|
On December 22 2009 23:36 TeWy wrote:
You're obviously missing the point, re-read what I said and put more thoughts into it.
I'm assuming we're talking about this:
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.
StarCraft is a game of imperfect information, with a fair amount of effort being devoted to seeing what your opponent is doing and in some cases hiding what you're doing. Because of this, your anticipations don't (Or shouldn't) change several times a second. Maximum useful APM is not infinite, although that ceiling is so high it's easy to imagine that it is infinite.
For example, If I'm playing as a Zerg against a turtling Terran, have overlords at all unoccupied expansions and key locations on the map, all my workers are working, all my larvae are morphing, I'm researching all the upgrades I want, and I have a lurker force moving out to contain (Say, on HBR, and I want to burrow at the spokes to delay his push), what else can I do? For the time being, my maximum usuable APM has peaked, and until some more units hatch, some more larvae come out, or my Terran opponent does something, my anticipations won't change.
|
Regarding maximum APM, if we call p(t) the number of actions performed at time t, we get the sum P(t) as our total actions performed (using the same notation as in the prev. post). Our average APtick would thus be P(t)/(T-t). Now, if you consider the fact that your input is sequential, it is reasonable to think that our max number of actions pr. tick is 1. Thus p(t) would be 1 if an action is performed at tick t, and 0 otherwise. Thus P(t) is bounded by the number of ticks in the game T-t=T if t=0. The number of minuteticks passed is m, so our max APM would be P(t)/m, with which it's just "plug and chug". Further specialization should be done by the time it actually takes to do things like form a box, etc.
|
We're talking about "math" and "statisics here", if some was checking and microing each of his unit individually each 0.1 rather than 0.01s, as a spectator you would see little difference for sure, however in terms of probability, it would be better to check them each 0.01sec and change their direction/action at this time, by a 0.00001% margin probably but it would still be an improvement.
Anyway, the "time" (and also "space") in a video game is not "continuous" as previously mentionned, therefore the APM can not be "infinite" stricto sensu, not to mention that our real limitation is our corpse so this discussion is kind of futile.
|
On December 23 2009 00:16 TeWy wrote:
We're talking about "math" and "statisics here", if some was checking and microing each of his unit individually each 0.1 rather than 0.01s, as a spectator you would see little difference for sure, however in terms of probability, it would be better to check them each 0.01sec and change their direction/action at this time, by a 0.00001% margin probably but it would still be an improvement.
When you're talking about margins that small, they don't have much of an outcome on the game. I would be more worried about critters blocking your scouting path or something equally absurd than be worried trying to squeak out the 0.00001% improvement in my APM. It's simply not useful, and thus would go above and beyond the maximum -useful- APM. The theory stands.
On December 23 2009 00:16 TeWy wrote:Anyway, the "time" (and also "space") in a video game is not "continuous" as previously mentionned, therefore the APM can not be "infinite" stricto sensu, not to mention that our real limitation is our corpse so this discussion is kind of futile.
I'm not the one who brought it up.
|
this is kind of OT, but I just need to say that I hate that cl sometimes looks like a small D
|
On December 23 2009 00:16 TeWy wrote:
We're talking about "math" and "statisics here", if some was checking and microing each of his unit individually each 0.1 rather than 0.01s, as a spectator you would see little difference for sure, however in terms of probability, it would be better to check them each 0.01sec and change their direction/action at this time, by a 0.00001% margin probably but it would still be an improvement.
Anyway, the "time" (and also "space") in a video game is not "continuous" as previously mentionned, therefore the APM can not be "infinite" stricto sensu, not to mention that our real limitation is our corpse so this discussion is kind of futile.
By the previous assumption that you can only perform one action pr. tick, P(t) is still bounded by T. I think there are something like 8 distinct directions that anything can actually move in StarCraft, and in general your choices from t to t+1 are very limited, so margins like 0.000001% do not exist, but I get your point.
EDIT: I just think the discreet case is more interesting, because with that you can shave off actions quite generously. We now know that P(t) is bounded by T for instance.
|
I really advise reading the rest of the interview. It's better than most would expect.
|
On December 22 2009 22:22 theSAiNT wrote:
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.
Why are those "extreme displays of micro" of any interest to you as a spectator, when you can execute them all easily yourself in every match you play, like EatThePath demands it when he speaks of "a world of perfect micro"?
On December 22 2009 22:05 EatThePath wrote:
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'd rather think SC would then be solved far earlier than chess. In a world of perfect micro, a certain fight is either always won or never, in any given situation. And then I think there would easily come a timing push that is unstoppable. And when you fix that one, there will immediately come another one. In SC as it is, the strength of timing pushes is ultimately balanced through the the micro mechanics needed to succeed with it.
Well - because SC is a game of incomplete information - the only other aspect would be luck and mindgames: Will my opponent do a build this push wins against or one where it fails against? Like rock/paper/scissors. No tension for the spectator.
But I can see what kind of gameplay you might actually have in mind, but then again I think it will still bear mechanical limitations. You speak of "an exhausting set of interactions based on an ongoing fight of harassment" and I imagine the guy who can multitask more harassments and micro them more effectively to win. You could control the game without an interface right through your mind and still be outpaced - but it gets less noticeable and hence less impressive. And at this point, in my opinion, something near the current state is more desirable, where you can be impressed by a variety of things: Amazing micro feats, huge armies macroed out of nowhere, cute mindgames, etc.
ps: Yeah, that's basically two opposing points I made here, depending on how you understand "perfect micro".
On December 23 2009 00:41 InFiNitY[pG] wrote:
this is kind of OT, but I just need to say that I hate that cl sometimes looks like a small D ^_^
hi Infinity
|
On December 23 2009 01:31 a11 wrote:
Why are those "extreme displays of micro" of any interest to you as a spectator, when you can execute them all easily yourself in every match you play
By definition, those are the displays of micro which CANNOT be easily executed in every match you play.
They are often the clutch points of battles and draw the greatest admiration from spectators.
Perfect split of 20 scourges in ZvZ?
10 carrier lockdowns?
Hero dragoon holding off a bunker rush?
If SC2 games have more situations like that because the interface doesn't artificially hinder players, then all the better.
|
Well, you seem to have intentionally cut off the part of the sentence that even made it an argument: Since you were quoting me on a reply to EatThePath, I assumed you were with him. Now you seem to disagree with yourself, saying (indirectly) AI should be enhanced, while you list a "perfect split of 20 scourges" as a micro feat - but the scourges will automatically split with enhanced AI. That's what the whole argument is about. The whole reason, Boxer's lockdowns are so impressive, is that they are not easy to pull of. In Dune, moving 24 units simultaneously is impressive - in SC, no one cares.
|
I'd rather think SC would then be solved far earlier than chess. In a world of perfect micro, a certain fight is either always won or never, in any given situation. And then I think there would easily come a timing push that is unstoppable. And when you fix that one, there will immediately come another one. In SC as it is, the strength of timing pushes is ultimately balanced through the the micro mechanics needed to succeed with it.
Well - because SC is a game of incomplete information - the only other aspect would be luck and mindgames: Will my opponent do a build this push wins against or one where it fails against? Like rock/paper/scissors. No tension for the spectator.
Basically perfect execution would turn the game into a turn-based game in a theoretic sense. But I dont know how you could prove that a turn-based game with hidden info would be easier to solve than one with full information. Chess is solvable whereas hidden information game turns into a game of odds when the perfect strategies are figured out (but thats way farther down the road than just the execution which just brought us into this turnbased-like state). And a game of odds can still be interesting, its not necessarily just rps. And if you figure all the valid things to do on some map - switch the map.
But there wouldnt come some unstoppable build, or if there is its due to racial/map imbalance (which is definitely there though). Defending is always easier than attacking if you know what to prepare for.
|
On December 23 2009 02:38 a11 wrote:
Well, you seem to have intentionally cut off the part of the sentence that even made it an argument: Since you were quoting me on a reply to EatThePath, I assumed you were with him. Now you seem to disagree with yourself, saying (indirectly) AI should be enhanced, while you list a "perfect split of 20 scourges" as a micro feat - but the scourges will automatically split with enhanced AI. That's what the whole argument is about. The whole reason, Boxer's lockdowns are so impressive, is that they are not easy to pull of. In Dune, moving 24 units simultaneously is impressive - in SC, no one cares.
To be fair SC1 has a lot of "look how dumb those scourge are 10 dropships just flew right past them!" moments, and dragoon AI is especially bad. But how will, say, lockdown, or making concaves, or using pylons/overlords as decoy targets, or any of the other micro we love in SC1 be easier?
|
ITT: People who don't know anything about Game Design argue against a proven professional.
Also ITT: People de-rail the topic to a crude representation of APM via statistics and math.
Every click should count. Every click should put one player one step closer to defeating the enemy. All what Sirlin argues against is that repetitive tasks should be replaced because there is no though involved. Skill, sure. Many thoughtless things require skill. Sirlin is simply an advocate that a game should be a pure extension of a player's reasoning.
Ultimately, this is the truth. I am no expert on SC pro's but from what I understand Boxer had fairly low APM and Boxer was the best.
|
On December 23 2009 07:14 Jazriel wrote:
ITT: People who don't know anything about Game Design argue against a proven professional.
Also ITT: People de-rail the topic to a crude representation of APM via statistics and math.
Every click should count. Every click should put one player one step closer to defeating the enemy. All what Sirlin argues against is that repetitive tasks should be replaced because there is no though involved. Skill, sure. Many thoughtless things require skill. Sirlin is simply an advocate that a game should be a pure extension of a player's reasoning.
Ultimately, this is the truth. I am no expert on SC pro's but from what I understand Boxer had fairly low APM and Boxer was the best.
Yes but even Boxer performed a million 'useless' clicks a game.
The problem with the useful vs useless clicks argument is that it is based completely on game theory and not human physiology.
One one hand you have the theory that selecting your probe and clicking 170 times while it travels in a straight line to the locaion you will create a pylon is useless clicks. You only needed one click to travel that straight line. The other side of the argument is that when shit goes down and you need your high apm to be fluid, controlled and precise. Going from 25 click a second performing 'useful' actions to 250 when your attacking and being counterattacked all at once would not be fluid, your internal timing would not be quick and steady, and the mussels in your arms and hands would be under more stress comparitivley.
The reason people make so many 'useless' clicks is that they are getting to a point where they are performing a steady even number of clicks per second and allowing their brain and body to attain a state of focus at that level, as well as a form of warm up for their mussles.
Unless you take non-game mechanics into account the 'useless clicks' argument is flawed.
|
On December 23 2009 03:11 jalstar wrote:Show nested quote +On December 23 2009 02:38 a11 wrote:
Well, you seem to have intentionally cut off the part of the sentence that even made it an argument: Since you were quoting me on a reply to EatThePath, I assumed you were with him. Now you seem to disagree with yourself, saying (indirectly) AI should be enhanced, while you list a "perfect split of 20 scourges" as a micro feat - but the scourges will automatically split with enhanced AI. That's what the whole argument is about. The whole reason, Boxer's lockdowns are so impressive, is that they are not easy to pull of. In Dune, moving 24 units simultaneously is impressive - in SC, no one cares. To be fair SC1 has a lot of "look how dumb those scourge are 10 dropships just flew right past them!" moments, and dragoon AI is especially bad. But how will, say, lockdown, or making concaves, or using pylons/overlords as decoy targets, or any of the other micro we love in SC1 be easier?
Well smart-cast will make lockdowns easier at least. Select your group of 10 ghosts and L click L click L click.
|
On December 23 2009 08:38 seppolevne wrote:Show nested quote +On December 23 2009 03:11 jalstar wrote:On December 23 2009 02:38 a11 wrote:
Well, you seem to have intentionally cut off the part of the sentence that even made it an argument: Since you were quoting me on a reply to EatThePath, I assumed you were with him. Now you seem to disagree with yourself, saying (indirectly) AI should be enhanced, while you list a "perfect split of 20 scourges" as a micro feat - but the scourges will automatically split with enhanced AI. That's what the whole argument is about. The whole reason, Boxer's lockdowns are so impressive, is that they are not easy to pull of. In Dune, moving 24 units simultaneously is impressive - in SC, no one cares. To be fair SC1 has a lot of "look how dumb those scourge are 10 dropships just flew right past them!" moments, and dragoon AI is especially bad. But how will, say, lockdown, or making concaves, or using pylons/overlords as decoy targets, or any of the other micro we love in SC1 be easier? Well smart-cast will make lockdowns easier at least. Select your group of 10 ghosts and L click L click L click.
I don't have a problem with smartcast if it's just that. I'd only be against it if it was the AI casting something other than a buff (like Bloodlust, not D-Matrix) or a heal automatically.
|
On December 23 2009 07:14 Jazriel wrote:
ITT: People who don't know anything about Game Design argue against a proven professional.
It's not so much that I say he's wrong trying all that macro mechanic stuff, but that given the arguments presented to refute it, it is probably not Blizzards time to rest just yet. We'll see in the Beta. Also, if you read the quote Archerofaiur restlessly quotes, you will see that it represents a rather shoddy argument, if an argument at all. However, I've come to the conclusion that IdrA's advice is probably learnt the hard way. A conclusion reinforced by aaiur's belief that "I don't know how to refute that" is equivalent to "addressing the argument".
On December 23 2009 07:14 Jazriel wrote:
Also ITT: People de-rail the topic to a crude representation of APM via statistics and math.
When someone says the max APM in a game is infinite, my discreet mathematics senses start to tingle. *shrug*
EDIT: I might add that ultimately, it seems to me that an esport friendly game should be the ultimate manifestation of the players reasoning OUTSIDE the game. In game, it is simply execution and adaption based on things the player is already comfortable with.
|
On December 23 2009 07:14 Jazriel wrote:
Sirlin is simply an advocate that a game should be a pure extension of a player's reasoning.
Ultimately, this is the truth.
Why, because you say so?
|
On December 23 2009 07:14 Jazriel wrote:
ITT: People who don't know anything about Game Design argue against a proven professional.
Also ITT: People de-rail the topic to a crude representation of APM via statistics and math.
Every click should count. Every click should put one player one step closer to defeating the enemy. All what Sirlin argues against is that repetitive tasks should be replaced because there is no though involved. Skill, sure. Many thoughtless things require skill. Sirlin is simply an advocate that a game should be a pure extension of a player's reasoning.
Ultimately, this is the truth. I am no expert on SC pro's but from what I understand Boxer had fairly low APM and Boxer was the best.
As if Sirlin being a professional game designer makes his opinions immune to fault. Don't use ethos to defend a logically flawed argument.
And arguing against a proven professional? Starcraft is far more popular, far more profitable, and supports are larger pool of skilled player than any of the SF series and especially Sirlin's little pet game. Yet, at the end of every paragraph he writes on starcraft (after he repeatedly proves his ignorance regarding it), he tacks on a nice little ad-hominen against professional starcraft players. How is this not taking a stance against far more 'proven' gaming individuals?
Ultimately, this is the truth.
The arrogance associated with the concept that games 'should' be 100% strategy is astounding and unsettling. So it's what Sirlin likes. So what? People that enjoy starcraft like to be tested in terms of mechanical capability as well. We get that Sirlin doesn't like that. I don't personally care why that is, but that he feels the need to bust into a game with so many gaps in his understanding and declare that the game would be 'better' if we took out the benefit of APM just makes him look ignorant.
|
On December 23 2009 08:54 armed_ wrote:Show nested quote +On December 23 2009 07:14 Jazriel wrote:
Sirlin is simply an advocate that a game should be a pure extension of a player's reasoning.
Ultimately, this is the truth. Why, because you say so?
On December 23 2009 07:14 Jazriel wrote:
Sirlin is simply an advocate that a game should be a pure extension of a player's reasoning.
Ultimately, this is the truth.
On December 23 2009 07:14 Jazriel wrote:
Ultimately, this is the truth.
On December 23 2009 07:14 Jazriel wrote:
this is the truth.
On December 23 2009 07:14 Jazriel wrote:
the truth.
On December 23 2009 07:14 Jazriel wrote:
truth.
Did I do it right?
Truth has nothing to do with what I say and the Truth is objective. Statistically, you and 99% of the people on these forums will disagree with that axiom.
|
|
|
|
|
|
EPT
