|
OP's Note: While I make specific references and criticisms of certain TL posts, it is not my intention to be insulting or "call someone out." I simply want to start a discussion on practice methods in Starcraft.
I: Introduction
Recently, qxc posted a blog on Starcraft Training DBZ Style, which was spotlighted on the TL front page. For those of you who haven't read it, qxc theorizes about a form of speed-training - increasing the speed of the game in a UMS map to improve one's game.
As someone who has an extensive background in piano and music education, I'm very surprised to see a professional player take this attitude towards practicing Starcraft (my original response), and I wanted to start a discussion on practice myths and methods in Starcraft. I've always considered the learning process on musical instruments and a computer keyboard to be fundamentally similar, yet conventional community "wisdom" goes against this.
This post attempts to highlight this learning process with analogies to music education. Keep in mind that one could create an analogy with many other activities and sports.
II: Practice in Music
By understanding how a musician approaches a new piece, we can better understand how we can learn to approach learning Starcraft mechanics. For example, when a professional musician approaches a new piece of music, these are the steps usually taken:
1) Break up the piece into smaller, simpler sections.
2) Practice each section slowly with an emphasis on proper technique and efficiency.
3) Put all the smaller sections together at a slow tempo.
4) Gradually bring the piece up to proper speed.
5) Polish the finished product.
There is a clear emphasis on breaking up the learning process into smaller fragments and prioritizing slow, mechanical mastery. You will never see a competent musician start learning a piece by beginning to play it at full speed.
This practice philosophy is not exclusive to music. Many other activities, such as sports, emphasize these same traits in practice. Does a basketball player learn how to shoot a three-pointer in the middle of a heated game? Of course not! He learned how to hit that buzzer beater by spending thousands of hours shooting in practice. These practice methods are universal in every sport... except for Starcraft!
(For added effect, watch this video.)
For some reason, the popular attitudes toward practicing and improving in Starcraft are nothing like other sports and activities. These attitudes are revealed below.
III: Practice Myths in Starcraft
A quick TL search for a typical "help me improve" post reveals much about the conventional wisdom of Starcraft practice.
[H]Improving my macro
If macro is what you want to focus on, you want to focus on playing more games.
This is the most common advice I see, and is also the most shocking. A lot of TL'ers think that by playing more games, you will eventually get better at the game. While there is some truth to it, I argue that this is a very inefficient way of learning.
Telling one of my beginner piano students to practice a piece by playing it at full speed multiple times would be horrible advice. They would barely be able to coordinate themselves properly, and if by some miracle they learn the piece, they would undoubtedly play it sloppily.
This "trial by fire" method would be considered insane in all other fields, yet it is exactly the advice I see being given to Starcraft players.
Something that helps me a lot with my macro is just to go 1v1 against the AI
This suggests that we should play games against an unrealistic opponent. A computer AI behaves nothing like a human, so what is the sense in playing against something you will never see?
LaLuSh's response makes an important distinction between using an AI to start a game and playing an AI to beat it. I am against the latter, not the former.
Play against the computer to concentrate on macro and injects.
Now we're getting somewhere. The idea of targeted practice is supported by authorities such as Day[9]. While this advice is a step in the right direction, I believe it doesn't go far enough.
IV: Practice Propositions
My ideal practice method would draw as many influences from music education as possible. As such, it is based around three components: dividing an activity into sub-sections, slow mechanical mastery of each sub-section, and mastery through real situations. As an example, what if we approached a new build-order in the same way that musicians approached new music?
1) Break down the build into "phases." For example, a beginning (up until you establish your main production structures), a middle (macroing out of your base production structures), and an end (your late-game plan and tech tree).
2) Establish your "macro cycles" for each phase. Hash out exactly what buttons you will be pressing in your typical macro cycle for each phase.
3) Practice this slowly until you are comfortable with the mechanics.
4) Polish against real human opponents (i.e. mass-gaming).
Note that mass-gaming has a part in the practice process, but only after one devotes some time to slow, mechanical mastery.
In addition to learning new builds, these same ideas can also be applied to maintaining and improving macro mechanics.
Iust like a musician practices scales, and a basketball player practices dribbling, why not devote a bit of your time to practicing macro mechanics as a Starcraft player? For example, a Zerg player could devote some time at the beginning of every session to practice their larvae-inject mechanics.
V: Addressing Counter-Arguments
Inspired by r.Evo's Response (Addressing Counter-Arguments)
Starcraft and Music are different because...
Stop right here! As bmn points out, the differences are irrelevant. Learning Starcraft and music are fundamentally the same because they involve learning a new skill. This learning process is common and well-established in many other fields (sports, music, etc.), but for some reason, hasn't carried over to Starcraft.
If you need extra convincing, I encourage you to read Ghol's post, which provides an excellent Muay Thai analogy.
What's the point of practicing in this manner when it all goes to hell when you get hellion dropped?!?
The goal of slow mechanical mastery isn't simply to be able to execute your mechanics. You want to be able to execute them without even thinking. If your macro-cycles (larvae injects, SCV building, warp-ins, etc.) are second nature to you, you'll be able to devote more of your concentration to that hellion-drop or DT harass instead of thinking about your macro.
But Koreans mass game!
There's an important distinction between mass gaming and Korean targeted mass gaming. A core of the Korean practice model is isolating strategic and mechanical situations in-game. Koreans don't play games for the hell of playing them; every game is directed at improving a facet of one's game.
I'm simply proposing the logical extension of the Korean model into mechanical practice.
Starcraft is just a game!
I'd like to remind you that you are posting on a community website dedicated to competitive Starcraft. Most TL'ers take a lot of joy in improving and winning games. If that's not for you, I suggest you look at this alternative game. You'll love it. I promise.
VI: Questions and Discussion
So TeamLiquid, how do you practice Starcraft? I'd be especially interested in hearing the opinions of some high ranked/professional players.
Do you think there is any "practice wisdom" to be gained from music education?
VII: Some Interesting Discussion
QXC's Response
Destiny's Response
Ghol's Response (Muay Thai and Sub-sets)
r.Evo's Response (Addressing Counter-Arguments)
RosaParksStoleMySeat's Response (On Controlled and Automated Processes in Education)
|
Russian Federation269 Posts
|
Great post.
I find a similarity in both music or Starcraft: when playing it should be treated as art, rather than SC2 people trying so hard to be competitive on ladder when you'll end up with 50% winrate anyways. This kind of mindset in a game helps a lot.
This kind of refined and methodical practicing plan sounds definitely better in many cases, as you are breaking your gameplay down and focusing on what you know you need to practice.
And once you've done this, just like said: you will need to ladder to see if your gameplay works in realistic situations and ones you haven't prepared for in your previous practice.
As music practice methods have been developed for thousands of years, it really seems a very solid regime and I find it safe to say that bringing this over to Starcraft 2 is an amazing idea.
|
Fully agree. This sort of approach is what I used to make the jump from being a 20 apm BW noob to a D level player that could macro decently in about a week.
If you break things up its much easier to focus on, as well as perfect, that aspect. Then from their you can integrate it into the whole and move on to the next item.
|
TLADT24920 Posts
Bravo, awesome post. Couldn't agree more. I don't have much time for games so I barely even play now. When I used to play and decided to improve by focusing on the specific idea of making more workers and units as well as scouting more, I jumped 2 leagues and felt that my game improved a lot to the point where I was pretty confident of all my matchups 
|
On October 08 2011 05:54 Hapahauli wrote:Now we're getting somewhere. The idea of focused practice has been mentioned by authorities such as Day[9], and I believe this a step in the right direction. However, I think this doesn't go far enough.
A way that worked for me to fix my sloppy early-game macro was very simple: Establish a benchmark on what you want to achieve, e.g. by watching a very strong player do a similar opening, and then repeatedly play the early game up to that point and measure how well you're doing.
In my case, it was playing up to ~5:30 minutes. I had far fewer drones than I should've had. I play Zerg.
To work on that specifically, what I did was:
1) Find a replay from a strong macro player who did a standard macro opening without facing any early pressure (no 2rax/2gate).
2) Count how many drones he had at ~6:00.
3) Play a game -- not a ladder game, either vs AI or a custom -- where I know I won't be attacked for 6 minutes. Focus on good macro.
4) Stop the game after ~6 minutes, count drones. If my number is significantly lower than the baseline, watch replay.
5) Repeat step 4 until I can reliably meet the desired goal.
This is an example of practicing a very specific part of the game in isolation. Early-game macro was a source of many of my losses, so that's what I did to improve it. Within 1-2 hours, I had done more to improve my early-game macro than months of playing did. It's surprising how effective it is if you have a *specific* and *measurable* goal, and you focus on just that and nothing else for a while.
It's harder to do the same for other things -- tech decisions, late-game multitasking -- because the early game is the only part that is trivial to repeat in isolation under the same conditions. Perhaps the same can work with the unit tester map to focus on unit control. I don't know what the best way to apply it to mid/late-game problems would be, but I never tried.
|
Just my 2 cents, i could be completely off base here.
Well I agree the focusing on areas is important, and that doing it slowly then bringing it all together can provide some surprising results.
But I think you're making the mistake in comparing Starcraft 2 to learning Piano. What is starcraft 2 to the vast majority of people (>99%)? It's a game. A fun game - yes. The funnest - says Artosis. But still a game.
Whereas Piano is not just a game, it's a life-skill essentially. The ability to play an instrument is a wonderful thing, and there are a ton of studies that show that it helps kids perform better in school, etc. But when practice Piano, you know you're doing something that takes a long time to master, and that is your general goal (I'd assume, or at least general competancy).
But when most people go play SC2, they just want to steal nerd's ladder points. They want to cannon rush or macro, or marine micro or baneling bust. Then don't want to become a pro. Maybe they want to get to the next league up, whatever.
Maybe i'm wrong idk. I know that when I go on to play SC2, I just want to win and have fun, and i'm not always sure the order of those  But I know that I'll never take a day to hash out one build and practice it over and over like a pro does. Obviously there are exceptions, but I think for most people, SC2 isn't smoething they want to turn into different drills or exercises and practice the different peices individually. They want to play.
I think this is where the advice "play more games" comes from. You can obviously tell people to do it your way, which is most likey a much more efficient and effective way to do it. But most people just want to play. So the practice their injects - while playing. Or they practice their marine micro - while playing. It's what we do.
If people could "just play" piano, you'd probably see the same thing occur, why practice specific parts over and over and over if you can just learn gradually while having fun most of the time? Of course, people can't just "play piano", put me in front of a piano and I'd probably make noise that sounds like a dying cow trying to faceroll a piano. But you can "just play" starcraft to a certain degree. And while pros/dedicated/motivated people wll of course always practice methodically, I think for the most part they just want to play more games
|
|
|
Your post is very valid and I agree with almost everything you said but there's also some truth to what qxc says...
For example, I feel that Starcraft also has an emphasis on quick thinking and reaction time that music may not and in that regard playing faster games might help you by forcing you to constantly stay on your feet and keep yourself attentive.
To be honest I don't think either of you are wrong lol.
|
@ galtdunn - I believe the comparison is valid, no matter what the difference in "magnitudes" may be. My post doesn't try to make an exact comparison between Piano and Starcraft, rather, I suggest that we can learn something about practicing Starcraft from practicing Piano.
You make a good point that the advice of "play more games" is derived from playing Starcraft for pure enjoyment. However, many people derive enjoyment from improving at Starcraft, and I believe my advice caters well to this.
@shishy - My initial response to qxc's blog discusses more on that subject. I believe qxc's method is valid for certain, specific goals, such as wanting one's hands to move faster. However, I believe this training method does nothing for improving one's mechanical game. After all, I can have blazingly fast hands, and yet still be horrible at macro.
|
On October 08 2011 06:37 Hapahauli wrote:
@ galtdunn - I believe the comparison is valid, no matter what the difference in "magnitudes" may be. My post doesn't try to make an exact comparison between Piano and Starcraft, rather, I suggest that we can learn something about practicing Starcraft from practicing Piano.
You make a good point that the advice of "play more games" is derived from playing Starcraft for pure enjoyment. However, many people derive enjoyment from improving at Starcraft, and I believe my advice caters well to this.
@shishy - My initial response to qxc's blog discusses more on that subject. I believe qxc's method is valid for certain, specific goals, such as wanting one's hands to move faster. However, I believe this training method does nothing for improving one's mechanical game. After all, I can have blazingly fast hands, and yet still be horrible at macro.
Yeah I agree with you, I guess I was just pointing out that a lot of people just won't want to put that much work in In a longwinded manner I suppose. Great write-up and interesting comparison, not the first I've seen of people comparing SC2 and piano though, must be some truth there!
|
While I agree with most of what you're saying, there are certainly phases in development where just "playing games" might indeed be the most beneficial method of improving one's game. One key difference between music and games is that games are an "open" environment, where adapting strategies and thinking on one's feet are part of succeeding. At the same time, there are skills embedded in the game that considered "closed skills" which benefit greatly from incremental repetition. A real challenge then lies in balancing open and closed elements within training.
In sports training, there is a tendency to put higher volumes of closed training (simple drills, physical training, basic technique work) in the preseason and early season. Then as the season progresses, the transition is towards playing out certain game situations, doing scrimmages, intra-club competitions, etc.
|
First off this is from my phone so please forgive any mistakes.
A slow methodical approach to learning or improving Starcraft seems very practical. A common element that Day9 mentions and is commonly referenced in psychology, is that you can hold ~7 items in your short term memory at any given time. This makes nearly several new things at once extremely difficult. Remembering to micro, macro, inject, an expand is hard to do if you're new to it.
Building habits slowly over time is more than likely not the most fun thing. Starcraft is a game but to improve with efficiency requires dedication.
|
If you don't enjoy playing the game in the first place, why would you continue playing it? Even if that leads to your enjoyment of getting better, there are plenty of other games to get better at. Pick one that you fundamentally enjoy before you worry about improving.
As far as the topic in the OP, I find it similar to playing music, but it is also very different.
Rarely do I have issues in music of not being able to move fast enough to hit the notes that I want to hit. Music is about performing things at the correct tempo, not performing them as fast as humanly possibly.
This is the major difference between StarCraft and music. In StarCraft, if you can do it faster, that is better. Always, with no exceptions.
There are always more notes to be hit, always more keys to press.
This is not the same in music. For comparison of that medium, it would be that you have this endless string of smaller and faster notes to hit, but most of them are optional.
I guess it kind of has to due with the difference in performance of the two. Music is performed for an audience to listen to your perfection of a previously existing song. Starcraft is performed to beat the pulp out of your opponent.
|
On October 08 2011 06:47 zylog wrote:
While I agree with most of what you're saying, there are certainly phases in development where just "playing games" might indeed be the most beneficial method of improving one's game. One key difference between music and games is that games are an "open" environment, where adapting strategies and thinking on one's feet are part of succeeding. At the same time, there are skills embedded in the game that considered "closed skills" which benefit greatly from incremental repetition. A real challenge then lies in balancing open and closed elements within training.
In sports training, there is a tendency to put higher volumes of closed training (simple drills, physical training, basic technique work) in the preseason and early season. Then as the season progresses, the transition is towards playing out certain game situations, doing scrimmages, intra-club competitions, etc.
I agree with you completely - there are certainly times where mass gaming is good for practice, just as it is good for a musician to play through his or her pieces in their entirety.
However, that's usually in the last step of the practicing process. After we have solidified our mechanics and foundations, mass gaming is great. Unfortunately, conventional advice suggests that we should hone our mechanical skill through mass gaming instead of simple practice. My objections are against this conventional wisdom.
|
4713 Posts
Very interesting post. I didn't know you could draw such a parallel between SC2 and playing a musical instrument, having had no experience playing a instrument. I guess that logically it makes sense.
I had the same feeling, while playing practice games just to nail my macro it felt like I was having some substantial improvements compared to playing ladder.
You gave me some interesting ideas for a practice regime now.
|
@Jermstuddog - You make interesting points, however, I believe most of the differences you list are in fact similarities!
On October 08 2011 06:50 Jermstuddog wrote:
Rarely do I have issues in music of not being able to move fast enough to hit the notes that I want to hit. Music is about performing things at the correct tempo, not performing them as fast as humanly possibly.
This is the major difference between StarCraft and music. In StarCraft, if you can do it faster, that is better. Always, with no exceptions.
This is true, however, it doesn't change the validity of a practice method. If I want to learn an incredibly fast musical work, I'll approach it by the practice methods that I listed. The same method can hold true for practicing Starcraft, only that you have the luxury of setting the "speed" bar as high as you'd like.
This is not the same in music. For comparison of that medium, it would be that you have this endless string of smaller and faster notes to hit, but most of them are optional.
I'm not suggesting that we should script every mechanic that we play during a game - that's impossible. However, we can practice macro mechanics that constantly repeat themselves throughout a game. For those, we can draw upon music education.
guess it kind of has to due with the difference in performance of the two. Music is performed for an audience to listen to your perfection of a previously existing song. Starcraft is performed to beat the pulp out of your opponent.
I'd actually suggest that the two are fundamentally the same. Some of us play Starcraft in front of an audience, whether or not our audience is at an MLG live event or the viewers of a replay that we post. In addition, can we not admire the "art" of starcraft, and the beauty of a perfect game?
|
I think there's a way of practice that includes both qxc's and your approach, and is superior to both at the same time.
Here's what I mean.
StarCraft is a strategy game, so ideally you'd want all your losses to come from simply being outsmarted strategically. You don't want to be superior to your opponent but him winning just because you couldn't execute the plan. So obviously to make sure that doesn't happen, your mechanics should be solid as steel.
-> Our goal, steel solid mechanics.
We get there by making sure our mechanics hold even in the heavier-than-normal circumstances.
That's the part QXC is accentuating. Ultra-fast speed surely is heavier-than-normal, so that's covered.
However, it's wrong to start at high speeds without making sure the technique is right.
If you draw parallels with real sports, the gym work athletes do regularly could be compared to mechanics practice in StarCraft.
You want to know the tactics of the sport, but when it comes to making it happen, you need to make sure you're able to complete the run/jump well/hit the ball powerfully.
So, you're in the gym now. Your goal is to get stronger. It would be foolish to lift your maximum weight right away if you wish to progress. You start by making sure your technique is spot on, and from there you increase resistance, making sure that the weakest link in your body is strenghtened, and then the next weakest link and then the next. You polish your quirks while increasing resistance.
So in the end, you have good technique, and are lifting huge weights, you are strong!
Then you can be a top player, if your strategy and tactics are top level too.
|
Sports and e-sports are different than your piano lessons or weight lifting analogies...
You want to learn over time in the most effective environment. If you want to be fast, then its a fast environment you want to place yourself in and grind it out. "Professionals", yet you do not brush upon this simple, natural concept.
|
I think partly the difference in opinion between the approach your suggesting and qxc's post is his is intended more for people that already very good, whereas you are talking about an approach that is suitable for beginners.
That's why he draws a metaphor to training done by professional athletes and you talk about teaching a beginner a new piece on piano.
The most helpful form of training for me though has definitely been to focus on specific areas and "drill" them, but then again, I'm a beginner...
|
|
|
|
|
|
EPT
