EPTIn my previous post (link), I ended with a list of topics I wanted to flesh out in future posts, the first being:
If you had to make a list of ‘learnable skills’ in Starcraft, what would be on it?
I got several replies to this post with players giving their answer to that question, which I indicates that it’s certainly a topic we’re all thinking about. Before I even started playing 1v1 games, this question was burning in my mind.
I recently finished the book ‘The Art of Learning by Josh Waitzkin, the chess-prodigy who was the inspiration for the movie ‘Searching for Bobby Fischer’. In it, Josh parallels the learning process he went through when he first learned chess and when he later began with Tai Chi Chuan. Both processes began with the practice of simple skills (extracted from the game / sport) and practiced until that skill became almost second nature. At that point another skill was introduced – which included the use of the first – and again practiced until it felt natural.
In both chess and Tai-Chi Chuan, it seems that there exists a cumulative body of knowledge which has defined what these primary skills are, and in what order beginners should learn them. I find two important characteristics here that are important to note:
- In both disciplines, there is clear reasoning behind why these skills are considered primary
- Each skill is clearly defined
So, for example, Josh describes how in learning chess with his first master they began by going over simplistic scenarios on the board with only a few pieces. The reasoning: to develop an understanding of the subtle relationship between the different pieces, in a way that’s simplistic and removed from the complexity of an actual game. The definition: practice a three-piece scenario with sets of pieces A, B, C, etc. Notice how the practice description wasn’t “work on subtle piece relationships” or something abstract like that – it had definition. With a game like chess, one can only imagine the amount of complexity involved in high-level play, and for a beginner, just like one in Starcraft, I’m sure that simply witnessing that play can be intimidating. This approach helps to remedy that, but also has a clear purpose to it.
Turning to Starcraft
My last post garnered some great responses, some of which I just want to quote here briefly:
+ Show Spoiler +
On January 16 2010 14:43 ella_guru wrote:
Now, best things to do when learning SC imo:
1. Dont stop making scvs. Have them command center lights always flashing
2. Keep minerals low
3. Keep scvs mining/gas/constructing. no idle scvs
4. Learn a BO.
5. Dont get supply stuck
6. Practice, practice, practice
To this properly can take a long time! I have logged many games and cant keep up with it all. Enjoy practicing these things and it is easier. Through in fancy dropship stuff or whatever when you want, but know that consistency comes from these things, as opposed to flukey cheese.
Now, best things to do when learning SC imo:
1. Dont stop making scvs. Have them command center lights always flashing
2. Keep minerals low
3. Keep scvs mining/gas/constructing. no idle scvs
4. Learn a BO.
5. Dont get supply stuck
6. Practice, practice, practice
To this properly can take a long time! I have logged many games and cant keep up with it all. Enjoy practicing these things and it is easier. Through in fancy dropship stuff or whatever when you want, but know that consistency comes from these things, as opposed to flukey cheese.
+ Show Spoiler +
On January 16 2010 08:08 Chef wrote:
I don't want to read your life story (sorry) but I will give my thoughts on your questions.
If you had to make a list of ‘learnable skills’ in Starcraft, what would be on it? How would you rate the skills in term of importance to winning?
All skills in StarCraft are learnable. Sometimes we have other hobbies that can transfer to StarCraft (and vice versa), but everything can be learned through StarCraft effectively. I don't think there is a most important skill. It's like trying to decide which is more important to living an active life, your heart or your brain? You need them both to at least some capacity. I would say the best thing for a beginner to focus on is the weakest parts of his game. If that means learning how to move a mouse around and select units, that's what's going to help him win the most. If he knows how to do that but is always choosing the wrong units, then he needs to work on his strategy.
When I was first learning, I'd always look for the most blatant holes in my game play and fix that. That's what's going to give you the highest risk of losing, and therefore what you need to work on to win. There isn't any one thing.
What’s the best way to approach learning a matchup?
Play against a friend who tries a variety of different builds. If you play the same matchup with them game after game you will learn much faster than say, playing a matchup every couple of games. Other things that will help you is reading a guide to the matchup you're playing before playing with your friend, so that you have a large variety of things to think about and incorporate into your game while you play.
I would avoid playing with someone who is going to 'teach' you StarCraft, because most StarCraft players are not teachers. You're better off playing with people your own skill who aren't going to play like retards just because they're bored. I'm not exactly speaking from experience here (I haven't had the dedication to get on a schedule with a 'teacher,' but I have played with rivals), but I really don't think most people have any clue about how to effectively teach someone. Pick up advice here and there, but don't listen to someone who is going to dictate every aspect of your game.
What’s the most efficient way to learn a build?
Play on an empty map to practice the build (if you cancel a computer's slot at countdown you will be in the game by yourself).
Numbers are typically the wrong way to go about learning a build (Day[9] even preaches this). Numbers won't tell you where to place your buildings or where to move your units. You can learn this from a VOD, but at your stage it's probably a good idea to enjoy making your own builds while they can still be effective. Just have solid goals, and ideas of where you want your build to bring you in the mid game and the late game. Refine it after practice games with real opponents.
I really think the fastest way to learn StarCraft is just to have a rival though. Nothing will motivate you to win like someone you know with pride on the line.
I disagree with that. I think that a player who cannot vary his strategies is not learning the game. Maybe it's the fastest way to win games on iCCup, but it becomes a real hindrance to the evolution of players like that when they plateau and realise they can't play more than 3 builds and they're uncomfortable trying new ones against players their skill level. Variation is important to keeping things interesting and therefore keeping the brain active while learning.
I don't want to read your life story (sorry) but I will give my thoughts on your questions.
If you had to make a list of ‘learnable skills’ in Starcraft, what would be on it? How would you rate the skills in term of importance to winning?
All skills in StarCraft are learnable. Sometimes we have other hobbies that can transfer to StarCraft (and vice versa), but everything can be learned through StarCraft effectively. I don't think there is a most important skill. It's like trying to decide which is more important to living an active life, your heart or your brain? You need them both to at least some capacity. I would say the best thing for a beginner to focus on is the weakest parts of his game. If that means learning how to move a mouse around and select units, that's what's going to help him win the most. If he knows how to do that but is always choosing the wrong units, then he needs to work on his strategy.
When I was first learning, I'd always look for the most blatant holes in my game play and fix that. That's what's going to give you the highest risk of losing, and therefore what you need to work on to win. There isn't any one thing.
What’s the best way to approach learning a matchup?
Play against a friend who tries a variety of different builds. If you play the same matchup with them game after game you will learn much faster than say, playing a matchup every couple of games. Other things that will help you is reading a guide to the matchup you're playing before playing with your friend, so that you have a large variety of things to think about and incorporate into your game while you play.
I would avoid playing with someone who is going to 'teach' you StarCraft, because most StarCraft players are not teachers. You're better off playing with people your own skill who aren't going to play like retards just because they're bored. I'm not exactly speaking from experience here (I haven't had the dedication to get on a schedule with a 'teacher,' but I have played with rivals), but I really don't think most people have any clue about how to effectively teach someone. Pick up advice here and there, but don't listen to someone who is going to dictate every aspect of your game.
What’s the most efficient way to learn a build?
Play on an empty map to practice the build (if you cancel a computer's slot at countdown you will be in the game by yourself).
Numbers are typically the wrong way to go about learning a build (Day[9] even preaches this). Numbers won't tell you where to place your buildings or where to move your units. You can learn this from a VOD, but at your stage it's probably a good idea to enjoy making your own builds while they can still be effective. Just have solid goals, and ideas of where you want your build to bring you in the mid game and the late game. Refine it after practice games with real opponents.
I really think the fastest way to learn StarCraft is just to have a rival though. Nothing will motivate you to win like someone you know with pride on the line.
I disagree with that. I think that a player who cannot vary his strategies is not learning the game. Maybe it's the fastest way to win games on iCCup, but it becomes a real hindrance to the evolution of players like that when they plateau and realise they can't play more than 3 builds and they're uncomfortable trying new ones against players their skill level. Variation is important to keeping things interesting and therefore keeping the brain active while learning.
I love this kind of discussion, because it allows us to see how others are thinking about the game, and to question our own approaches to learning.
For a fun analytical experiment, and for the sake of my own training, I wanted to see if I could break down the matchup that I’ve been working on (T v P) in the way that Waitzkin did: brainstorm primary skills important to a beginner with clear reasoning behind them, clearly define those skills in real terms, and rank those skills based on an order of importance.
To do this I needed to have some criteria for figuring out which skills in the matchup should be considered ‘primary’. I decided on the following:
A - How relevant is the skill to developing a player’s understanding of the matchup’s flow (and consequently the higher-level game)?
For example: If the only thing your practicing is an early-game rush, then you’ll either win early or die early each game. This does little to help you understand how T v P ‘flows’, and thus would rate low based on this criterion.
B - How relevant is the skill to a novice-level match?
For example: Science vessel micro is an essential skill for higher-level players, but at the D-level most games don’t even reach the point where you need to do very much with your vessels. So should you really be spending your time practicing that?
C - How relevant is the skill to winning the game?
For example: A good, well-timed push is more likely to win you more games than good sim-city.
On their own each of these criteria has at least one logical hole, which is why I decided on three. For example, a perfectly executed BBS Rush can probably increase your win rate at the D-Level if you only do that, so it ranks high with criterion C. However, that kind of rush will end the game extremely early (one way or the other), so it violates criterion A.
My Terran versus Protoss Skill List
A few nights ago I took out a piece of paper and started working on this list. Please note: of course this kind of thing is highly-debatable. This is my initial logical progression, and I’m sure I’ll be able to refine it better with more thought and time.
#1 – Learning one build order that is standard for this matchup, that transitions into the mid- and late-game well. In my case, I chose the 1 Fact FD build. (Criteria fulfilled: A, B, C)
Reasoning: A good build order, if selected properly, scores high on all three criterion: it allows you to get to the later stages of the game (thereby allowing you to experience the game’s flow); most novice Protoss players will have well-timed opening builds, so in the early game it is important to be accurate with your own to keep pace; a well-executed build is, of course, the way to set yourself up to win…
More… + Show Spoiler +
Notice how I defined it as ‘one build order, standard for this matchup, that transitions into the mid- and late-game well’. I didn’t say ‘learn T v P builds’, and that I defined the type of build to learn. Again, learning a BBS rush violates criterion 1, and really doesn’t give you any exposure to the flow of the game.
Personally I picked the 1 Fact FD build, for all the reasons I stated. I think it’s a well-rounded build that works well for most scenarios. My experience with it is that at the D-level, the only huge change I have to make to the build is when a Protoss early expands, but that seldom happens in my experience.
Personally I picked the 1 Fact FD build, for all the reasons I stated. I think it’s a well-rounded build that works well for most scenarios. My experience with it is that at the D-level, the only huge change I have to make to the build is when a Protoss early expands, but that seldom happens in my experience.
#2 – Using an SCV scout to read the Protoss build and to determine what specific adjustments are required as a result. (Criteria fulfilled: A, B, C)
Reasoning: In my opinion, D-level Protoss players on ICCUP are as sneaky as they come, and so if it were for only to deter all sorts of cheese, scouting would be important. However, my reasons go deeper than that...
More… + Show Spoiler +
Of course scouting is important to winning games (C), but I wanted to avoid just accepting it as an immutable fact. In T v P, scouting allows you to adjust to survive early cheese (novice..) and aggression, so it satisfies (B). It also allows you to survive to the mid- and late-game, which satisfies (A). However, it also satisfies (A) because scouting a Toss should set-off a logical progression in your own mind:
How to win: Toss is doing this -> I will respond this way -> they’ll act this way -> I’ll counter this win
The flow of the game should (in some simplistic way) unfold in your mind as you scout. Of course you’ll need to make adjustments as the game goes on, but you should at least have some idea of how things will unfold.
The reason I placed build as #1 and scouting as #2 (though it seems obvious) is that when you have a good build, you might just get lucky and anticipate what the Toss will do without proper scouting. However, it doesn’t work in the opposite direction: without a good build, you’ll lose, even if you can do fancy twirls with your scouting SCV.
I also said ‘specific adjustments’ to more precisely define a plan of action. So instead of this:
SCV scouts 2 gates and range running -> get ready for early aggression
…in your mind it should be something like:
SCV scouts 2 gates and range running -> get two extra marines, position tank properly
This definition might change as you practice, but at least you now have something specific and definable to practice.
How to win: Toss is doing this -> I will respond this way -> they’ll act this way -> I’ll counter this win
The flow of the game should (in some simplistic way) unfold in your mind as you scout. Of course you’ll need to make adjustments as the game goes on, but you should at least have some idea of how things will unfold.
The reason I placed build as #1 and scouting as #2 (though it seems obvious) is that when you have a good build, you might just get lucky and anticipate what the Toss will do without proper scouting. However, it doesn’t work in the opposite direction: without a good build, you’ll lose, even if you can do fancy twirls with your scouting SCV.
I also said ‘specific adjustments’ to more precisely define a plan of action. So instead of this:
SCV scouts 2 gates and range running -> get ready for early aggression
…in your mind it should be something like:
SCV scouts 2 gates and range running -> get two extra marines, position tank properly
This definition might change as you practice, but at least you now have something specific and definable to practice.
#3 – Being able to produce units efficiently for the mid-game timing push, or mid-game macro. (Criteria fulfilled: B, C)
Reasoning: I didn’t want to use the term ‘macro’, because its definition in common usage is vague at best. What I mean specifically is achieving the goal of having X number of units by Y time, for a specific in-game action (the timing push). I didn’t say ‘work on the timing push’, ‘work on timing’, or anything similarly vague. There is a clear goal here...
More… + Show Spoiler +
It satisfies (B) since I would imagine that most D-level TvP games are won or lost by the push, and having enough units with the correct composition at the correct time are all required ingredients for a successful one. In the same way, it satisfies (C).
I didn’t indicate that it satisfies (A), because I see it as more of a set skill which doesn’t directly allow you to experience game flow. So, it’s part of the flow, but not something that leads to you learning it.
I didn’t indicate that it satisfies (A), because I see it as more of a set skill which doesn’t directly allow you to experience game flow. So, it’s part of the flow, but not something that leads to you learning it.
#4 – Being able to use your vultures, tanks and goliaths comfortably to perform an effective push. (Criteria fulfilled: B, C)
Otherwise known as ‘push micro’, being able to coordinate your units in a push required for a push to be successful. A perfectly-timed and well-composed push can fail miserably if your tanks are caught un-sieged, or your vultures get stuck in your minefield, etc. Again I stayed away from simply saying ‘push micro’ because it is vague, and doesn’t give a new player guidance as to what exactly they should be practicing...
More… + Show Spoiler +
With the same reasons as in my explanation for #3, ‘push micro’ satisfies both (B) and (C), and doesn’t really satisfy (A). The reason I put it after mid-game macro is that at the D-level you can often get away with poor micro if you have all the right units at the right time. However, even if your micro is great, It’s harder to get away with a good push when wrong number of units at the wrong time.
Applications to a New Starcraft Player
Though I hope to use my own experiences to test this analysis, it’s easy to speculate on how this kind of thing can be used to help a new player improve:
- Knowing which skills are most important to your stage in learning helps focus your energies and attention where you’ll see the biggest return.
- Defining each skill as specifically as possible, with clear goals, helps you know exactly what you have to learn, with the ability to use progress markers.
- By doing both of the first two, you can easily filter through the seemingly endless amount of Starcraft-related information available, to find the information that is relevant to you.
--
Phew. Hopefully in the next week or so I can post some of my own replays and talk about ways of approaching each skill. I’m experimenting with a few right now, and I want to see them through before I post. It’ll be fun!
Until then -
P.S. One final note:
Yes, of course I left out quite a few details, and I’m sure I’ll have at least one reply to this telling me that I forgot to list one thing or the other. That’s fine! – all I ask is that you go through a similar logical progression yourself: figure out if you agree with my criteria or if you have your own, and based on that, why should a skill be included or excluded? How does it help a new player understand the flow of the game? Develop their mechanical skills? Etc.
