EPT+ Show Spoiler +
I am TheKingRoc, an avid follower of Starcraft 2 and eSports, and am diamond on the NA ladder.
Often when I listen to professional players or commentators, they seem to make a connection between the speed at which a player is able to complete a task and their Actions Per Minute; the purpose of this post is to debunk this myth, and provide a more accurate description and metric for a players ACT (more on this later). I will also provide some simple neurological techniques that can help any player with 15 minutes to increase those same numbers.
Disclaimers:
+ Show Spoiler +
1. This post will not help you to increase your APM, and may even decrease it.
2. This post is not related to any specific strategy, and should not be viewed as a tool to help you beat other players. This guide is here to make you more efficient with the keyboard and mouse.
3. This post contains fairly technical terminology. No prior knowledge is required to comprehend all written information, and I do try to be very concise and transparent, but please take the time to understand everything outlined before bashing in the comments section.
4. This post is in numerous areas may not technically coincide with relevant scientific research. Many things were changed in order to make the subject matter more apparent in it's relevance to Starcraft, please understand that all of the concepts are indeed concurrent with preexisting scientific data.
Definitions:
+ Show Spoiler +
SC2 APM:
+ Show Spoiler +
Most people know the answer to this question; your Actions per Minute in Starcraft 2 is the amount of times you depress a key with a relevant hotkey, right click, or left click every 1.39 minutes.
Layman's terms: How fast you hit the keyboard and mouse.
Layman's terms: How fast you hit the keyboard and mouse.
ACT:
+ Show Spoiler +
ACT, or action completion time, is the amount of time (generally measured in milliseconds) is the time it takes an individual to mentally process a set of instructions and physically complete a relevant action. This is actually split in to two metrics ACT and PCT (Process Completion Time) but for the sake of comprehension, this post will be categorizing both as the former.
ACT is made up of your MCT and PCT.
Layman's terms: How long it takes you to do a specific task.
ACT is made up of your MCT and PCT.
Layman's terms: How long it takes you to do a specific task.
Handspeed:
+ Show Spoiler +
The speed at which a player's hand is able to move from a starting point to a desired destination; this should be metricized with three to five point times (this is basically measuring how long it takes your hand to go from a key to another in milliseconds; generally QP, AL, Z>, QM, and ZP are used as starting points and desired destinations), but this in a practical sense cannot be done to a highly accurate degree.
Layman's terms: How fast your hands can move from one place to another on the keyboard.
Layman's terms: How fast your hands can move from one place to another on the keyboard.
MCT:
+ Show Spoiler +
This is a bit more complicated. Mental Classification Time (MCT) is the time it takes for your brain to pick up a stimuli from your senses (the pixels that you identify as a Thor, for example), send it through your mental decision making process, decide upon a response (lets say this is a magic box), and send the relevant actions required to complete that same response to a selected grouping of muscles. Sometimes, this response is no response.
MCT is half of what makes your ACT.
Layman's terms: The time it takes your brain to process information, and tell your hands to do something with it.
MCT is half of what makes your ACT.
Layman's terms: The time it takes your brain to process information, and tell your hands to do something with it.
PCT:
+ Show Spoiler +
The time it takes your fingers and hands to complete the tasks assigned to them by your brain. This includes all actions required to get to the point where you can complete the task (sometimes the human brain fucks this step up) and the actual completion of the task itself.
PCT is the other half of your ACT (not literally half, just for simplicity's sake).
Layman's terms: How long your fingers take to do what your brain says.
PCT is the other half of your ACT (not literally half, just for simplicity's sake).
Layman's terms: How long your fingers take to do what your brain says.
PIT:
+ Show Spoiler +
The time your muscles spend idling (physical idle time), generally during MCT.
Why APM is Truly (sort of) a Useless Statistic:
+ Show Spoiler +
To answer this question, let's first recall what APM is: How many times you hit the keyboard and mouse in about a minute.
Now, you also need to realize that for the entire duration of a SC2 match, all you do is create and complete these little, short tasks; you do it hundreds of times each and every game.
Therefore, using common sense, we can deduce APM is actually a measure of [(Number of completed tasks * average actions included per task) completed per 1.39 minutes], which makes it basically a rough measure of how many tasks you assign yourself. You can't really improve this number (it's pretty much defined by how mentally "quick" you are; if you can find a way to make yourself smarter, be sure to tell me), and you definitely don't want to increase the length of each task, so APM is down the shitter as far as a measure of speed.
APM is a poor and inconsistent measure of hand speed, because many players do not utilize the entirety of their handspeed (none are required to) for a significant portion of the game, so APM sucks again here.
The only real use for APM would be to measure the MCT and PCT (moreso the PCT) of one player against the other; keep in mind this statistic is really, really skewed as well, because of that same lack of full mental and physical application across the board for the whole duration of measurement.
Now, you also need to realize that for the entire duration of a SC2 match, all you do is create and complete these little, short tasks; you do it hundreds of times each and every game.
Therefore, using common sense, we can deduce APM is actually a measure of [(Number of completed tasks * average actions included per task) completed per 1.39 minutes], which makes it basically a rough measure of how many tasks you assign yourself. You can't really improve this number (it's pretty much defined by how mentally "quick" you are; if you can find a way to make yourself smarter, be sure to tell me), and you definitely don't want to increase the length of each task, so APM is down the shitter as far as a measure of speed.
APM is a poor and inconsistent measure of hand speed, because many players do not utilize the entirety of their handspeed (none are required to) for a significant portion of the game, so APM sucks again here.
The only real use for APM would be to measure the MCT and PCT (moreso the PCT) of one player against the other; keep in mind this statistic is really, really skewed as well, because of that same lack of full mental and physical application across the board for the whole duration of measurement.
How individual or compiled ACT is a better measure, even if there isn't really a way to measure it:
+ Show Spoiler +
In simple and perfectly adequate terms, ACT is a more specific measure of your speed, with better delimiters and no possible inaccuracy.
The reasoning behind this is that ACT is a very small-scale measurement (generally used to find only the speed of three or four events in sequence) that inaccuracies are unlikely, because you're not comparing 15 minutes of data versus 15 minutes of data.
ACT is your "true" speed, in the sense that it measures the timeliness in which you can begin and finish just about any task. In SC, we can use ACT on a larger scale to measure tasks commonly required for effective macro, like making an SCV, but more on that later.
The reasoning behind this is that ACT is a very small-scale measurement (generally used to find only the speed of three or four events in sequence) that inaccuracies are unlikely, because you're not comparing 15 minutes of data versus 15 minutes of data.
ACT is your "true" speed, in the sense that it measures the timeliness in which you can begin and finish just about any task. In SC, we can use ACT on a larger scale to measure tasks commonly required for effective macro, like making an SCV, but more on that later.
How does one improve their ACT?
+ Show Spoiler +
Like nearly everything a human does, the only way to improve the speed at which one completes any particular action is to complete that action a batrillion times. In essence, every task that you want to do faster, you have to do a lot to do faster (I apologize for that sentence).
However, not all practice is truly efficient. For instance, if you load up a ladder game and try to do everything required for good macro, drop, and micro and the same time, your brain will learn how to do all of those things at a much slower rate, and the acquired knowledge will not independently assort; this basically means that you learn things in the context of other things, and if you don't do those other things you suck at the thing you learned.
Therefore, the best way to learn any particular task is to take it to the smallest level, and surround it with itself. So, to put that into english: if you make thirty SCVs in a row you know how to make SCVs better than if you, say, make 15 SCVs, a marine, then 15 more SCVs.
Make your goal to isolate problem areas where your fingers fumble (this is where we start to apply some of those far away definitions), and work on completing those motions a million times.
Let's say that you have a problem double tapping your two hotkey, then your 1 hotkey, and queuing a probe. You should work on doing those exact motions repetitively to best isolate the problem and train your fingers to quickly complete the motion. Make sure to iron out all of the really obvious keystroke problems you have, before moving on to the MCT area.
It's also extremely beneficial to practice making quick responses to stimuli in the game to decrease your MCT: load up a few custom games, and tell a friend to do a different 6:30 push ever game. Scout, and try to read the information and make the correct decisions to defend the push every time.
Truly, the best practice for improving your ACT is to be consistently aware of stutters in your process; if you know you have to do something but can't seem to get to it, there's a PCT problem and you should fix it by repetition. Watch your replays; if you see that you could have better defended a push, make sure you know the correct responses and properly execute them; one key factor in a good MCT is familiarity with the desired response. Always be sure to know what to do; if you don't, than go read the rest of the strategy forum before finishing this post
However, not all practice is truly efficient. For instance, if you load up a ladder game and try to do everything required for good macro, drop, and micro and the same time, your brain will learn how to do all of those things at a much slower rate, and the acquired knowledge will not independently assort; this basically means that you learn things in the context of other things, and if you don't do those other things you suck at the thing you learned.
Therefore, the best way to learn any particular task is to take it to the smallest level, and surround it with itself. So, to put that into english: if you make thirty SCVs in a row you know how to make SCVs better than if you, say, make 15 SCVs, a marine, then 15 more SCVs.
Make your goal to isolate problem areas where your fingers fumble (this is where we start to apply some of those far away definitions), and work on completing those motions a million times.
Let's say that you have a problem double tapping your two hotkey, then your 1 hotkey, and queuing a probe. You should work on doing those exact motions repetitively to best isolate the problem and train your fingers to quickly complete the motion. Make sure to iron out all of the really obvious keystroke problems you have, before moving on to the MCT area.
It's also extremely beneficial to practice making quick responses to stimuli in the game to decrease your MCT: load up a few custom games, and tell a friend to do a different 6:30 push ever game. Scout, and try to read the information and make the correct decisions to defend the push every time.
Truly, the best practice for improving your ACT is to be consistently aware of stutters in your process; if you know you have to do something but can't seem to get to it, there's a PCT problem and you should fix it by repetition. Watch your replays; if you see that you could have better defended a push, make sure you know the correct responses and properly execute them; one key factor in a good MCT is familiarity with the desired response. Always be sure to know what to do; if you don't, than go read the rest of the strategy forum before finishing this post
Do I need to measure my ACT in order to improve it? Hell, what am I actually improving?
+ Show Spoiler +
No! Knowledge of the defintion of ACT (or any controllable statistic, for that matter) is enough to positively effect your ACT.
It does help, however, to know that you can affect three key areas that make up ACT: PCT, MCT, and indirectly, PIT (These terms are all defined above). Every time you increase the speed at which you preform an action, you are almost positively decreasing the length of all three of these things independently of one another.
It's also good to know that up to around 275 SC2 APM, an increase in a consistently present action like the production of an SCV will reflect in your APM. The reasoning for this is that for the majority of the mid and late game, you have tasks that make it through the MCT phase, but are bumped off at the PCT phase because of the limited queue of the physical reflex (sorry about how french that probably sounds).
It does help, however, to know that you can affect three key areas that make up ACT: PCT, MCT, and indirectly, PIT (These terms are all defined above). Every time you increase the speed at which you preform an action, you are almost positively decreasing the length of all three of these things independently of one another.
It's also good to know that up to around 275 SC2 APM, an increase in a consistently present action like the production of an SCV will reflect in your APM. The reasoning for this is that for the majority of the mid and late game, you have tasks that make it through the MCT phase, but are bumped off at the PCT phase because of the limited queue of the physical reflex (sorry about how french that probably sounds).
FAQ:
+ Show Spoiler +
[i]You don't give APM enough credit. The top pros all have 300 APM, and you need it to be fast enough to macro effectively!
The reason I say APM is invalid is again because it is a [i]product of your ACT. Please, go read the "Why APM is useless" section again, and notice how APM can be simplified to a formula where the only variables are in direct correlation to ACT; the top pros all have 300 APM because they process a hell of a lot of tasks in a hella short time (this is particularly true for BW). This requires high hand speed which, when fully utilized, produces APM.
The Fact is, you don't need to have all of the extra translation required and risk for misinterpretation that APM has inherently, when it really tells you nothing except for a players handspeed. Again, even that statistic is likely invalid, because the vast majority of players do not use the entirety of their handspeed for the entirety of the measurable period.
Edit History:
+ Show Spoiler +
