UC Berkeley Starcraft Class - Page 20

On February 10 2009 18:24 MJAnoname wrote:
Yes, I'm from Sirlin.net. No, I'm not Sirlin. Yeah, I'm kinda late in this comment.

I will make one statement that agrees with Sirlin (In my first two paragraphs), and then one clarification that I think will appease most of your discontents with what he's said (The rest of my post).

I just want to point out that saying that Starcraft should remove some of it's unncessary clicks and keystrokes by streamlining its UI doesn't necessairly mean that Starcraft should become a "STRATEGY (in real-time)" game. It needs to be strategy AND real-time. The challenge of Starcraft isn't just choosing the right play, it's making all the important decisions you must make at the extremely rapid pace you must make them at, while splitting your attention between 10 different places on the map all at once.

I believe that any given string of actions in a game represent one larger decision. Imagine this: Player A moves his screen to see the location where all his gateways are built, selects them one by one, pressing z after selecting 6 of them, and pressing d after selecting the other 4. That is the long way of saying "Player A decided to build 6 zealots and 4 dragoons." The in-game representation of that is "move screen - click, z (times six) - click, d (times four)." That requires a movement of the screen, ten clicks, and ten keystrokes. (There is also an implied movement of the screen at the end, because looking at the gateways while they're building isn't too useful.) We all know that in SC2 that decision will be ten keystrokes -"0zzzzzzddd", which is over 50% more efficient than in SC1. (I know the debates over MBS have raged here for hundreds (thousands?) of pages, so I'll try to tiptoe around the subject and just use it as an example instead of making it the brunt of my argument.)


The problem here is that it's hard to give a good reason why building 6 zealots and 4 dragoons should require ten clicks, ten keystrokes, and two screen movements (one explicit, one implied) when it could be just ten keystrokes. Saying the first is somehow better would open up all sorts of ridiculous logical arguments, like arguments that say that Dune 2 is better because you have to select units individually instead of in groups. That is artifical difficulty. Imagine if the ten keystrokes from SC2 is a plain old hoop that you have to jump through. By comparison, SC1 requires you to jump through a flaming hoop, and Dune 2 just plain requires you to walk on open flames. What does it really add to the game that is interesting? That's not a hypothetical question- I believe there is an real-world answer for it that is present in SC1, and I believe there is an potential answer for it that is better than SC1's system.

What makes building six zealots and four dragoons 50% less efficient than "0zzzzzzdddd" interesting is the fact the SC1 system greatly assaults the player's attention-dividing ability and hand speed, the "real-time" part of the RTS genre. If you reduce the amount of APM it takes a Terran player to do the same stuff he already does by half, you free up about 200 APM for him to use elsewhere. There are all sorts of interesting attack methods in SC1 that are based on the concept that it is hard for a player to pay attention to everything that is going on all at once, and have the speed in both physical movement and decision-making to respond to it all. If you split your forces into 3 groups, two "main armies" and one shuttle with a reaver, you can assault your opponent's attention and speed. If you are attacking two of his expansions while reaver dropping his SCV line, you are attacking him on three fronts, while he still has to deal with his 4th front of macromanagement. He may make a critical mistake, like not noticing your reaver drop until all his SCVs are already dead, and if so you have just split his attention in a strategically interesting manner - you are using attention splitting as a strategy to beat him. But if he has 200 APM just freed up with nothing to put it towards, this strategy would not be nearly as effective. By greatly reducing the strains the UI places on his physical speed and attention reserves, all sorts of interesting strategies in Starcraft would be nullified by the fact that your opponent has such deep reserves of attention and speed. You could (almost?) never overwhelm his ability to pay attention to everything and react to it in a reasonable time frame. So what to do?

I feel the answer is that both communities are right in their own way here, and a combination answer best solves the problem. Sirlin is right that desiging a UI to increase APM is artifical and not an interesting way of keeping the players busy. However, pro Starcraft communities are right that attention-splitting and quick decision-making (and as a by-product speed) are the "real-time" component, and are a huge part of what makes Starcraft the best and most successful competitive RTS in gaming history. I think the answer is to cut out all the "unnecessary" clicks - reduce building four zealots to "0zzzz" and increase efficiency by over 5% - and replace all of that "hollow" APM that has been carved out of the game by increasing the Decisions Per Minute (DPM) that the player has to face. While APM is well and good, DPM is the number that the "real-time" component of an RTS should strive to measure. DPM is what drives APM - your APM (or EAPM, for those who spam commands) is a reflection of the number of decisions you make per minute and your speed/efficiency in executing them - the number of decisions you make a minute is what drives you to perform a number of actions that minute.

For example, the actions required to build 10 zealots is 20 actions - 10 clicks, 10 keystrokes. SC2 will cut the clicks out and reduce that to 10 keystrokes. If that's all you do for one minute, your APM in SC1 is 20, and APM in SC2 is 10. However, your DPM in both games is only 1 - you made the one decision to build 10 zealots, and that's it. (That might be oversimplifying it - let's assume for example's sake that you have nothing building in any of the gateways at the moment, and they're all in one location so you don't have to consider where. If those two aren't given then your DPM might be higher than 1.) At this rate, a 20 DPM player in SC1 would have an APM of 400. (I know that this only applies to macro and not to troop movement; again, it's for examples sake, and I think if we drew out a more fleshed-out real-game example this would certainly still hold true, though the numbers used in the example might be different). A 20 DPM player in SC2 would have an APM of 200. However, the lowered APM in SC2 gives us room to design the game in order to increase the number of decisions the player faces per minute. Since we freed up 200 APM, the equivilent of 20 DPM in SC2 conversion rates, we can add another 20 DPM, and in fact we must add 20 DPM in order maintain the speedy-decision-making component of the "real-time" part of Starcraft, Futhermore, more DPM means more strategic decisions the player must face at each turn, which means there are more opportunities to increase the weight on the "strategy" side of the game. The end result would be that three of the four main attributes that Starcraft measures to determine your success - decision-making-speed, attention-dividing, and strategy - would recieve increases in their overall contribution to the winner of the game, while the hand-speed limitations would remain.


This post has already been long as hell, so I'll spare you another lengthy outro paragraph. To sum it up quickly: "Requiring high APM, and therefore high hand speed, is just an unfortunate side effect of attempting to require high DPM. The goal of any RTS should be to ensure that the APM and hand-speed it requires to sufficiently test the player's attention-splitting abilities should remain as decision-full as possible, or in other words, represent a high DPM. A high DPM increases both the strategic and real-time aspects of the game. The fact that it requires good mechanics is just an unfortunate side effect of the fact that we cannot plug the game directly into our brains."

There's also no way I'm going to thoroughly proofread something this long when I'm not being graded on it or paid for it, so I'm sorry if there's any tyops. Feel free to ask any questions you want, I'll check this thread every now and again. Also feel free to flame away, it doesn't bother me. I anticipate a wave of "WHATS YOUR KESPA RANK? IF IT'S BELOW 20 YOU HAVE NO RIGHT TO TALK" - I am not the best SC player in the world, but I would find it hard to refute this analysis, unless you don't want to grant me the few axioms I have taken here. (Namely, I think they are that Starcraft measures decision-making speed, attention-splitting, and strategy primarily, and hand speed secondarily; that achieving higher necessary DPMs required to play at the competitive level is always a good goal for an RTS game, and that the DPM should be what is filled out to test speed and not the APM; and that increasing your APM is pointless if the increased number of actions still represents the same DPM)

On February 10 2009 18:24 MJAnoname wrote:
Yes, I'm from Sirlin.net. No, I'm not Sirlin. Yeah, I'm kinda late in this comment.

I will make one statement that agrees with Sirlin (In my first two paragraphs), and then one clarification that I think will appease most of your discontents with what he's said (The rest of my post).

I just want to point out that saying that Starcraft should remove some of it's unncessary clicks and keystrokes by streamlining its UI doesn't necessairly mean that Starcraft should become a "STRATEGY (in real-time)" game. It needs to be strategy AND real-time. The challenge of Starcraft isn't just choosing the right play, it's making all the important decisions you must make at the extremely rapid pace you must make them at, while splitting your attention between 10 different places on the map all at once.

I believe that any given string of actions in a game represent one larger decision. Imagine this: Player A moves his screen to see the location where all his gateways are built, selects them one by one, pressing z after selecting 6 of them, and pressing d after selecting the other 4. That is the long way of saying "Player A decided to build 6 zealots and 4 dragoons." The in-game representation of that is "move screen - click, z (times six) - click, d (times four)." That requires a movement of the screen, ten clicks, and ten keystrokes. (There is also an implied movement of the screen at the end, because looking at the gateways while they're building isn't too useful.) We all know that in SC2 that decision will be ten keystrokes -"0zzzzzzddd", which is over 50% more efficient than in SC1. (I know the debates over MBS have raged here for hundreds (thousands?) of pages, so I'll try to tiptoe around the subject and just use it as an example instead of making it the brunt of my argument.)


The problem here is that it's hard to give a good reason why building 6 zealots and 4 dragoons should require ten clicks, ten keystrokes, and two screen movements (one explicit, one implied) when it could be just ten keystrokes. Saying the first is somehow better would open up all sorts of ridiculous logical arguments, like arguments that say that Dune 2 is better because you have to select units individually instead of in groups. That is artifical difficulty. Imagine if the ten keystrokes from SC2 is a plain old hoop that you have to jump through. By comparison, SC1 requires you to jump through a flaming hoop, and Dune 2 just plain requires you to walk on open flames. What does it really add to the game that is interesting? That's not a hypothetical question- I believe there is an real-world answer for it that is present in SC1, and I believe there is an potential answer for it that is better than SC1's system.

What makes building six zealots and four dragoons 50% less efficient than "0zzzzzzdddd" interesting is the fact the SC1 system greatly assaults the player's attention-dividing ability and hand speed, the "real-time" part of the RTS genre. If you reduce the amount of APM it takes a Terran player to do the same stuff he already does by half, you free up about 200 APM for him to use elsewhere. There are all sorts of interesting attack methods in SC1 that are based on the concept that it is hard for a player to pay attention to everything that is going on all at once, and have the speed in both physical movement and decision-making to respond to it all. If you split your forces into 3 groups, two "main armies" and one shuttle with a reaver, you can assault your opponent's attention and speed. If you are attacking two of his expansions while reaver dropping his SCV line, you are attacking him on three fronts, while he still has to deal with his 4th front of macromanagement. He may make a critical mistake, like not noticing your reaver drop until all his SCVs are already dead, and if so you have just split his attention in a strategically interesting manner - you are using attention splitting as a strategy to beat him. But if he has 200 APM just freed up with nothing to put it towards, this strategy would not be nearly as effective. By greatly reducing the strains the UI places on his physical speed and attention reserves, all sorts of interesting strategies in Starcraft would be nullified by the fact that your opponent has such deep reserves of attention and speed. You could (almost?) never overwhelm his ability to pay attention to everything and react to it in a reasonable time frame. So what to do?

I feel the answer is that both communities are right in their own way here, and a combination answer best solves the problem. Sirlin is right that desiging a UI to increase APM is artifical and not an interesting way of keeping the players busy. However, pro Starcraft communities are right that attention-splitting and quick decision-making (and as a by-product speed) are the "real-time" component, and are a huge part of what makes Starcraft the best and most successful competitive RTS in gaming history. I think the answer is to cut out all the "unnecessary" clicks - reduce building four zealots to "0zzzz" and increase efficiency by over 5% - and replace all of that "hollow" APM that has been carved out of the game by increasing the Decisions Per Minute (DPM) that the player has to face. While APM is well and good, DPM is the number that the "real-time" component of an RTS should strive to measure. DPM is what drives APM - your APM (or EAPM, for those who spam commands) is a reflection of the number of decisions you make per minute and your speed/efficiency in executing them - the number of decisions you make a minute is what drives you to perform a number of actions that minute.

For example, the actions required to build 10 zealots is 20 actions - 10 clicks, 10 keystrokes. SC2 will cut the clicks out and reduce that to 10 keystrokes. If that's all you do for one minute, your APM in SC1 is 20, and APM in SC2 is 10. However, your DPM in both games is only 1 - you made the one decision to build 10 zealots, and that's it. (That might be oversimplifying it - let's assume for example's sake that you have nothing building in any of the gateways at the moment, and they're all in one location so you don't have to consider where. If those two aren't given then your DPM might be higher than 1.) At this rate, a 20 DPM player in SC1 would have an APM of 400. (I know that this only applies to macro and not to troop movement; again, it's for examples sake, and I think if we drew out a more fleshed-out real-game example this would certainly still hold true, though the numbers used in the example might be different). A 20 DPM player in SC2 would have an APM of 200. However, the lowered APM in SC2 gives us room to design the game in order to increase the number of decisions the player faces per minute. Since we freed up 200 APM, the equivilent of 20 DPM in SC2 conversion rates, we can add another 20 DPM, and in fact we must add 20 DPM in order maintain the speedy-decision-making component of the "real-time" part of Starcraft, Futhermore, more DPM means more strategic decisions the player must face at each turn, which means there are more opportunities to increase the weight on the "strategy" side of the game. The end result would be that three of the four main attributes that Starcraft measures to determine your success - decision-making-speed, attention-dividing, and strategy - would recieve increases in their overall contribution to the winner of the game, while the hand-speed limitations would remain.


This post has already been long as hell, so I'll spare you another lengthy outro paragraph. To sum it up quickly: "Requiring high APM, and therefore high hand speed, is just an unfortunate side effect of attempting to require high DPM. The goal of any RTS should be to ensure that the APM and hand-speed it requires to sufficiently test the player's attention-splitting abilities should remain as decision-full as possible, or in other words, represent a high DPM. A high DPM increases both the strategic and real-time aspects of the game. The fact that it requires good mechanics is just an unfortunate side effect of the fact that we cannot plug the game directly into our brains."

There's also no way I'm going to thoroughly proofread something this long when I'm not being graded on it or paid for it, so I'm sorry if there's any tyops. Feel free to ask any questions you want, I'll check this thread every now and again. Also feel free to flame away, it doesn't bother me. I anticipate a wave of "WHATS YOUR KESPA RANK? IF IT'S BELOW 20 YOU HAVE NO RIGHT TO TALK" - I am not the best SC player in the world, but I would find it hard to refute this analysis, unless you don't want to grant me the few axioms I have taken here. (Namely, I think they are that Starcraft measures decision-making speed, attention-splitting, and strategy primarily, and hand speed secondarily; that achieving higher necessary DPMs required to play at the competitive level is always a good goal for an RTS game, and that the DPM should be what is filled out to test speed and not the APM; and that increasing your APM is pointless if the increased number of actions still represents the same DPM)

On February 10 2009 23:32 ruXxar wrote:
The problem is this : If we free up the macro side of the game, the balance will undeniable shift.
This is what we do not want, we want the balance between micro and macro to stay roughly the same.
If the balance shifts towards micro the game will become something ala wc3, where the macro decision making does not require enough attention to actually having the player overloaded.
This is what makes SC so great, the capacitiy of attention or "thoughts per minute" needed to play a "perfect" game of starcraft is in some situation more than players currently can handle, and this is where the game comes down to prioritation, shall i focus on micro or shall i focus on macro, aswell as thinking about your strategy at the same time.

A game like wc3 for instance has almost no macro.

However, this is not solely due to the UI interface, because wc3 was limited in other ways(max food supply reduced, less resources availible(gold mine vs mineral field)).

I think what most people are afraid of is to see SC2 turn into something like wc3, when we start deviating from what is the "tried and tested" formula.

The thing is that since it's a successor in a franchise, it should try to keep the elements which peopled liked about its predecessor, so that you can keep the fanbase($$$). At the same time you also want to incorporate new features so that you can attract a new crowd, and expand your market.

If you take away or change those elements, you get some people that adapt, and some that get alianated.

I just hope I won't be one in the latter group.

EDIT : On topic,

Real shame second class didn't get recorded Q.Q

On February 10 2009 19:32 toopham wrote:
Good stuff. you're very good at explaining. I didnt bother reading about the MBS debate or even read up on all the details about SC2 so I'm far behind on this, but your reasoning seems very good.

If blizzard really is putting a lot of decision making factors in SC2 that requires higher apm then I guess MBS is just to let the players use that extra "200 APM" for somewhere else that is needed in SC2. And thus in the end you can still go for reaver drop while attacking the front to use ur opponent's limited APM to ur advantage.

On February 11 2009 22:25 maybenexttime wrote:
Ooookaaaaay, welcome to two years ago.

On February 11 2009 22:25 maybenexttime wrote:You’re also missing the fact that the side effect of SBS & manual-mining is the variety of playstyles and negative feedback mechanic associated with expanding (the more you expand, the more overwhelmed with the macro side of the game you get, rendering you more vulnerable the the player who’s behind economically).

On February 11 2009 22:25 maybenexttime wrote:“If you reduce the amount of APM it takes a Terran player to do the same stuff he already does by half, you free up about 200 APM for him to use elsewhere.”

It’s not really the APM per se – it’s the attention requirements of various tasks (in this case macro-related).

On February 11 2009 22:25 maybenexttime wrote:It’s not like the idea is new – we’ve already established that two years ago... The problems is that SBS & manual-mining have several really positive side effects, and MBS/auto-mining remove those.

On February 11 2009 22:25 maybenexttime wrote:The divide between being either “pro-MBS” or “anti-MBS” (providing you’re familiar with arguments of both sides, which Sirlin looks completely oblivious of, by the way) is determined by how skeptical towards the new macro mechanics’ success they are. The latter are very skeptical and would rather see MBS and/or auto-mining removed (and thus the game keep that ‘ “hollow” APM’) than risk the game (SC2) being less skillful and much more one dimensional (micro-focused), whereas the former believe preserving those positive side effects with new (attention-demanding) macro mechanics (this time ‘round in form of conscious gameplay design choices) is after all possible.

On February 11 2009 22:25 maybenexttime wrote:(There are also those who believe the physical aspect of StarCraft should be the one emphasized on, while the mental side is just an addition. From what I understand, this is what e.g. NatsuTerran wants. Those are a small minority, though.)

On February 11 2009 22:25 maybenexttime wrote:Your example is a gross oversimplification, but I’ll bear with you.

However, you’re, again, beating a dead horse here. What you’re saying has been established like two years ago.

This is what happens when you cut in into a discussion without familiarizing yourself with the topic beforehand. This is exactly what Sirlin did, and the reason why he completely misrepresented the points of the “anti-MBS” crowd, arguing against arguments from right after the initial announcement, making himself look like an arrogant douche in the process.

On February 11 2009 22:25 maybenexttime wrote: –MJAnoname

I disagree here.

The UIs we’re using in modern games are in my opinion much better than plugging the game directly into our brains.

The reason for that is the fact that, currently, we let our subconsciousness take care of a variety of tasks (we subconsciously issue orders by pressing keys and moving our mouse – keeping up with popcap, for instance), while we’re using our conscious mind for more involving problems (designing a proper Swarm push with late-game Zerg, e.g.).

In brain-computer UI all those tasks we currently do subconsciously (like “mindlessly” building Supply Depots, etc.) would have to be conscious decisions, severely taxing our DPM.

On February 11 2009 22:25 maybenexttime wrote: –toopham

Not “somewhere else” but in BASE, or at least doing macro/base related tasks.

On February 08 2009 18:56 Nimue wrote:
actually, you guys didn't miss much. There were several comments after class that the material was too basic for them. So, next week, I think I'm gonna turn it up a lot, skip the simple "what is a dragoon" "how much health does an Archon have" kind of stuff and go straight for the heavy stuff.

On February 12 2009 06:25 f10esqftw wrote:
lol wtf hapened to this thread? How did it turn into an essay writing contest about all the topics that have been argued to death already about SC2? Never mind, don't answer that question; it will just cause more of it.