with the development of “Aeon of Strife” / “Defense of the Ancients” (DOTA). This community has often been described as toxic. The DOTA game mode, itself, is incredibly fun and engaging, but the learning curve is steep and most games had the “No noobs” + “Only Pros” disclaimer, to the point where you couldn’t find DOTA games if you weren’t an established player. When a new player is finally able to find a game, they will frequently get booted because of their poor performance in-game. The DOTA community was an instance when the autonomy granted to server administrators and communities made the game inaccessible to players who weren’t a member of that exclusive club.

“Aren’t gamers and gamer communities entitled to playing the game how they want and to the exclusion of other players?” I think that in niche instances (or dedicated creative communities) then they absolutely should have that level of autonomy. If the community is toxic, then so be it.

However, as I’ve already stated players don’t necessarily enter a contract with the game’s community when they hand over $60 in purchase of a new video game. The gamer is making a purchase in good faith, and they are buying an entertainment product from the publisher/developer responsible for the title. The game is selling the promise of entertainment complete with the promises of features (as communicated through marketing promotions and crowdsourced reviews).

The feature we are discussing is the promise of multiplayer gameplay. As the developer, you have a responsibility to deliver upon the promise that you made and that the would-be player accepted in good faith. Other individuals or groups of individuals aren’t entitled to depriving the would-be player of that experience. The only way that they would gain the power to do so is if they were granted such power (by the developer) or if there were omissions in the game design that enabled them to have such power over the would-be player. Otherwise, it is the developer’s responsibility to make good on their promise.

It is easier to let communities manage themselves, because their demands and enthusiasm are tireless and they know exactly what they want. The manpower to appease a community is utterly insane. The only thing that could sustain it would be the community itself. However, the community didn’t sell me the product. The developer did. In which case, the developer is the middle man in all of my interactions with the community. They determine and guide the nature of all of our interactions. When more power is handed over to the community, remarkable things can happen, but not all of those outcomes are to the benefit of the would-be player, nor are they necessarily consistent with the promises made by the developer to the player.


[QUOTE=Atavax;266333]Why don’t justice systems the world over just develop algorithms to determine what is and isn’t enough proof to be convicted of a crime?[/QUOTE]We do have such algorithms. We codify it, and we call it Law, Legal Practice, the Legislative System, and the Justice System: What is a crime and what isn’t. What is admissible as evidence and what isn’t. The algorithm is roughly: impartial discretion and judgement applied to precedent as weighed against the body of evidence through due process.

Peer review relies on the computational power of the human mind to weigh the relative merits and come up with a determination of who is better. Peer review is computationally expensive and time consuming. The administration of peer review needs to be monitored in order to prevent collusion or the abuse of power.

Algorithms provide useful, time-saving methods to make determinations. For the sake of argument, the aggregate decision-making ability of peer review could be coded into the parameters of an algorithm and used to make predictions or to evaluate future instances in a manner consistent with the original peer review system. Obviously, such an algorithm wouldn’t have the breadth of scope that peer review enables, but the algorithm would be much more efficient when applied to a large volume of instances that only ever varied within a defined, predictable range.

What constitutes “the best” is entirely subjective, depending on what virtues you consider worthy of acknowledgement. That is why people can’t come to a consensus about who is or isn’t the best at a given sport.

TL : DR
ELO is NOT about policing players who are disruptive on a server (although there may be implications of varying ELO ranks upon the nature of the players themselves). ELO is about matchmaking (and ranking to a lesser extent). ELO is not necessarily about who is “the best”.
Tournaments are the way that we gauge who is the best, and even then “the better team doesn’t always win”. Tournament results are, ultimately, only accurate and relevant for that given moment in time under those specific tournament conditions.


Addendum:

Abusing Systems in Videogames and Sports:

In videogames, Griefing is much easier to do when ELO systems aren’t in place. When an ELO system is responsible for matchmaking, you would need to lose many games in order to be eligible to play lesser players. As tokamak noted, wins against lesser players will cause your rating to rise and you will start to be matched up against more challenging and more evenly matched opponents.

ELO matchmaking systems seem to effectively curb Smurfing. The only way that dummy accounts and smurfing are still relevant are when set matches are administered outside of matchmaking. Otherwise, the ELO matchmaking algorithm wouldn’t match a significantly lesser-ranked player with a higher-ranked one.

Halo 3’s ELO system was famously exploited by teams of players to help one (or more) of the team’s members reach the Skill cap of 50. Halo 3’s ELO matmaking algorithm created a team ranking by weighting each member’s ranking in concert. The team ranking was a blend of the individual player rankings. In this manner, you could game the system by having a number of highly ranked players on a team with an extremely low rank player. The low rank player would significantly shift the team’s overall rating down. Since the low ranked player was playing on a dummy account (a new account) but was actually a highly skilled player, the team was not only likely to be matched against lesser opponents, but the performance bonus awarded would be even greater than normal when victory was achieved against other players of high rank.

A Korean Starcraft league scandal arose when it was revealed that matches were fixed, by featuring a higher proportion of terran-favoured maps, to the benefit of an otherwise unaware fan-favourite terran player.

There are countless examples of match fixing in sport. It is a testament to human intelligence and enterprise that people will find a way to game any system. Does that mean that we should throw our hands up in the air and declare that since no system is perfect that all systems should be abolished? Of course not. We do the best we can. We iterate and improve, and we try to apply and regulate our flawed systems with wisdom and justice.

“Men must be governed. Often, not always wisely, I will grant you, but governed nonetheless.”


(Atavax) #176

there are two main points i disagree with you on.

#1 is that the developer is responsible for providing a quality multiplayer experience. and Developers Are Responsible For All Aspects of the Gamer’s Interaction with Their Product

#2 ELO is not about policing players who are disruptive on a server.

the multiplayer experience is largely completely out of the developer’s hand. Even if a developer made the best possible multiplayer platform, the developers can’t provide a quality multiplayer experience without the help of their customers because the customers are the majority of the multiplayer experience, without them, there is no multiplayer. It can’t be the developer’s responsibility to provide a quality multiplayer experience, because it is largely out of their hands. When you buy a multiplayer game you do so knowing that the quality of experience you will have is largely based on how well a community the customers have established.

ELO like many other things can have multiple functions. Finding someone or a group of people of roughly similar skill is relaitvely simple feat… the controversy and complexity of ELO’s are their policing.


(tokamak) #177

A rated system has nothing to do with policing. It doesn’t need any.


(Coolaguy) #178

[QUOTE=Atavax;266372]there are two main points i disagree with you on.

#1 is that the developer is responsible for providing a quality multiplayer experience. and Developers Are Responsible For All Aspects of the Gamer’s Interaction with Their Product

the multiplayer experience is largely completely out of the developer’s hand. Even if a developer made the best possible multiplayer platform, the developers can’t provide a quality multiplayer experience without the help of their customers because the customers are the majority of the multiplayer experience, without them, there is no multiplayer. It can’t be the developer’s responsibility to provide a quality multiplayer experience, because it is largely out of their hands. When you buy a multiplayer game you do so knowing that the quality of experience you will have is largely based on how well a community the customers have established.[/QUOTE]It would be unreasonable of me to insist that a developer is completely responsible for all elements of online interaction. My argument, however, is centered on the idea that ensuring a consistent player experience is paramount to the precedent of the more, shall we say, ‘coloured’ interactions that have been the standard of the online experience to this point.

Thus, what I am really saying is that the opportunity for developers (and other ~social engineers) lies in trying to curb the negative aspects and elements of online interpersonal interactions while still fostering the benefits. Videogame innovation in social engineering (e.g. MMO, Farmville social reciprocity, etc.) represents a growth opportunity for developers, just as innovations in game mechanics did (e.g. experience points, variable reward, etc.). This is where the market wants to go. This is what consumers want. Consumers (i.e. gamers) want a game that is remarkably fun to play; that they can play with their friends; and that frees them from the hassles of online interactions, if they so choose. However, this is not the current state.

My personal opinion on the backlash of Farmville is that it became stigmatized in the media. The social goldrush was too rapid, and when people caught on that all of these supposedly non-core gamers were spending excessive, compulsive amounts of their time AND their hard earned money on virtual goods, there was a story to be had. Zynga must be duplicitous. Video games had to be inherently wrong. Right? Certainly it wasn’t normal. Farmville gave its users everything they wanted for a time, but Farmville succeeded with the wrong market, because Farmville was more of a guilty pleasure for its users than it was something they could openly embrace. Farmville didn’t jive with the values of the mainstream audience that it appealed to, at least not with respect to their perceptions and opinions of online interactions. However, market perceptions are not static, but can be shifted. These perceptions can be brought in line with the underlying values of the demographic in question.

I’m not intimately familiar with the mechanics in Farmville, but it seems to me that they simplified and customized the available interactions to suit the Farmville agenda. For instance, you could either remain apathetic (a non-player or one who refused to help others to your own detriment) or your could take action and be helpful. As a gamer, you couldn’t sabotage, and you couldn’t send a “stfu, f@g!” or “u mad, bra?” messages. Furthermore, it was a game that you played with your friends… the real ones, not a bunch of dudes in your clan or on a server. Those relationships mattered to you, and you didn’t like the prospect of letting them down, even if it meant that you had to log in to Farmville to water their crops. (PS. - I’m not saying that you can’t form real-life friendships online, just that it is a dramatically different experience than real-life friendship and that forming long-term friendships online is not the norm of online interactions)

Now, there’s a fine line between embedding social mechanics in your game that encourage players to play and making them a slave to your online community. I think, sometimes, games like World of Warcraft and Farmville cross that line. However, the first challenge is getting the players to come. It’s not as simple as: “If you build it, they will come”. You need to be able to ensure a consistently beneficial player experience, especially when trying to appeal to all but the most traditionally core of online gamers.

[QUOTE=Atavax;266372]there are two main points i disagree with you on.

#2 ELO is not about policing players who are disruptive on a server.

the controversy and complexity of ELO’s are their policing.[/QUOTE]Your intended point, here, completely escapes me. I get that you don’t agree, but I have no idea what you are saying.


(tokamak) #179

Starcraft however, seems to have quite a lot of status even amongst non-gamers, no matter how weird the addiction stories from South Korea get.


(tokamak) #180

Pretty excited about this, an 11 winning streak. While for unrated games I really wouldn’t care less.