EPT[G] APM Vs. Ability Completion Time - Page 3
| Forum Index > StarCraft 2 Strategy |
|
Sotamursu
Finland612 Posts
| ||
|
DarQraven
Netherlands553 Posts
APM is not useless. There's a good chance that a player with high APM also has a lower ACT. The problem is that it is not a measure that actually allows a meaningful comparison between players. Two reasons, as detailed in the OP: 1) It's a long-term average. If one player has insanely low ACT and does stuff at twice the speed other players do, that's going to leave him with extra time in which there is relatively little to do. If that player doesn't spam during that time, his APM is going to be exactly the same as the other players, even though his reactions were twice as fast. Starcraft is not a game that goes faster as the player gets faster: factors like building time, move speed, DPS, etc all bottleneck the speed of events, and at a certain point there simply isn't anything useful to inbetween the moments when an action is needed. That's when players need to start spamming to have APM reflect their actual speed. 2) It doesn't actually relate to quality of gameplay all that much. Watch VODs of high APM players and see how, without fail, most of them give each move order between 3-8 times, even when the location of that move order is almost identical. Bam, APM is now 3 to 8 times higher than the actual "useful apm" (making the army move, only 1 action). On a similar note, I've seen players with lowish APM react immediately to a drop, while some people are hovering around the 200APM mark and react 2 seconds late. Which one is the better player and which one does APM suggest is the better player? By factoring in mental processing time instead of just raw keystrokes, you get a much more useful measure of a player's speed than APM gives you. I can get an almost pro-level APM just by mashing keys all over the place while playing cookiecutter builds, but I'm fairly sure I don't have nearly the same level of control and awareness they do. On August 01 2011 20:24 Sotamursu wrote: I read the OP and glanced through the replies. How exactly would you go about measuring ACT? I'm not sure it's possible from within SC2 itself, because the game would have to know which events you find relevant and what your reaction to them has to be. As far as I understand the measure, it comes down to the time between you being notified of something important, and you having completed your reaction to that event. For instance, the time between the "probe complete" sound playing and you having queued up another probe. However, you'd have to measure a series of these events as well, since it's fairly easy to prime yourself for a certain event and then react quickly. You can't do the same if you sequentially get "probe complete" sound, "construct additional pylons", and a drop in your base, especially if you don't know the order or timing they're gonna come in. | ||
|
NASAmoose
United States231 Posts
My question would be--how do you go about really practicing this? I think I'm at the point in my gameplay where sitting down and making an SCV every 17 seconds isn't really going to help me (in fact, I'm overproducing them of late). So what are specific ways to get your ACT higher? | ||
|
Micket
United Kingdom2163 Posts
| ||
|
Emperor_Earth
United States824 Posts
On August 01 2011 22:36 Micket wrote: But cycling through your control groups is perhaps the foundation of good macro. You aren't doing any ability completion there, you just going through your mental checklist to make sure you are remembering to do everything. So APM is still very important and not a useless stat. Edit: Took out super long rant. Micket has grasped part of it. APM is a great indicator because of how many things actually need to be done/checked up on. There's so many things that need to be done that you will get to the point that there's no delay between actions. Once you understand the game fully enough. It's not about knowing what to do. It's about prioritizing it. And seeing how many things you can check off that priority list. Hence APM will invertly scale with ACT. Also. Some actions are far more important than others but not have the same ACT and will affect APM the same way. ACT is truly only a worthwhile stat if you find yourself staring at the screen spamming because you're clueless and don't know what to do next. TLDR: Until you actually know how to play or at least commentate, don't argue with a pro like Red[NaDa] and think that APM is not as viable a stat as your newfangled "ACT". | ||
|
Sina92
Sweden1303 Posts
| ||
|
TheKRoc
United States74 Posts
On August 01 2011 16:50 Macpo wrote: I am a bit skeptical about the demonstration on why APM is not an adequate indicator... It just says it is not good because it doesn't measure your "true speed". Of course it does! it just counts the number of actions per minut, nothing deep or strange there... if you can give 300 orders per minut and control what you are doing, then you are going fast. When you are driving a car, speed is in miles or kilometers per hour. When you're in front of your computer, it's actions per minut... Then of course, high apm doesn't mean you play the most efficient way, in the same way that you can drive a car faster and yet take more time to get to a certain point, because you made a detour. But that doesn't mean your car wasn't going fast... APM measures a players speed in the context of one game. It doesn't measure his maximum speed by any means, because for a large portion of the game players do not utilize their full speed. Because we don't know how far below a player is below their maximum, or how that relates to other player's maximum when using APM, we cannot tell how fast a player truly is by using APM. Please read the OP in it's entirety. | ||
|
TheKRoc
United States74 Posts
On August 01 2011 17:19 elgranbasio wrote: I sympathize with you about all the people who blindly comment about APM on here without reading your detailed/well thought out post. Stay strong, and keep coming up with well presented/organized posts like these :D Thanks so much! | ||
|
TheKRoc
United States74 Posts
On August 01 2011 18:51 JustAGame wrote: You could sum up the full OP with: 1.) APM includes spam, and therefore doesnt measure your speed. 2.) To get faster you need to practice the action you want to be faster at. 3.)+ Show Spoiler + spoilers for the win This, fortunately enough, isn't accurate. I actually explain a whole lot of things other than those in the OP, such as: Helping to clarify the reasoning behind the invalidity of APM. Several key technical terms relating to handspeed and the quickness with which a player can complete any given objective. Providing clear guidelines and hopefully fresh advice on how to best increase that same quickness Introduced the concept of mental isolation to a community previously unfamiliar with it, that deals with it a whole lot. etc. | ||
|
Macpo
453 Posts
On August 02 2011 01:32 bonerificus wrote: APM measures a players speed in the context of one game. ... Please read the OP in it's entirety. Well, no offense, but I don't really see the point of measuring one's speed outside the game, as if it were its "true" speed... You just kind of build some abstract thing that never applies and therefore doesn't help much. it's a bit like religious sects telling you you are only using 10% of your brain capacities. You could also open paint and see how many dots you can make in one minut with your mouse... And also, please, avoid remarks like "read the OP in its entirety" blablabla. people may know how to read and still not agree  With no intention to be a hater , Macpo | ||
|
TheKRoc
United States74 Posts
On August 01 2011 19:06 Herculix wrote: well before i reply, i just wanna say that your whole concept is stupid because you're saying APM should be replaced by this alleged measurement in which there is no legitimate way to calculate. WTF? if you don't care about APM, just ignore it, it's not a huge deal, but this is pretty much useless theorycrafting. I don't think APM should be replaced by ACT, and nor is it ever said in the OP. The fact of the matter is, ACT is a real, accurate, statistic (whether or not we can consistenty measure it); if you understand it, working towards improving it will give you a tangible improvement in-game. random bronze zerg player has 40 total APM (200+ at one point if he spams), but at any given passive moment during the game (after 10 minutes), he only has 20-40 current APM during passive moments, and 60-100 during active situations where all the tasks he assigns himself are called upon to be completed simultaneously. the tasks that he assigns himself are building units, rallying, managing queens, spreading creep, scouting, positioning overlords, army movement, and developing tech when necessary. idra has 200 total APM, but at any given passive moment during the game (after 10 minutes), he has 150-200 curent APM, and 200-300+ during active situations. idra assigns himself the same exact tasks as the bronze player. both of these are averages taken from many games i've watched both of idra and of random bronze players. the bronze player is able to complete the tasks he's given to some extent. you asked how to make yourself smarter, i.e. improve your APM? do what random bronze player does, but do it consistently, accurately, and quickly, like idra does. that's what APM shows. your math formula is poor and based off of that i would assume your understanding of APM in general is incredibly poor. What you say here doesn't even make sense... did you read the post? In the OP, I mention how an increase in ACT on a highly represented action will reflect in increased APM, until about 250 APM. The "random bronze zerg" you mention A. does not have anywhere near the same number of processed mental tasks as IdrA (a professional player has much smaller, more precise guidelines, like "send an overlord to the left side of his ramp" rather than "send an overlord to his base"), and B. cannot physically complete all of those tasks, because he takes too long and the oldest uncompleted tasks fall off the memory queue (also mentioned in OP). Then, you claim APM shows how consistent, accurate, and quick you are? This is unfortunately just not true, and I'm tired of answering similar questions, so please go back and read the OP if you are wondering the validity of APM as a measure of speed and... consistency.. people who talk about APM being useless usually only talk about total APM because they're stupid and they're stupid because they know total APM is easily inflated, yet they try to act like people arguing that APM matters are actually referring to total APM. actually good players demonstrate how good their mechanical skills are with their current APM which you can check sporadically during different paces of the game and different scenarios of the game (i.e. managing 1 base vs 2 vs 3, etc). I've also answered this complaint a ton. The reason a player who is consistently faster than another player has higher APM is because they complete key tasks faster than the other player. If you have faster ACTs, then you will complete more actions total (at least in a complex game like SC2), and therefore have a higher in-game APM. APM is a byproduct of fast ACTs, explained in OP... it is not a 100% accurate indicator of any specific skill, but it is a solid general indicator of a player who is potentially weak in an area, and conversely that a player is competent in all the necessary areas to play high level SCBW/2. usually once you develop most of your skills, your APM will reach an "average current" APM of a certain level depending on your race, usually 130-200+ during passive play at the current stage of SC2, though Nada has a similar amount of APM in BW as he does in SC2 for example. some players inflate this by performing actions with extreme excess even well into mid-game like ViBE, but most don't. strong players who have low APM are usually notoriously poor multi-taskers, including players like Axslav, Minigun and Sjow. they cannot handle the tasks which get your APM high, such as complex army movement. they often crumble when they are not setting the pace of the game and they are being harassed in multiple locations or having to deal with an army with a diverse set of threats happening on separate parts of the minimap. they also lack the multi-tasking to harass multiple locations at once if they have to deal with a split army that can potentially defend both threats. The first paragraph is fine, but again, APM is a byproduct of ACT so we can easily make that conclusion. You still don't show the worth in APM over ACT here. This is a false conclusion. Just because you can name three players that have problems multitasking and also have low APM, does not mean that their is an actual correlation therein. White-ra has like 107 APM, and he multitasks as much as SeleCT. | ||
|
TheKRoc
United States74 Posts
On August 02 2011 01:47 Macpo wrote: Well, no offense, but I don't really see the point of measuring one's speed outside the game, as if it were its "true" speed... You just kind of build some abstract thing that never applies and therefore doesn't help much. it's a bit like religious sects telling you you are only using 10% of your brain capacities. You could also open paint and see how many dots you can make in one minut with your mouse... And also, please, avoid remarks like "read the OP in its entirety" blablabla. people may know how to read and still not agree With no intention to be a hater , Macpo No, This isn't even close to correct! I don't think ACT is something you should ever measure, as you don't need to. if you know what it is and seek to improve it, then you will see a direct correlation in-game regarding the speed at which you macro. I cover this in the OP............ | ||
|
TheKRoc
United States74 Posts
On August 01 2011 22:36 Micket wrote: But cycling through your control groups is perhaps the foundation of good macro. You aren't doing any ability completion there, you just going through your mental checklist to make sure you are remembering to do everything. So APM is still very important and not a useless stat. Yes, you are. Every time you click a hotkey, a full task is initiated and completed. Please remember, as I said in OP: EVERYTHING you do in the game of Starcraft is defined in actions. Not for one second are you not cycling through action completion. Even spam is 100% action completion; to explain, I will put an example in terms of the OP: When you're not doing anything in the early portion of the game, your brain recognises this lack of input as a stimuli. It formulates a physical instruction list, which is likely to spam a few keys over and over, and sends it to your fingers for completion. When you're macroing in the midgame, your brain's internal timer will signal a specific portion of the cerebellum, which then formulates an physical instruction list (in this case, probably clicking a few hotkeys to check your production) and sends it off to your fingers. | ||
|
Macpo
453 Posts
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 Now, speed is a relation between a quantity and a time (like kilometers per hour, or numbers of pages written per day, actions per minut, whatever) right? So, what is the quantity and what is the time you are using to elaborate your new speed indicator? (because you claim to give a metric, not only "ideas") | ||
|
TheKRoc
United States74 Posts
On August 01 2011 20:24 Sotamursu wrote: I read the OP and glanced through the replies. How exactly would you go about measuring ACT? Well, it's really hard and you don't need to When someone really feels a desire to measure their ACT, a psychologist can administer a test that presents a few hopefully apparent stimuli and provoke a predictable, consistent response. This is often something along the lines of "write "The brown fox jumps over the lazy dog."" It's iffy though, for a whole lot of really complex reasons. For one, it's hard to isolate the ACT from other components of the thought process (most of which are not present in an instinctive game like Starcraft) that occur in the same timeframe. It's also nearly impossible to isolate MCT and PCT, which is the real desired statistic in the medical field. Normally, it's used to diagnose a few rare post-processing disorders. Feel free to PM me if you have any other questions :> | ||
|
TheKRoc
United States74 Posts
On August 02 2011 02:04 Macpo wrote: Well, If I am correct, what you want to do is to introduce a way to evaluate speed. Let me quote you, so that we agree at least on that. this is about the first sentence of the post. Now, speed is a relation between a quantity and a time (like kilometers per hour, or numbers of pages written per day, actions per minut, whatever) right? So, what is the quantity and what is the time you are using to elaborate your new speed indicator? (because you claim to give a metric, not only "ideas") That's really not what I mean in the quote. The purpose of the OP is less to replace APM and more to provide a statistic that is relevant to actually improving your macro speed. I'm going to go back and edit that, because I think it's a little bit misleading. You shouldn't ever have to evaluate anything with ACT; if it quickens in many key tasks, the improvement will be apparent in other areas (like floating money). If you want the actual methodology for measuring ACT in the psychological field, I can give that to you as well. Normally, a patient is given a clear, hopefully easily comprehensible stimuli, and is timed on how long it takes him/her to process and complete relevant PMTs (processed mental tasks). This is often "Write "The quick brown fox jumps over the lazy dog."". The test is repeated with nearly identical stimuli 5 to 10 times, and then averaged. The final number is measured in seconds, using the form (completed actions/second). This test is generally used to diagnose rare post-processing disorders in the medical field. | ||
|
Macpo
453 Posts
It may clarify things a bit (which are still pretty confusing at least in my head ) | ||
|
integral
United States3156 Posts
I can think of several isolated tasks that could be practiced separately. Many of these already exist as custom micro maps. - injecting larva for all of your hatcheries - creep spread using creep tumors - macro cycles with all races - splitting and loading an army into dropships - spreading marines (such as vs banelings) etc. It also seems like there is going to be a significant correlation between typing speed and ACT for e-Sports, so a player could work on increasing their hand/typing coordination/speed outside of SC2 as well. Mouse precision, speed, and accuracy could also be worked on and measured outside of SC2. | ||
|
TheGreenMachine
United States730 Posts
On August 02 2011 02:31 integral wrote: Let's say you were trying to come up with a standardized test similar to the one for measuring ACT in psychology: what tasks in SC2 would be easily quantifiable? I can think of several isolated tasks that could be practiced separately. Many of these already exist as custom micro maps. - injecting larva for all of your hatcheries - creep spread using creep tumors - macro cycles with all races - splitting and loading an army into dropships - spreading marines (such as vs banelings) etc. It also seems like there is going to be a significant correlation between typing speed and ACT for e-Sports, so a player could work on increasing their hand/typing coordination/speed outside of SC2 as well. Mouse precision, speed, and accuracy could also be worked on and measured outside of SC2. Yeah this sounds way more useful than some hypothetical ACT thing. Also add the speed at which you do these things splitting army against drops and adjusting army hotkeys using initial lings to scout for proxies The average speed at which people do these things+ things listed above = a better measure of the skill of a player in most circumstances. | ||
|
TheKRoc
United States74 Posts
On August 01 2011 22:30 NASAmoose wrote: Really interesting original post, thank you sir. My question would be--how do you go about really practicing this? I think I'm at the point in my gameplay where sitting down and making an SCV every 17 seconds isn't really going to help me (in fact, I'm overproducing them of late). So what are specific ways to get your ACT higher? If you feel you have basic building blocks down pat, try doing them on a bigger scale For instance, if you feel comfortable making SCVs, Marines out of three rax, checking your control groups, dropping mules, and building supply depots consistently, try putting the whole thing on a larger scale. Load up a custom game, and keep your money low on 7 bases and pure mineral tech Doing things like this force your brain to do be highly familiar with repetitive actions, because you're going to push yourself really hard to stay low on money from an income higher than you're used to (I hope!). You can also try to reduce mental processing time, by having a friend do one out of three or five predetermined 6 minute timing pushes in a custom game. Scout at the same, consistent time, and push yourself to make necessary preparations faster than you have in the past. Like I said in the OP, there is a point where your ACT gets low enough that you end up processing a fuckton of actions - at this point, you're going to see your handspeed increase substantially. The best way to do this is those multitasking UMS - they all do a great job of forcing a shitload of stimuli onto you, and making you react quickly enough to do them all. Alternatively, you could try to micro two or three dropships around in a specific pattern (maybe a ven diagramm), and macro at home. The thing you really want to aim for when your keystrokes are accurate, precise, and quick is a consistent strech on the speed at which those strokes occur. Thanks so much for the question and the kind words! | ||
| ||
