|
I believe it was Artosis who said that APM measure by SC2 is around 30% lower than in BW. Hearing that, I figured that maybe it was because SC2 doesn't require as much APM to be successful as is needed for Brood War pros, and I left it at that.
Then I started to wonder just how fast I could click, and from there I discovered exactly why APM is ~30% lower, and it's actually not the player's faults.
The questions to ask.
1) How do I produce a perfectly reproducible test bed for APM?
2) Does computer performance impact APM readings?
3) Does internet connectivity (latency) impact APM readings?
4) Does game speed impact APM readings?
1) To make this test as accurate as possible, I used two keyboard macros on my Razer Blackwidow.
To get the maximum APM possible, the first I used a macro that spams 1, 2, 1, 2 as long as I hold the button down (1 and 2 are control-grouped workers), with 0 delay between presses. My keyboard has a polling rate of 1000Hz, so the maximum APM it can output is 60,000 per minute, or 1000 per second.
The second macro was set to put a 0.2 second delay between every keystroke, giving the player exactly 300APM.
Below is a thumbnail and a link to the full-size image of the 'Infinite APM' macro:
![[image loading]](http://i.imgur.com/Hz8T8l.jpg)
http://imgur.com/Hz8T8
2) Below are two tests using the "Infinite" APM macro key. The test is running on Q6600 with GTX260 at 1920x1200 on high and 1024x768 on low. The reason I'm using such extreme APM for these tests is that at higher APM, the difference in APM for each settings should be more obvious.
High Settings, Slowest Game Speed: 46488 APM
![[image loading]](http://i.imgur.com/OICxgl.jpg)
http://imgur.com/OICxg
Low Settings, Slowest Game Speed: 50256 APM
![[image loading]](http://i.imgur.com/syY2Ol.jpg)
http://imgur.com/syY2O
What can be concluded here is that while computer performance may have an impact on APM, the difference is only slight, and is irrelevant at more natural APM.
3) For this test I compared the Low Settings Test above to a test done on single-player.
Low Settings/Online, Slowest Game Speed: 50256 APM
http://imgur.com/syY2O
Low Settings/Offline, Slowest Game Speed: 50940 APM
![[image loading]](http://i.imgur.com/qGilbl.jpg)
http://imgur.com/qGilb
Internet connection also appears to have a negligible impact on APM readings. Lets move on.
4) Having eliminated variables, we now have the final experiment, and the part that will probably make some people angry. The following are results from the 300 APM timing experiments:
Low Settings, Normal Game Speed (which is equal to real time):
![[image loading]](http://i.imgur.com/MOl0jl.jpg)
http://imgur.com/MOl0j
Low Settings, Faster Game Speed (what we all play at):
![[image loading]](http://i.imgur.com/Xs1rkl.jpg)
http://imgur.com/Xs1rk
Conclusions: Why people should be annoyed at Blizzard about this.
Test 4 demonstrates that APM is calculated relative to GAME time instead of REAL time. In the Broodwar days, 3rd party APM calculators read APM in real-time.
If we take the 216 APM of 'Faster Speed' and divide by the 300 APM of 'Normal Speed', we get:
216 / 300 = .72, meaning SC2 APM = 72% of your actual APM, which is nearly a 30% decrease in APM, hence why SC2 players are 30% slower.
I would hope that someone at Blizzard will read this post and change the APM system to measure APM in real-time instead of game time. Means those 300 APM players should actually have 417 APM.
Thanks for reading!
-Veratule
|
I dont want to be rude, but istn this common knowledge ?
But nice test anyways 
|
On March 24 2011 06:01 RmpL wrote:I dont want to be rude, but istn this common knowledge ? But nice test anyways
I actually didn't know if it was common or not. Either way I think Blizz should fix it.
|
The reason why its lower, is because we play at Faster modus, which is the normal speeld x1.38. But yeah, its kinda common knowledge but thanks for the efford 
|
Good job OP, I was unaware of this.
Seems silly to have the M in APM not stand for "minute" but "blizzard minute." I just assumed the were using real units of time.
|
This has been known since beta, where do you think Artosis got the 30% figure from?
Anway, it's a nice experiment you did, and I definitely agree that APM should be gauged on fastest game speed.
|
I think what you've found is that because time is accelerated in the faster game mode, APM calculated based on game time is bound to be less, but their realtime APM will be the same as it's always been (as we see, a gosu gosu 300); Not because they are slower, but just SC2 players are being measured by differing time scales.
|
Cool test, but this was known since March or April of last year (search function). Your actual APM is 1.33 times faster than the displayed APM.
|
Yea this is common knowledge since the beta. Program like SC2gears can also convert your game APM to real APM. Or you can multiplie your in game APM by 1.398 .
|
People have been bringing this up since the beta. Blizzard hasn't done anything so far, so I wouldn't expect it to change for a while, if at all.
|
On March 24 2011 06:04 TrickyGilligan wrote:
Good job OP, I was unaware of this.
Seems silly to have the M in APM not stand for "minute" but "blizzard minute." I just assumed the were using real units of time.
Thanks
Yeah, that's exactly what I'm saying. Doesn't make sense.
I didnt post it, but that test with 50940 APM, I did the same test on faster instead of slower and got around 21000, and normal was just under 30000.
|
United States12248 Posts
Yeah this was already known. Faster speed runs at 1.3999x speed, and so your APM is measured at that speed. Your reported APM is equal to 1/1.3999 of your actual APM. This has been known since beta.
|
Yep, this was known throughout beta. I think it's better this way. It's good to have a separation between Game Seconds and Real Seconds. Otherwise, everything gets confusing.
|
The problem is that everything in the game is based on the "blizzard minute". So do you have APM completely unrelated to the clock time and build times? Or do you readjust everything, which would completely ruin all the even 30/40/50 ect. build times.
|
1 - Everyone knows this already.
2 - It doesn't really matter. I mean unless you are just trying to feel super good about yourself and your 80 APM is actually 110 or whatever, we are usually comparing APM's with players being measured by the same system.
So what we know - Anything 150-200 is high. Anything about 200 is really high(at least in my opinion).
If you compare 2 players and one has an APM of 150 and the other 200, the one with 200 is doing more APM. If it is actually 200 and 250 in real time it doesn't change the fact that the player at 200 is doing more and that is really all you need to know.
So does it really matter if it is exact real time or not? As long as the point of comparison is the same we get everything we need. This to me is a non-issue and something everyone already knows and is brought up often when talking about SC2 APM. Just don't compare 3rd party APM counters from BW(real time) with SC2(game time) APM because you are then comparing 2 different measurements at that point, again something we all know which is why people don't do it.
|
It was kinda established that APM was approximately 30% lower, but I haven't actually seen anyone do an in depth test regarding it, so good job on that.
|
everyone knew this but it is still a good demonstration and good op, it makes no sense to show a decreased apm.
|
Known to people or not, i have always believed Blizzard implemented APM in a wrong way. If a game is played at "Faster" speed, APM be interpreted in terms of Faster speed, not Normal. Calculating apm based on a magic number called ingame minutes is just plain dumb. This shouldnt be too much of a fix, and would reflect the true meaning of APM.
|
On March 24 2011 06:07 FLuE wrote:
1 - Everyone knows this already.
2 - It doesn't really matter. I mean unless you are just trying to feel super good about yourself and your 80 APM is actually 110 or whatever, we are usually comparing APM's with players being measured by the same system.
So what we know - Anything 150-200 is high. Anything about 200 is really high(at least in my opinion).
If you compare 2 players and one has an APM of 150 and the other 200, the one with 200 is doing more APM. If it is actually 200 and 250 in real time it doesn't change the fact that the player at 200 is doing more and that is really all you need to know.
So does it really matter if it is exact real time or not? As long as the point of comparison is the same we get everything we need. This to me is a non-issue and something everyone already knows and is brought up often when talking about SC2 APM. Just don't compare 3rd party APM counters from BW(real time) with SC2(game time) APM because you are then comparing 2 different measurements at that point, again something we all know which is why people don't do it.
I understand that it's all relative, I agree with your point, but I would rather the APM be right instead of relative.
|
So if my apm is 140... that means its really 200? Cool. I'll take it 
|
|
|
|
|
|
EPT
