|
@drKobalt:
Interesting results. It is good to see that OL, O9 and O0 have the same travel time. With our system of calculating keyscores, L is considered quite a bit better than 0. So I think we might find some other interesting results, when running this on a bigger scale.
Also interesting is the difference between H and M. I thought they would score more equally, because the pinky is not the fastest finger (at least for me). So I thought they would both score kind of bad.
After seeing those results, I have some questions:
First I would like to know, with which finger you press U (aka T in the left handed layout). The key is normally assigned to the ring finger, but some people press it with the pinky.
Also I am interested in why you chose the median instead of a mean value and how much the measured values deviate from this median.
Correct me, if I am wrong (which is likely, because I do not know much about statistics), but I feel like we are not searching for a median value, rather than the lowest value. You can always have a variation in reaction time and thus be slower than usual, but normally you should not be able to press the button faster than your best. And with TheCore we usually assume optimal play, which would also mean we assume you are always pressing the button with the smallest possible reaction time. But just picking the lowest value could also lead to errors. So maybe we could use something like:
Mean={1/n*sum[ (X)^(1/2) ] }^2
This should put more emphasis on the lower values and less on the higher values. Maybe we even have to go to x^(1/4) instead of x^(1/2).
Thanks for putting time into helping this project.
edit: changed ^-2 to ^1/2 in the equation.
|
should I be using core 1.0 or 2.0 ?
|
WOW. Some next level shit going on right here. I'm so impressed with how far you're going to try and maximise the layout.
To all involved: 
|
This is quite interesting, I'll have to take a look into it and perhaps switch over. I feel like my play is slower than it's supposed to be partially due to improper hotkey setup.
Thanks for this thread.
|
@KDot2: If you never used TheCore, I would advise you to start with 2.0 . AFAIK there are no major changes coming up for 2.0, as those would trigger the beginning of the work at 3.0. And we released 2.0 as an improvement over 1.0, so this is the way to go.
If you are already using 1.0 then you can decide, if you really want to switch. I can tell you from 1st hand experience, that the switch from 1.0 to 2.0 is quicker than the initial learning process, but it still takes some time. I personally enjoyed learning 2.0, but it will not give as much improvement in usability as the swtich from standard to 1.0.
|
I'm a new player and figured i would try this as i am going to do the staircase method.
About setting camera locations. Intuitively Shift+Ctrol + something feels like such an awkward combo. Is it actually supposed to be pressed like it sounds or is it like with alt + ctrl = alt gr?
If it is supposed to be like that, what fingers do you use for such a combo? Left thumb + index finger + middle finger (i am right handed)?
|
![[image loading]](http://i.imgur.com/stLa5Zk.gif)
It's right there in the OP 
Save yourself a headache, and make sure you read every word in the OP and watch all the vids 
|
By all means but that does have ctrl and shift using 1 finger. So i take it i should press two buttons at once with my poor old thumb?
|
Yes. If you wish you can remove the buttons between ALT and CTRL (remove the keycaps) So you can find ALT / CTRL quicker
|
Is there a list of contributors to The Core / The Core Light anywhere per chance?
|
On December 20 2014 22:18 YoTcA wrote:
First I would like to know, with which finger you press U (aka T in the left handed layout). The key is normally assigned to the ring finger, but some people press it with the pinky.
Also I am interested in why you chose the median instead of a mean value and how much the measured values deviate from this median.
Correct me, if I am wrong (which is likely, because I do not know much about statistics), but I feel like we are not searching for a median value, rather than the lowest value. You can always have a variation in reaction time and thus be slower than usual, but normally you should not be able to press the button faster than your best. And with TheCore we usually assume optimal play, which would also mean we assume you are always pressing the button with the smallest possible reaction time. But just picking the lowest value could also lead to errors. So maybe we could use something like:
Mean={1/n*sum[ (X)^(1/2) ] }^2
This should put more emphasis on the lower values and less on the higher values. Maybe we even have to go to x^(1/4) instead of x^(1/2).
Thanks for putting time into helping this project.
edit: changed ^-2 to ^1/2 in the equation.
I press U (T) with the ring finger.
The median is considered more robust for small sample sizes and data that has the potential for noise spikes (like the subject getting distracted and recording an unusually long response time). The median, unlike the mean (average), is dependent on the ordinal values of the data list and not on the actual values of the data. So an unusually long or short value will only displace the median an ordinal value up or down.
In some respect you are right, we are trying to consider the lowest POSSIBLE time. But in reality we face an important issue. If we assume that the data gathered conforms to a normal distribution (Bell curve), which is very likely true, then the bulk of the data will accumulate around the average value. Lowest and higest values will have increasingly lower incidence along each side of the curve. Because of this, the lowest possible time might not even happen in the curse of the testing session. We would be relying on chance and it might happen within some tests and not within others. It is much safer to look at values that happen often: mean or median.
Regarding your scaling, it is possible to apply any such ad hoc weighting function but this is better done after the data analysis, not during. Just because it is an artificial manipulation of data that would be difficult to interpret statistically.
I want to post another chart with response times for the standard SC2 hotkey configuration just to compare it to TheCore.
Also want to make an experimental protocol for testing out all possible (or at least reasonably possible) key combinations, just to have the scores. Maybe new insights will show up.
Finally, I have a couple of questions:
1. Is there any data of building construction incidence? In other words, how often does a particular building/structure is built within a game.
2. How are you assigning the building keys? It is easy to see that you wanted the fastest or easiest finger combination for a pylon, but what about the rest? Is there a list you use of building/structure priorities?
3. Is there a similar criterion for unit training?
4. And spells?
|
Hi, I am a new user of The Core (just downloaded it and tested it 5 minutes)
I would like to Know, How should I Inject with The Core 2.0? I've watched the video from TutorialCentral, but it only works for 1.0 as the camera base doesn't have an alt bind anymore, and I want of course the fastest method.
Thanks for your help guys!
|
On December 23 2014 09:25 Kenjin38 wrote:
Hi, I am a new user of The Core (just downloaded it and tested it 5 minutes)
I would like to Know, How should I Inject with The Core 2.0? I've watched the video from TutorialCentral, but it only works for 1.0 as the camera base doesn't have an alt bind anymore, and I want of course the fastest method.
Thanks for your help guys!
Find your base cam and find your inject hotkey they will be next to eachother for all zerg core 2.0 profiles. Test if you can press both of them at the same time for me it is - and = go in a text document and press both at the same time -=-=-=-=-=-=-=-=-= is something it will look like. You do not need to use any mouse buttons to inject with this method just have your mouse in the middle of the screen. Lastly this method works best when you have same amount of queens and hatches. Keep all queen on a single control group.
|
So i can't do the rapid fire inject method due to not having the proper keyboard to press two buttons alternatively at the same time.
Should i use v.1 instead (v.2 seems to be extremely unpleasant for my fingers for camera injects) or does anyone have the same problem and have rebounded some keys? Let me know!
Edit: I am new to sc2 and zerg, so i think the problem may be just horrible mechanics on my part.
Additional questions: Is it advisable to just bind the base key (for the rapid fire inject) to my mouse side button, so i can push that + inject at the same time?
In the 2013 Rapid fire inject video using thecore, it says i should rebind the AI-something button to also work as my inject button. Is this still the case in v.2? My alternative button for that is already bound (P). (ZRM)
I'm trying to get around all of this by using the backspace method of injecting but i get my queens wandering all over the place.
|
On December 23 2014 10:59 reps)squishy wrote:Show nested quote +On December 23 2014 09:25 Kenjin38 wrote:
Hi, I am a new user of The Core (just downloaded it and tested it 5 minutes)
I would like to Know, How should I Inject with The Core 2.0? I've watched the video from TutorialCentral, but it only works for 1.0 as the camera base doesn't have an alt bind anymore, and I want of course the fastest method.
Thanks for your help guys! Find your base cam and find your inject hotkey they will be next to eachother for all zerg core 2.0 profiles. Test if you can press both of them at the same time for me it is - and = go in a text document and press both at the same time -=-=-=-=-=-=-=-=-= is something it will look like. You do not need to use any mouse buttons to inject with this method just have your mouse in the middle of the screen. Lastly this method works best when you have same amount of queens and hatches. Keep all queen on a single control group.
Damn, I know about this method, but I have a "-=================" kind of keyboard 
I think it's time to buy a new one.
|
@drKobalt: I do not think that the results will show a bell curve. I think it will rather look like some kind of saturation, where most values will be close to the ‘minimum key press time’ and some will be higher. And the farther away from the minimum, the less results will show up. But I agree with you that we should wait for the results until we decide how to evaluate them
I think a reasonable first step to take, would be to test all the building hotkeys for all layouts, as those are always a combination of 2 keys (build basic key + building key), which are always pressed in the same order and are relevant for the game. And with all the different layouts we get many different key combinations already.
Now to answer your questions:
1. Yes there is: https://docs.google.com/spreadsheet/ccc?key=0AtFSjIb2ibJTdDJUT19PTmR1TVhhTklQWm5aVFl3YUE#gid=10
But I do not know, when this data was collected. I think it is from WOL, as there is no data for the HOTS units.
2. Hm, first of all, we try to put the buildings you need more than 1 on a better key. So a fusion core will not go to the home row, but a supply depot will. Second, we try to create connections from build basic to the advanced structures (like barracks and factory on the same key). Overall we are not looking at data for this process, because there are only a few buildings for each race you are building more than once or twice. So it is kind of easy to order them in priority. (Although I am not sure, why the hatch in ZRM is on the home row)
3. Units are most of the time in the order that they appear in the production building. But here goes the same as for buildings. Most of the time you know which to units of a production facility need to go on the home row and the other units do not have that big a difference. For Zerg we went with the massable units first (drones lings roaches) and then with the higher tech units.
4. Here we have rapid fire to take into consideration. Some spell are good for rapid fire some aren’t. You do not want to rapid fire 10 storms on one location by accident, but rapid fire infested terrans is a must have. So we have some ability keys with rapid fire, some without. And the abilities are assigned accordingly.
@Kenjin38: You can always use another inject method. I bind all my queens to 1 CG and jump to each hatch individually via the camera locations. Works like a charm, when you got used to it.
|
Zerg players, how do you use your control groups? I'm still not sure, i'm used to use only 1 for queens, 4 for hatches, 3 for main army and 2 for harass/spells, but now i'm kind of lost.
|
One quick suggestion: I would make the name of the files have the version in them. Yea... I can add it myself but it might help others.
e.g. instead of "TheCore PLM .SC2Hotkeys", call it "TheCore PLM v2.SC2Hotkeys"
|
@pedz the script we use that generates the files reaaally doesn't do well with that unfortunately.
|
Is there a "visualizer" for The Core v2 ?
I'm starting the adventure of moving from core v1 to core v2.
|
|
|
|
|
|
EPT
