|
On December 22 2009 15:32 TeWy wrote:
By the way, the notion itself of "maximum useful APM" is a non-sense, every move you make in a RTS is based on a certain anticipation thus probability, and those probabilities change each +epsilon time, the maximum useful APM is INFINITE, no matter what the RTS is.
Not that this really matters but the value wouldn't be infinite. By the loosest definition of "maximum useful APM" that I can imagine. The number of actions that would actually do something in the game i.e. not selecting a group 1000 times every frame or issuing the exact same command multiple times each frame or issuing different commands that cancel each other before they are executed. Then a simple version of the value would be (average number of possible action in each frame) * (number of frames in that particular game) / game length in seconds. By frame I mean each cycle where starcraft takes new input from the user.
Of course what we should mean by "maximum useful APM" would be the subset of those actions that are actually helping us win the game which is just a small fraction. Even this is a huge value that no player would ever get close to. You can view the useful APM as a logarithmic function where increasing the input early on has great affect on the output but later on it has very little effect. Having 100 APM instead of 50 is a huge difference but having 950 instead of 900 probably makes very little difference. What I meant and what I think Sirlin meant wasn't that exact APM. The value I'm interested in is the point where the curve starts straightening out, where an increase in useful APM stops mattering that much. Not that that point could be defined, but I don't want anyone to ever get close to it and I think that Sirlin believes people have already reached it.
|
I think arguing for a really low skill (mechanics) ceiling in Sc2 is like saying soccer was a better game if all players had exactly the same body, because it'd be less about dexterity and more about strategy then. But, frankly, soccer - unlike Go or Chess - doesn't possess enough strategic depth to be still interesting when all players got the same mechanics, and neither does StarCraft.
And I think this has got inherently to do with the basic principles of StarCraft and is not a flaw. If you take a look at Chess or Go, the prime examples for purely strategic games, there is one important difference regarding how you deal damage: In Chess, you can beat a bishop with a bishop without getting a scratch on yours. In StarCraft, if you send a marine against a marine, you end up with a 5hp marine. There is no way around this.
While in Chess an advantage is achieved through careful strategic planning - because otherwise you will lose your own bishop, too - it is in StarCraft achieved through having more units or through micro that ain't easy to pull off (e.g. stacked mutalisks vs. m&m's) in most cases - which are exactly the most apm-heavy goals to accomplish. Of course there are also aspects like unit placement (flanking) and army composition involved, which one would call rather strategic choices. But to hope that there would even be the possibility to create an RTS that resembles SC, but in which a player's skill is defined solely through the latter and that is still enjoyable and as interesting as chess, imho is blue-eyed at best.
+ Show Spoiler [some examplifying war3 bashing] +Somebody mentioned War3 and I think it might serve as a very fitting example for what I'm trying to say. There were very good players, and there still are. Though Moon was said to have quite impressive APM, I think the whole game could be seen as laid out way more strategically than SC - low mechanic requirements (fewer things to do and more time to pull it off), but many, many choices: On paper, the sheer amount of items, mercenaries and heroes should make for a very "deep" game, should'nt it? I certainly think so. But it's too simple. War3 isn't chess. Everything's been figured out. Though I'm certain, dominant players still constantly outsmart their opponents in subtle mind-games that allow them to always fight in the favorable position etc., there is no real beauty in that.
In the end, it's units hitting each other. There is thinking involved as to which of your own units should best hit which of our enemies units at a given time, but, unlike real strategic games, it doesn't go any further, at least not in a noticeable way. And that's why, when you don't have to divert your actions, because you can do everything easily, things get stale in a game like this, in a game, where there is no hitting without being hit.
So Browder got it right, while Sirlin didn't.
Of course, the best way to add more clicks, without making the game look like a forced clickfest, is, like everyone here already knows, the addition of recurring choices (like in the macro mechanics) rather than the addition of recurring tasks (like sending your workers to mine).
(This was mainly to comment on + Show Spoiler +If Starcraft were, at its core, a terrible game then it would need all these execution barriers separate good from bad players. Maybe it's not a terrible game though. Maybe there is actual strategy in it and letting everyone be on about the same level of APM (by making the maximum useful APM fairly low) would result in a fine game that's inclusive rather than exclusive.
It's ironic that hardcore starcraft players take the stance that their own game is terrible and that only by adding a layer of intentionally bad interface does it become good.
-Sirlin
|
The ideal to which competitive computer games should aspire is strategic depth unmitigated by interface limitations. This means I should be able to perform anything I can think of within the rules. Wanting the goal to be something besides perfect control is just a backwards attitude.
If you want a game to also be a sport, by which I mean that part of the skill is athletic performance beyond what your mind alone can do, why would you advocate something arbitrarily bad like a unit selection cap, which makes more busy work to do what you want within the game? If you could choose any bodily feat, why choose clicking a mouse? That is not inherently matched to strategic thinking, it just happens to be the best-choice optimization of technology ten years ago.
Sirlin may not know very much about Starcraft or appreciate it, but his point is valid in a larger sense than you guys take it, even if it rubs you the wrong way.
|
On December 21 2009 15:38 maybenexttime wrote:
As usual, he doesn't even bother to try to understand what we're actually arguing for. I had little respect for him to start with, but now I have even less.
Somebody please explain what the other side is arguing for.
Could somebody link me to a thread from so 2007 where this might be elucidated?
|
On December 22 2009 18:13 EatThePath wrote:
The ideal to which competitive computer games should aspire is strategic depth unmitigated by interface limitations.
no. seriously. that is what you want. there is no reason it has to be that way.
i don't think anybody disagrees that fps' are a mix of mechanical skill and tactics/strategy. i'm pretty sure nobody thinks it should be any other way. or maybe you're advocating built-in aimbots?
why should starcraft be any different? oh, because it has the word strategy in the genre and randoms who don't even play the game are convinced they would if only they could 'out think' their opponents.
there is no way starcraft is deep enough to survive without mechanics, it is not nearly deep enough and no other 'rts' is or will be, at the very least while the current stock rts template is followed. starcraft is, by a ridiculous margin, the most successful 'rts' ever as an esport. to suggest it's doing it wrong and a sirlin and a bunch of casuals hold the key to taking sc2 to the next level is frankly laughable. protip: there have been loads of mechanically easy rts' since the release of starcraft. where are they now? are you going to argue its the strategic depth of starcraft that allowed it to succeed where they all failed?
ps. the idea that mechanics are in the way of you executing your brilliant strategy is a horribly one-sided way of looking at it.
pps. people aren't really talking about a unit selection cap, apart from in the context of sirlin bringing it up disingenuously to discredit people who believe mechanics are a vital part of the game. educate yourself. it is pretty insulting to enter a thread, tell everybody how it should be, then ask what they're talking about. you could try searching for "MBS" to start.
|
As long as the "maximum useful APM" is unreachable do any of the little details matter?
|
On December 22 2009 18:40 Lachrymose wrote:
there is no way starcraft is deep enough to survive without mechanics
Of course the wow factor of pro gamers' superlative mechanical feats helps a game be worth televising. I'm not sure how you're evaluating how deep the game is though. Lots of other things are factors in Starcraft's success, anyway, like the evocative universe and feel of it.
ps. the idea that mechanics are in the way of you executing your brilliant strategy is a horribly one-sided way of looking at it.
Why? That's why I wrote the paragraph about sport. Why did you just rant instead of respond?
pps. people aren't really talking about a unit selection cap, apart from in the context of sirlin bringing it up disingenuously to discredit people who believe mechanics are a vital part of the game. educate yourself. it is pretty insulting to enter a thread, tell everybody how it should be, then ask what they're talking about. you could try searching for "MBS" to start.
Which is why I asked. Sorting through flamey TL threads is slow going. I appreciate the motivation for the vitriol at Sirlin when you put it like that, so thanks. I don't mind being insulting on the internet if it gets people to change their minds when they read the truth™.
edit: This is all in the context of design principles and philosophies, not the way Starcraft 2 should turn out actually really. I view that as passé and out of my hands anyway, right? Besides, Starcraft "isn't that game", I agree. My point is to defend Sirlin's larger correct point. But now I see he's being unnecessarily snobby. He isn't saying you should dumb things down to a grandma APM status. You should make the controls as good as possible for the game you have so that busy work APM is minimized. That's all.
|
|
|
without mbs and automine blizzard can't bring in new gamers, and the old sc1 faithful won't be around forever
|
On December 22 2009 19:30 EatThePath wrote:Show nested quote +On December 22 2009 18:40 Lachrymose wrote:
there is no way starcraft is deep enough to survive without mechanics
Of course the wow factor of pro gamers' superlative mechanical feats helps a game be worth televising. I'm not sure how you're evaluating how deep the game is though. Lots of other things are factors in Starcraft's success, anyway, like the evocative universe and feel of it..
I don't think there's actually a large audience gathered because of "the evocative universe and feel of it". Those might be benefiting aspects, but I doubt anyone would come to cling to a game like SC because of these points. That's hardcore RPGs we're speaking of, where the flair takes precedence over gameplay.
Point being, the only reason watching people play chess is impressive, is how they sometimes think a dozen turns in advance and pull of astonishing plays like sacrificing both rooks, a bishop and a queen and then winning with the three remaining minor pieces. In SC, you sacrifice one of your expansions for a doom drop into your opponent's mainbase and that's it. It doesn't get any deeper.
Hence, if everybody was able to pull off everything he could think of, pro-SC'd be just plain boring for the most part. Unfixable. You know how many openings there are in chess in comparison to SC? I'm afraid I can't pinpoint quite why this is, yet, but I don't see a way for SC to become some kind of real-time chess, where mechanics would be equally non-existant while the gameplay would be equally deep.
If you want a game to also be a sport, by which I mean that part of the skill is athletic performance beyond what your mind alone can do, why would you advocate something arbitrarily bad like a unit selection cap, which makes more busy work to do what you want within the game?
Well, I already addressed this point, saying ideally there should always be a choice involved, so there will be a reason why you don't want the AI to do it.
|
Even with the reduced difficulty mechanics you will not see anyone playing perfectly. The skill ceiling is still beyond any human attainability.
|
On December 22 2009 18:13 EatThePath wrote:
The ideal to which competitive computer games should aspire is strategic depth unmitigated by interface limitations. This means I should be able to perform anything I can think of within the rules. Wanting the goal to be something besides perfect control is just a backwards attitude.
This almost looks cut and pasted from Sirlin's book. Why should this be the ideal to which competitive computer games, or any games at all, aspire? This is 100% Sirlin's opinion of how games should be. It is the unprovable linchpin that he bases all of his game theory on. Many people, possibly the majority of people that play computer games and certainly the majority of people that enjoy starcraft also value the ability to execute strategy mechanically. The goal is still perfect control, the difference is that starcraft doesn't give it to you as a baseline - you attempt to achieve it personally.
Honestly, I try not to let my biases of Sirlin affect my opinion of what he writes, but it really just irks me that he says things like this with the background of being the 'move spammer' street fighter player. He comments again and again in his blogs and book about how he isn't as mechanically capable as other professional SF players, therefore valuing strategy over execution. This just colors all of his literature for me. It's an oversimplification, but it feels like he always wants to eliminate the physical aspects of gameplay just because he isn't as adept at them.
|
It is not actually the skill ceiling that has been lowered. It is more accurate to say that the skill floor level has been raised. Lower skilled players will be better. But higher skilled players will still require and use as much skill as SC1 takes, just in different areas of the game, and not all of them micro orientated.
The gap between a noob and a pro is not smaller. It's just that the noob is actually able to play the game.
|
Chess is a useful analogue, say if you were trying to explain to a complete outsider about the general dynamics of a game of starcraft. But it's so much different than the RTS genre in many important ways. I could go on and on so I think further discussion re: chess/sc should be moved elsewhere, but briefly...
The biggest two things from a game mechanisms (as in design), NOT game-playing perspective:
-Starcraft is a game of imperfect information, which comes out more in the moment to moment tactics than the strategy, but chess has perfect gamestate information.
-Chess is solvable. Starcraft is not because it isn't quantized, in time, space, and more generally number of possible gamestates.
I find starcraft to be richer than chess in essence, which is a natural conclusion from how I'm looking at it because starcraft has more "moving parts". If you were able to control your units perfectly, starcraft would have an exhausting set of interactions based on an ongoing fight of harassment and unit positioning, while trying to see what the opponent was doing and nullifying their threats.
There would be more perceived openings for starcraft if any given unit could be considered a game piece, as opposed to armies of certain compositions and sizes. This would make sense in a world of perfect micro, where each unit gives you a stream of choices.
I want interfaces and AIs that let you control your game pieces with the least busy work, not make decisions for you. For instance, why can't starcraft have troop formation commands? It's just dumb that I have to line up dragoons manually. That doesn't reduce my choices, it expands what I can feasibly do. I see MBS and UUS in the same way.
I realized how I can phrase this, and I think it really clarifies a lot of these arguments. When I play a game, the fun and competition comes from interacting with the opponent. I don't see "macro" in starcraft as part of interacting with the opponent. However, army composition is part of interaction, and it is facilitated by macro; I want the macro shortcutted so I can just do army composition with the least fuss, so I can interact, not play real-time solitaire.
edit: @TPS - Well I was careful to phrase what you quoted in the way that I did. Interfacing will always be a part of computer games, and so a part of the skill will be masterful interfacing with the game. I think it's absurd to intentionally include interface tasks when they naturally come up anyway, and will differentiate players no matter. I can see excepting this tenet if the interface tasks are particularly artful or cohesive with the vibe of the game, which quickly becomes subjective. :\ I guess I approach things from a more abstract direction, like a lot of designers will. I totally hear you about Sirlin. Defending the idea not the man.
|
On December 22 2009 17:01 a11 wrote:
But, frankly, soccer - unlike Go or Chess - doesn't possess enough strategic depth to be still interesting when all players got the same mechanics, and neither does StarCraft.
And that's why, when you don't have to divert your actions, because you can do everything easily, things get stale in a game like this
This depresses me.
I'm with Sirlin. I would like to think that SC does have and SC2 will have sufficient strategic depth to not have to have pointless actions.
Also, if anything the 'extra clicks' that professionals have over the rest of us won't disappear. They will be diverted towards more useful things. If anything, I look forward to seeing even more extreme displays of micro.
SC2 Hwaiting!
|
On December 22 2009 21:33 DeCoup wrote:
It is not actually the skill ceiling that has been lowered. It is more accurate to say that the skill floor level has been raised. Lower skilled players will be better. But higher skilled players will still require and use as much skill as SC1 takes, just in different areas of the game, and not all of them micro orientated.
This.
As much as I dislike Sirlin's blog, he is right imo that the cap on unit selection is just a gimicky, convoluted UI addition to force more clicks and that it is better game design, if having several small control groups results from players' choices, since it pays off in game (which up to a certain extent is the case in SC anyway). The case for MBS and automine is less clear however and that's probably why the discussion was/is so fierce about it. The want for mechanics is definitely legit in any sport/game, so the only decisive questions imo are: Does the "new" interface reduce the slope towards mechanical perfection too much or not? And if it is the case, are there no better ways to increase the slope again than an outdated interface? I personally think that Blizzard is on the right track with the macro mechanics, even though they still need work.
What I totally dislike about Sirlin's comment, is the lack of respect it shows with regard to Starcraft's hardcore player base. He wants to turn every game into a contest of wits, which - even though it is a central part of Starcraft - is and should not be what the game is all about.
Btw. what happened to the interesting grammar discussion from page 4? Maybe an own thread for the merits of semicolons would be in place?
|
On December 22 2009 22:05 EatThePath wrote:
Chess is a useful analogue, say if you were trying to explain to a complete outsider about the general dynamics of a game of starcraft. But it's so much different than the RTS genre in many important ways. I could go on and on so I think further discussion re: chess/sc should be moved elsewhere, but briefly...
The biggest two things from a game mechanisms (as in design), NOT game-playing perspective:
-Starcraft is a game of imperfect information, which comes out more in the moment to moment tactics than the strategy, but chess has perfect gamestate information.
-Chess is solvable. Starcraft is not because it isn't quantized, in time, space, and more generally number of possible gamestates.
I find starcraft to be richer than chess in essence, which is a natural conclusion from how I'm looking at it because starcraft has more "moving parts". If you were able to control your units perfectly, starcraft would have an exhausting set of interactions based on an ongoing fight of harassment and unit positioning, while trying to see what the opponent was doing and nullifying their threats.
There would be more perceived openings for starcraft if any given unit could be considered a game piece, as opposed to armies of certain compositions and sizes. This would make sense in a world of perfect micro, where each unit gives you a stream of choices.
I want interfaces and AIs that let you control your game pieces with the least busy work, not make decisions for you. For instance, why can't starcraft have troop formation commands? It's just dumb that I have to line up dragoons manually. That doesn't reduce my choices, it expands what I can feasibly do. I see MBS and UUS in the same way.
I realized how I can phrase this, and I think it really clarifies a lot of these arguments. When I play a game, the fun and competition comes from interacting with the opponent. I don't see "macro" in starcraft as part of interacting with the opponent. However, army composition is part of interaction, and it is facilitated by macro; I want the macro shortcutted so I can just do army composition with the least fuss, so I can interact, not play real-time solitaire.
edit: @TPS - Well I was careful to phrase what you quoted in the way that I did. Interfacing will always be a part of computer games, and so a part of the skill will be masterful interfacing with the game. I think it's absurd to intentionally include interface tasks when they naturally come up anyway, and will differentiate players no matter. I can see excepting this tenet if the interface tasks are particularly artful or cohesive with the vibe of the game, which quickly becomes subjective. :\ I guess I approach things from a more abstract direction, like a lot of designers will. I totally hear you about Sirlin. Defending the idea not the man.
Why do you approach these concepts from a design position based on game theory rather than a gameplay position based on experience? If gaming for most of my life has taught me anything, it's that your knowledge of a game and understanding of its mechanics is only as good as the time you invest in working with them. You can watch pro gamers play and criticize their strategic play as an amateur because you don't understand the strain of operating the interface - you haven't done it enough or at the proper level of play. When I was learning to play, I asked a lot of questions that occurred to me naturally with almost zero experience actually playing. Looking back at these, I see that they were generally regarding totally irrelevant aspects of gameplay. Without the experience of playing, I had absolutely no idea what I was talking about, even though to the best I could figure I understood the game. In game theory, you have to work with a series of assumptions. It only bothers me when these assumptions are untempered by actual gameplay experience.
I wouldn't advocate the inclusion of something like troop formation controls in sc2. It shifts the game from being capable of controlling your army effectively to knowing when and how to hit the 'box formation' button. It shifts the game away from mechanical capability, but not necessarily towards strategy. Whether or not you want options like this depends on how you want to be tested by the game. And, it's a weak point but it always occurs to me, it makes seemingly simple feats that much less impressive when the game engine automatically does it for you. When I watch starcraft, I don't want to just be dazzled by the brilliance and strategy of the players, it's important to me to appreciate the mechanical control that they exert over the interface.
As for macro, it is unquestionably an interaction with the opponent. What you build and when you build it depends utterly on what you know about your enemy. That you can't physically see them is simply a byproduct of the imperfect information aspect of the game. What it sounds like you actually want is a system that provides you with more total information. I don't qualify any action that helps you win as 'busy work' - even if you aren't having a direct effect on your opponent by building a unit, you are determining the outcome of the game through your actions. By and large, I don't want these tasks to be left to an automatic interface, but there is a careful balance that makes a game better. For example, I like that starcraft has rally points. Without them, I might enjoy the game a tiny bit less. That doesn't mean that I feel that automining will make the game a tiny bit better, though. If I had to choose, I'd probably keep automining out. The fact that starcraft has such a careful balance of mechanics like these that appeal to me is really miraculous. For the most part, I'd just as soon see most of it stay the same.
|
EDIT: TeWy suggested the best APM was infinite. This is my response.
TeWy, that would be true if the space of a StarCraft game was continuous. That would be a sensible idealization, but it's not strictly true, as StarCraft transitions discreetly from time t to t+epsilon. What this means is that epsilon cannot be arbitrarily small, so we might as well call it 1. Thus, the number of things you can control at any moment is a finite sequence k(t). The maximum number of stuff you can do in an SC match lasting from t to T would thus be the sum from t to T of k(t), which you will find is very finite, and which we will call K(t).
k(t) itself is the apm/tick if "game ticks" are our units. Also, when you consider that you issue orders over longer periods of time, most of your units are already doing something every moment, so it would make sense to divide by some average "activity coefficient" a, which would be the ratio of units already in action. K(t)/a would be our total amount of actions, while k(t)/a would be our apm/tick, which is convertible to apm/minutes, and importantly enough, finite.
|
On December 22 2009 22:57 edahl wrote:
EDIT: TeWy suggested the best APM was infinite. This is my response.
TeWy, that would be true if the space of a StarCraft game was continuous. That would be a sensible idealization, but it's not strictly true, as StarCraft transitions discreetly from time t to t+epsilon. What this means is that epsilon cannot be arbitrarily small, so we might as well call it 1. Thus, the number of things you can control at any moment is a finite sequence k(t). The maximum number of stuff you can do in an SC match lasting from t to T would thus be the sum from t to T of k(t), which you will find is very finite, and which we will call K(t).
k(t) itself is the apm/tick if "game ticks" are our units. Also, when you consider that you issue orders over longer periods of time, most of your units are already doing something every moment, so it would make sense to divide by some average "activity coefficient" a, which would be the ratio of units already in action. K(t)/a would be our total amount of actions, while k(t)/a would be our apm/tick, which is convertible to apm/minutes, and importantly enough, finite.
Pretty convoluted, but you're absolutely right. The notion of infinite useful APM is pretty ridiculous. If all my units have their orders and all my production structures are building, any "clicks" I do at that point are wasted spamming.
|
On December 22 2009 22:57 edahl wrote:
EDIT: TeWy suggested the best APM was infinite. This is my response.
TeWy, that would be true if the space of a StarCraft game was continuous. That would be a sensible idealization, but it's not strictly true, as StarCraft transitions discreetly from time t to t+epsilon. What this means is that epsilon cannot be arbitrarily small, so we might as well call it 1. Thus, the number of things you can control at any moment is a finite sequence k(t). The maximum number of stuff you can do in an SC match lasting from t to T would thus be the sum from t to T of k(t), which you will find is very finite, and which we will call K(t).
k(t) itself is the apm/tick if "game ticks" are our units. Also, when you consider that you issue orders over longer periods of time, most of your units are already doing something every moment, so it would make sense to divide by some average "activity coefficient" a, which would be the ratio of units already in action. K(t)/a would be our total amount of actions, while k(t)/a would be our apm/tick, which is convertible to apm/minutes, and importantly enough, finite.
Starcraft transitions are not absolutely continuous that's true, we're on a computer, we might be limited by the engine etc... I was giving an answer to someone who believed that the "maximum useful apm" was not limited by the law of physics, which is wrong, and that it was around 1000 apm.
On December 22 2009 23:30 ComradeDover wrote:
Pretty convoluted, but you're absolutely right. The notion of infinite useful APM is pretty ridiculous. If all my units have their orders and all my production structures are building, any "clicks" I do at that point are wasted spamming.
You're obviously missing the point, re-read what I said and put more thoughts into it.
|
|
|
|
|
|
EPT
