[D] Practice Myths and Methods - Page 10

On October 14 2011 19:30 Crusnik wrote:
Interesting read, I'll have to go over all of the comments later on when I actually have time.

Personally though, and I know this is does follow conventional wisdom, but I do prefer looking at my replays and just seeing what I did wrong, even when I win. I feel like that gives me specific areas to work on in games, both on the ladder (currently re-teaching myself through bombed placements and working back up the rankings) and against teammates/friends in custom made games.

On October 14 2011 19:43 ApocAlypsE007 wrote:
To the OP: what do you think about the practice methods showed by day9 in daily #360?
I plan to spend an entire week at least practicing like that so do you think it's good?

On October 18 2011 04:05 gatorling wrote:
I'll have to agree with the op
Practice basic mechanics until they become automatic and worry about stringing it all together after you've attained mastery of the basics.
Unfortunately most of us don't have the time or patience to tirelessly practice the basics.
I read somewhere that it takes around 10,000 hours to master a skill. That's roughly 6000 games

On October 08 2011 05:54 Hapahauli wrote:
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.

On November 13 2011 19:58 RosaParksStoleMySeat wrote:
I would like to contribute some educational theory as to why, while training, we break complicated tasks down into smaller components to achieve mechanical mastery.

Now, when you look at a Starcraft II game played at a high level, it is extraordinarily complicated. Let's look at some of the stuff that goes on in the first few minutes of a game:

*Keeping peon production constant.
*Putting peons on closer mineral patches for optimal mining.
*Scouting with a peon in a manner in which that unit is kept moving at all times.
*Building production buildings, not missing supply buildings/units, researching necessary upgrades.
*Timing expansions and maynarding peons to them.
*Setting building waypoints.
*Microing units against scouts/early aggression.
*(Zerg) Injecting larvae constantly and spreading the creep highway.
*(Terran) Calling down MULES for additional minerals.
*(Protoss) Chrono Boosting structures and upgrades for maximum productivity.

How can we possibly perfect all of these timings and actions? The answer is in our internal neural networks. According to connectionist theory, the human brain is composed of a complex system of neural networks, in which pieces of knowledge are connected between these networks, and become associated with each other.

When we are first learning how to do something, we learn it as a controlled process. For fun, let's take a game example: Final Fantasy VIII's gunblade system. Our main character Squall wields a gunblade, and to maximize damage, you have to hit R1 immediately as he strikes. The first couple hundred times we do this, we think about the timing, pay attention to the sound effects, watch him carefully, and hit R1 when we feel it is appropriate.

As we practice this action, it becomes what is called an automated process. In connectionist terminology, the association between two actions (selecting a physical attack and his gunblade hitting the mob) has become so strengthened in our neural network that we do it automatically. It would even be remarkably difficult not to activate his gunblade in this attack. As a matter of fact, on several occasions in my play through of Final Fantasy VIII, Squall had a confusion spell cast on him, attacked his own teammate, and I had already learned the process of activating his gunblade to such automaticity that I hit R1, inflicting additional unwanted damage on my own teammate. Obviously, most of the time knowing something to the point of automaticity is a good thing, but this is beside the point. Concentrated repetition made me learn this action. Practice made perfect.

This is why we should practice Starcraft II games in smaller segments. As we focus on one task, we can change a controlled process into an automatic one. This automatic process will be integrated into our main gameplay, and improve our performance better than any amount of random ladder games could.

I would expand past what the OP is talking about for this, though. Does anybody remember in Starcraft: Brood War, during B.Net attack, we saw Firebathero reset the rally points of 15 barracks in something like two seconds? He basically hit F2 (to position his camera to the barracks), selected the first one, hit F3 (to position his camera to the desired rally point), right clicked, then repeated with F2->left click->F3->right click for each individual barracks. I'm willing to bet that he did not learn how to do this in normal games, no matter how numerous they were. He sat there in a custom game for hours practicing this sequence, and then incorporated it into his main play. While Starcraft II is a little bit easier mechanically, I feel the same thing can be done for other tasks.

On November 13 2011 19:58 RosaParksStoleMySeat wrote:
I would like to contribute some educational theory as to why, while training, we break complicated tasks down into smaller components to achieve mechanical mastery.

Now, when you look at a Starcraft II game played at a high level, it is extraordinarily complicated. Let's look at some of the stuff that goes on in the first few minutes of a game:

*Keeping peon production constant.
*Putting peons on closer mineral patches for optimal mining.
*Scouting with a peon in a manner in which that unit is kept moving at all times.
*Building production buildings, not missing supply buildings/units, researching necessary upgrades.
*Timing expansions and maynarding peons to them.
*Setting building waypoints.
*Microing units against scouts/early aggression.
*(Zerg) Injecting larvae constantly and spreading the creep highway.
*(Terran) Calling down MULES for additional minerals.
*(Protoss) Chrono Boosting structures and upgrades for maximum productivity.

How can we possibly perfect all of these timings and actions? The answer is in our internal neural networks. According to connectionist theory, the human brain is composed of a complex system of neural networks, in which pieces of knowledge are connected between these networks, and become associated with each other.

When we are first learning how to do something, we learn it as a controlled process. For fun, let's take a game example: Final Fantasy VIII's gunblade system. Our main character Squall wields a gunblade, and to maximize damage, you have to hit R1 immediately as he strikes. The first couple hundred times we do this, we think about the timing, pay attention to the sound effects, watch him carefully, and hit R1 when we feel it is appropriate.

As we practice this action, it becomes what is called an automated process. In connectionist terminology, the association between two actions (selecting a physical attack and his gunblade hitting the mob) has become so strengthened in our neural network that we do it automatically. It would even be remarkably difficult not to activate his gunblade in this attack. As a matter of fact, on several occasions in my play through of Final Fantasy VIII, Squall had a confusion spell cast on him, attacked his own teammate, and I had already learned the process of activating his gunblade to such automaticity that I hit R1, inflicting additional unwanted damage on my own teammate. Obviously, most of the time knowing something to the point of automaticity is a good thing, but this is beside the point. Concentrated repetition made me learn this action. Practice made perfect.

This is why we should practice Starcraft II games in smaller segments. As we focus on one task, we can change a controlled process into an automatic one. This automatic process will be integrated into our main gameplay, and improve our performance better than any amount of random ladder games could.

I would expand past what the OP is talking about for this, though. Does anybody remember in Starcraft: Brood War, during B.Net attack, we saw Firebathero reset the rally points of 15 barracks in something like two seconds? He basically hit F2 (to position his camera to the barracks), selected the first one, hit F3 (to position his camera to the desired rally point), right clicked, then repeated with F2->left click->F3->right click for each individual barracks. I'm willing to bet that he did not learn how to do this in normal games, no matter how numerous they were. He sat there in a custom game for hours practicing this sequence, and then incorporated it into his main play. While Starcraft II is a little bit easier mechanically, I feel the same thing can be done for other tasks.

On November 13 2011 19:28 Honeybadger wrote:
I agree with everything you say, but I wanted to point out that I'm an extremely talented musician (tenor, alto, baritone, bass saxophone, cello, won "best saxophone soloist" in the state all four years of high school) and I start learning a piece by first playing once through, in its entirety, at double speed just because it makes me laugh. Nothing is funnier than classical music at a Benny-Hill tempo.

Though before that, I try to listen to it before playing anything.

Damnit, now I want to pull out my cello and try learning "through the fire and flames." There goes my night. Thanks. You jerk.

And on the starcraft topic, QXC is 100% correct. I'm going to compare Starcraft to sixteenth and thirty second-th note runs. As you mechanically are capable of playing the piece, the way you prepare for possible errors is to train pure muscle memory. if I'm playing a sixteenth note-primary solo (coletrane for example) at an 80 tempo, I'll learn it at 40, then 60, then 80. But then I'll push it to 100. And 120. 140. 160. 200.

If I can't play a piece (within reason) at a 200 tempo, I don't consider myself proficient at it. A problem many people encounter is that by only working with normal speed and below is that your body attunes itself to the speed you're most used to. If you train at "faster" all the time (or 120bpm on a piece written for 120) when the time comes for a wrench to be thrown in the works, you won't have the reaction time to deal with it. But if your brain is used to 180bpm or "fastest" then suddenly "faster" or 120 feels like freaking bullet time. Play a game of starcraft at "normal" speed and tell me it doesn't feel like the easiest thing in the world. Your brain is capable of reading three lines of text at once, and scanning an entire page of text in three seconds or less, solving a rubicks cube in under 10 seconds, etc. Starcraft is just another form of speed chess.

(this is all assuming you have basic mechanics and fundamentals down.)