As someone who does this for a living, I always feel like fairness, balance, and skill are just tools. The thing that really matters is player psychology. The game doesn't actually have to really be fair or balanced. It just has to *feel* fair and balanced to the player.
Skill and balance are a means to that end, but I don't think they should be the primary focus.
It's tempting to think just making it perfectly balanced will make it fun, but that's only part of it. If the player experience isn't considered, it's easy to make something perfect mathematically, but boring or frustrating to players.
As Jeff Kaplan (former WoW lead) put it, Perfect skill matching would mean you lose half your games. Is that fun? Is there any other way to do it and be fair? Probably not. So they found another way to make it feel fun anyways.
Whenever I've played with friends who were used to RPG-type games where they're always the hero, they get confused when they play against real people and die a lot instead of easily killing everything like they did before. That makes me think designers should think about how they transition these players from killing computer opponents to real ones.
As far as measuring skill goes, a perfect skill-based system would determine your skill quickly and accurately and it wouldn't change much. That would feel pretty boring and probably wrong to a player.
So if your goal is to make a game fair and balanced, you're probably barking up the wrong tree. Instead, think of fairness and balance as some of the tools to make it fun, and focus most on the perception of the average player.
The counter-example would be Consumer Reports. The problem is people would have to be willing to pay for a subscription to read "unbiased" video game reviews. It might be hard to get people to pay for them when there are so many free ones and "meta" review sites like Game Rankings and Metacritic.
One thing that worries me about the comments on the Halo 3 beta are that the game modes "Land Grab" and "Territories" "Gets intense when time's running out, and only one or two markers are left to defend/capture." or "Very fun when it's tied 2-2 and everyone's going after the last one!"
What about the rest of those matches? Why can't it be "very fun" and "intense" the whole match? I'm thinking of games like Wolfenstein: Enemy Territory where all of the action is always focused, so it's almost always "very fun" and "intense".
I guess I'm just biased towards focused objective-play vs. splitting everyone one up around the map until the very end. Especially when there are only 4 players per team, you really need them all focused in one place.
I mean, it almost makes sense on a 64-player Battlefield server to split the players up among 8 control points, but you still have 4v4 action at each point. If you have multiple control points but only 8 players, things will of course be sparse until the end.
More complex modeling than means and standard deviations have gone into improving online games (unless the author was simplifying things by saying means and standard deviations).
For example both Trueskill and etpub use a Bayesian form of Arpad Elo's rating system to rank and match players.
I did some work modeling kills and wins in Enemy Territory that yielded interesting insights about map- and weapon- balance in that game.
At arena.net, there is at least one employee whose sole job is to model the association between skills winning probabilities in order to balance the skills across the classes.
I'd like to see an RPG that rated players and mobs based on statistical models and made sure the game always gave a near constant level of difficulty. RPGs present, I think, more difficult problem in this area than some of your standard first-person-shooters, etc., because of the variety you have in builds.
I'm a statistician / machine learning guy, so I am all for modeling aspects of games to improve them.
I think it would be even better to have the single-player "wander" into 2-team multiplayer competitions. The single-player may have a completely different goal than the 2-teams, but will feel like he or she is crossing through a more realistic conflict.
I'll use Enemy Territory as an example since I've played that more than most multiplayer FPS.
The single player has a mission to just "escape" by going from one end of the map to another. Or something simple like that. He has to start on, say, the Axis side of the map, maybe not even that far from the current Axis spawn, and make his or her way across to some point on the Allies side of the map.
I'm thinking of a map, like, tc_Base. The Allies are trying to Dynamite the Axis radar, meanwhile, Joe SinglePlay is trying to escape from the same compound. He will have to dodge through Axis soldiers who perhaps have some incentive to shoot the player.
You could just use current multiplayer servers, and drop the occasional single-player into the action as if the player "wandered in"
I think this would give the multiplayer players "something to do" besides sitting around and talking about the latest Stormtrooper rifle until the Jedi force pushes them down the shaft (Jedi Knight II).
The Offset Editor allows all aspects of the game and game content to be created and modified inside a single application. It is built from the ground up with the goal of decreasing development time in every aspect of game creation.
Also:
Editor Framework notable features:
* For designers that prefer to work close to the metal, all content is editable external to the editor via easily readable XML files
* "Hot loading" is supported for all content files. If a file, like an image, is modified externally, the editor will automatically detect the change and reload it instantly
* Game code can be run while the designer edits a map
I think the reason they're focusing on this guy is because he is or is one of Final Fantasy's original creators who left Square around the time it merged with Enix or when they lost all that money on FF: Spirits Within.
I think it's significant that Microsoft is making inroads into one of the largest RPG franchises (not necessarily FF, but it's creators) we've seen. Both with Sakaguchi and with the annoucement about FF XI.
Square switching to Sony from Nintendo seemed to be one of the major factors in the Playstation's success over the Nintendo 64 (Final Fantasy VII), and while this may not make the Xbox 360 come out the victor in the next bout of "console wars" it will, I belive, be an influential factor, and therefore any news about what's going in these aspects is interesting.
We run an Enemy Territory Server called Beginner's Park 3. We have a website at http://bpark3.com/. We use a slightly customized version of the systats package to track our stats. You can even chat "!stats" in-game on our servers (see links to our servers from the webpage) and get a summary of your stats including your ranking, accuracy, favorite weapon, etc.
systats is a fairly active project and has a really nice stat tracking system but it is very person centric. IMO, it's great for public servers but for clan matches games I think I'd prefer a program called StatWhore which is designed to summarize stopwatch games.
For our stats system, to try and deal with "stat-whoring"---where, for example, players allow themselves to be killed over and over by another to boost the latter's stats---we developed a weak secondary dominance statistic that basically awards more points to those who kill players with higher K / D ratios. This has helped to encourage the better or more stat-conscientious players to play against each other instead of ganging up on the weaker players.
So I recommend systats for public-type servers, it's very flexible, and I recommend StatWhore for matches. As for whether or not stats are important, I think they're fun as long as you don't find yourself worrying everytime you play about how it'll affect your stats. I think a good stats system should try and prevent that. We've tried to do this, but are still considering improvements.
Slashdot Mirror
As someone who does this for a living, I always feel like fairness, balance, and skill are just tools. The thing that really matters is player psychology. The game doesn't actually have to really be fair or balanced. It just has to *feel* fair and balanced to the player.
Skill and balance are a means to that end, but I don't think they should be the primary focus.
It's tempting to think just making it perfectly balanced will make it fun, but that's only part of it. If the player experience isn't considered, it's easy to make something perfect mathematically, but boring or frustrating to players.
As Jeff Kaplan (former WoW lead) put it, Perfect skill matching would mean you lose half your games. Is that fun? Is there any other way to do it and be fair? Probably not. So they found another way to make it feel fun anyways.
Whenever I've played with friends who were used to RPG-type games where they're always the hero, they get confused when they play against real people and die a lot instead of easily killing everything like they did before. That makes me think designers should think about how they transition these players from killing computer opponents to real ones.
As far as measuring skill goes, a perfect skill-based system would determine your skill quickly and accurately and it wouldn't change much. That would feel pretty boring and probably wrong to a player.
So if your goal is to make a game fair and balanced, you're probably barking up the wrong tree. Instead, think of fairness and balance as some of the tools to make it fun, and focus most on the perception of the average player.
The counter-example would be Consumer Reports. The problem is people would have to be willing to pay for a subscription to read "unbiased" video game reviews. It might be hard to get people to pay for them when there are so many free ones and "meta" review sites like Game Rankings and Metacritic.
One thing that worries me about the comments on the Halo 3 beta are that the game modes "Land Grab" and "Territories" "Gets intense when time's running out, and only one or two markers are left to defend/capture." or "Very fun when it's tied 2-2 and everyone's going after the last one!"
What about the rest of those matches? Why can't it be "very fun" and "intense" the whole match? I'm thinking of games like Wolfenstein: Enemy Territory where all of the action is always focused, so it's almost always "very fun" and "intense".
I guess I'm just biased towards focused objective-play vs. splitting everyone one up around the map until the very end. Especially when there are only 4 players per team, you really need them all focused in one place.
I mean, it almost makes sense on a 64-player Battlefield server to split the players up among 8 control points, but you still have 4v4 action at each point. If you have multiple control points but only 8 players, things will of course be sparse until the end.
More complex modeling than means and standard deviations have gone into improving online games (unless the author was simplifying things by saying means and standard deviations).
For example both Trueskill and etpub use a Bayesian form of Arpad Elo's rating system to rank and match players.
I did some work modeling kills and wins in Enemy Territory that yielded interesting insights about map- and weapon- balance in that game.
At arena.net, there is at least one employee whose sole job is to model the association between skills winning probabilities in order to balance the skills across the classes.
I'd like to see an RPG that rated players and mobs based on statistical models and made sure the game always gave a near constant level of difficulty. RPGs present, I think, more difficult problem in this area than some of your standard first-person-shooters, etc., because of the variety you have in builds.
I'm a statistician / machine learning guy, so I am all for modeling aspects of games to improve them.
I think it would be even better to have the single-player "wander" into 2-team multiplayer competitions. The single-player may have a completely different goal than the 2-teams, but will feel like he or she is crossing through a more realistic conflict.
I'll use Enemy Territory as an example since I've played that more than most multiplayer FPS.
The single player has a mission to just "escape" by going from one end of the map to another. Or something simple like that. He has to start on, say, the Axis side of the map, maybe not even that far from the current Axis spawn, and make his or her way across to some point on the Allies side of the map.
I'm thinking of a map, like, tc_Base. The Allies are trying to Dynamite the Axis radar, meanwhile, Joe SinglePlay is trying to escape from the same compound. He will have to dodge through Axis soldiers who perhaps have some incentive to shoot the player.
You could just use current multiplayer servers, and drop the occasional single-player into the action as if the player "wandered in"
I think this would give the multiplayer players "something to do" besides sitting around and talking about the latest Stormtrooper rifle until the Jedi force pushes them down the shaft (Jedi Knight II).
If you look at their technology page, you'll notice they are developing an editor. There are pictures of this editor in action.
http://projectoffset.com/technology.html
From that page:
I think the reason they're focusing on this guy is because he is or is one of Final Fantasy's original creators who left Square around the time it merged with Enix or when they lost all that money on FF: Spirits Within.
I think it's significant that Microsoft is making inroads into one of the largest RPG franchises (not necessarily FF, but it's creators) we've seen. Both with Sakaguchi and with the annoucement about FF XI.
Square switching to Sony from Nintendo seemed to be one of the major factors in the Playstation's success over the Nintendo 64 (Final Fantasy VII), and while this may not make the Xbox 360 come out the victor in the next bout of "console wars" it will, I belive, be an influential factor, and therefore any news about what's going in these aspects is interesting.
That's all.
We run an Enemy Territory Server called Beginner's Park 3. We have a website at http://bpark3.com/. We use a slightly customized version of the systats package to track our stats. You can even chat "!stats" in-game on our servers (see links to our servers from the webpage) and get a summary of your stats including your ranking, accuracy, favorite weapon, etc.
systats is a fairly active project and has a really nice stat tracking system but it is very person centric. IMO, it's great for public servers but for clan matches games I think I'd prefer a program called StatWhore which is designed to summarize stopwatch games.
For our stats system, to try and deal with "stat-whoring"---where, for example, players allow themselves to be killed over and over by another to boost the latter's stats---we developed a weak secondary dominance statistic that basically awards more points to those who kill players with higher K / D ratios. This has helped to encourage the better or more stat-conscientious players to play against each other instead of ganging up on the weaker players.
So I recommend systats for public-type servers, it's very flexible, and I recommend StatWhore for matches. As for whether or not stats are important, I think they're fun as long as you don't find yourself worrying everytime you play about how it'll affect your stats. I think a good stats system should try and prevent that. We've tried to do this, but are still considering improvements.
--Zaedyn (aka auachapan)