|
In the OP, 1. step of TheStaircase currently have some bracket problems. I'll colour the brackets so you can see what I mean: "Step 1: Play a melee game* of Starcraft 2 (Mine only Minerals. (no gas) Attack Move on the minimap only. (do not look at battles)" The red one is missing. I think only using brackets for "no gas" would make it more easily understood.
Also, it says a melee game. I'm quite convinced you mean that the "4 out of 5 to pass" rule applies to all steps. 
Another note on the OP: All titles do not have equal space above and below them. For example, Getting Started have a lot of space below it.
Oh, and, finding the Spreadsheet to check is pretty hard now. Not that it's going to used all that much anymore, but a big, shiny link somewhere in the post would be nice ^^
About given units, I'm curious as to why Terran have Hellion/Hellbat. My guess is that it's because they look like Warp Prisms in the sense that they're gas free, but requires gas tech - and that they fit well into both bio and mech compositions, and thus will be a preferred choice of many Terrans ascending TheStaircase.
But, they're still very different from every other given unit. Besides Mothership/Core they're the only given that can attack, but cannot be built in step 1. The Mothership/Core, however, can only be build once, so as a chosen unit it wouldn't feel like it could contribute as much as most other units. Oh well, I suppose I can just opt not to make them 
|
i need some help with this... i registered and everything, chose the staircase view, but i cannot find the option of how to set my League Goal. Where is this option located? can you please advise?
also, what does it mean to play a melee game??
pls help.
|
hey hey!
ggtracker isn't tracking your progress on TheStaircase just yet. Right now it will just be your source for measuring how well you did each game. League goal just means that to have a pass for a game, you have to have that league badge or higher across the board for spending skill and saturation speed.
For a game to count for TheStaircase you must play a 1v1 melee game. Melee games mean custom games against friends or AI, VS ai ladder, unranked ladder, or ranked ladder games.
Let me know if you need anything else
|
Just a question for your saturation problem.
The amount of geysirs should be in the replay, aren´t they?
Whould it be possible for GGtracker to take the amount of them into account for optimal saturation? In that case the "instant" saturation of the next base would not be an issue anymore?
Just don´t use fix numbers for saturation.
Try an approach like this:
Base 1 saturation = 672 + (number of geysirs *100)
Base 2 saturation = 1344 + (number of geysirs *100)
Base 3 saturation = 2016 + (number of geysirs *100)
In this case every strat would work as it should, while the benchmark changes as the strat changes.
|
That is an interesting idea. We have the birth/death times, as well as gas/mineral income. I wonder if this would work. I will ask dsjoerg.
|
Another thing:
Right now, you are measuring the time from completing a Nexus/ Hatchery / CC until saturation.
Does this take macrohatch and CC on the highground into consideration? Do you know, if these buildings are placed at the right location?
If not, why don't you just measure the time between 1 base and 2 base saturation? Without taking the time, when the second CC/ Hatch/ Nexus was finished?
As it is right now, you could build 32 Workers. Then finish your CC. And get an Instant saturation. Thats not, what you want i think.
But you can't trick numbers. And i think, that these numbers would better reflect, if the players forget their macro while playing.
A 1-Base player focusing on his Army, after 1-base saturation... And after 12 minutes building his second CC, then saturating it in GM Level would still have a bad macro.
Do you see my point?
|
the structure must be in mining position for it to count.
measuring the time between 1 and 2 base saturation would definitely be an option to test. I still need someone that has python knowledge and can work with Graylin Kim's Sc2 reader library to test these things before we can incorportate them!
workers are a means to an end, that end is income. If we measure workers we only get a piece of the puzzle. Ultimately the goal is to generate income, no use having a bunch of workers if income isn't being generated.
"A 1-Base player focusing on his Army, after 1-base saturation... And after 12 minutes building his second CC, then saturating it in GM Level would still have a bad macro."
bad macro is a very much undefined term at this point in Sc2. if that player were going for an all-in build, did enough damage to stabilize and not enough to kill, then carried out the game normally I wouldn't say had bad macro based on that information alone.
|
But this starting player would analyse his game that way.
I would not look at the 2-base saturation timing in this example, if I do a 1-base all in. It's irrelevant, cause I did not plan to take a second base initially.
But if a starting player, for which your staircase is made for, looks at his saturation speed (2 times Master for the sake of it) with my example above. This player would ask why am I so bad? My spending skill and saturation speed is master level.
The answer would be( in a vacuum): He expanded too late, but when he did it, he did it very fast on a master level. But this is not enough. You have to expand earlier to get your whole macro going.
And I think that this is a benchmark, you are missing right now. When do I have to expand, to be a real master? The time from completing your second main structure to its saturation is a side effect of this.
|
Expansion timing is based on the individual game and situation. How do we make a benchmark for this that is consistently meaningful?
|
This is also the case with the time you need for the saturation of the 2nd and 3rd.
Early pushes/ Cheese will affect the second.
Timings might hit the third while saturating.
But that does not change the fact, that you just have to expand as early as possible without being too greedy.
|
"expand as early as possible without being too greedy" is difficult for a computer to interpret.
I think it is true that having an expansion without saturating it is inefficient play. It follows that the longer it takes to saturate, the more risk you are taking, and the less efficiently you are playing. If we could measure oversaturation (16-24 as well as 24+) that would be really fucking cool, but I haven't come up with anything yet.
|
To find out when pros expand, just watch a few hundred vods and you know, what a fast third is. Or the casters will tell u. But how fast do Gold Players oder Silver Players expand? How fast are they saturated on three bases? These are the things i want to know as a beginner. Am i good with my saturation speed on three bases? Am i too slow?
Attacks or harass will put me behind. But these are the lessons, all new players have to learn. Deal with this attack, defend well and get your three bases going. If you are hit by early attacks very often and lose because of this, build more army, delay your second/ third... but try to not delay it by much.
Just take a look at the terran benchmark for completing a third and getting the income as it is right now. 53 seconds? 28 Seconds? What is this? The player, who is the first to hit his mules wins? Not being rude here. Just a question.
I am a newb. Maybe getting promoted to gold, maybe not. And i want to use your system. But telling me how fast workers are send to an existing building is not a benchmark i would look for. (Queue workers, rally them on the mineral field, not more than 2 at a time, being within the rules, just remember to hit 4 (CC hotkey), maybe missing 1-2 seconds per worker, profit)
|
On July 09 2013 02:23 JaKaTaK wrote:
"expand as early as possible without being too greedy" is difficult for a computer to interpret.
I think it is true that having an expansion without saturating it is inefficient play. It follows that the longer it takes to saturate, the more risk you are taking, and the less efficiently you are playing. If we could measure oversaturation (16-24 as well as 24+) that would be really fucking cool, but I haven't come up with anything yet.
1. You don´t need a computer for this. Do it just like as it is right now. With slight differences.
Don´t measure the time between finishing the CC and saturation. Measure the time between the different income levels. It is in the replays. The players will tell you. And if you do it this way, you won´t have someone cheating your system, as u try to find out in your spreadsheet right now.
2. You could theoretically. Just count the workers active. If there is only 1 CC, but 32 workers were already created, you would have an oversaturation.
|
"But these are the lessons, all new players have to learn."
I feel like you are speaking from a point of authority, but don't have evidence to back up what you are saying. Again, I need someone who knows python to help me analyze possibilities for metrics, we can't just keep changing things in GGTracker to see how they work. We have to be able to prove its better than what we have now with data. I like all of your ideas, and am pumped to try them out asap
|
On July 09 2013 03:51 JaKaTaK wrote:"But these are the lessons, all new players have to learn." I feel like you are speaking from a point of authority, but don't have evidence to back up what you are saying. Again, I need someone who knows python to help me analyze possibilities for metrics, we can't just keep changing things in GGTracker to see how they work. We have to be able to prove its better than what we have now with data. I like all of your ideas, and am pumped to try them out asap
Life tells me not to try to reach everything at once. Starcraft tells me not to take a third without any army to defend it, if i am not able to scout, what my opponent is doing and how big his army is... :D
My only evidence would be the hundreds of games i watched the last months. Even Pros being too greedy have this lesson to learn.
And i really do not want to sound authoritative. I just like your approach and want to help if i can. And try it on my own.
|
I didn't think your intention was to sound authoritative. Glad to see I guessed correctly The search continues for a person with python experience that can help up test these ideas. Once we can do that, we'll be able to see which ideas show more separation between leagues and therefore are more likely to be indicative of skill or success.
|
On July 09 2013 05:23 JaKaTaK wrote:I didn't think your intention was to sound authoritative. Glad to see I guessed correctly The search continues for a person with python experience that can help up test these ideas. Once we can do that, we'll be able to see which ideas show more separation between leagues and therefore are more likely to be indicative of skill or success.
Jak, I know a little bit of Python, not sure how much use I can be but I can take a look.
|
@teuthida
what is going to happen is we'll get a huge dump of replays, and then we will analyze them en mass. I'll give you parameters, and then we see what happens when we apply those parameters to the games. Namely which leagues get what scores on average and the separation between them. Do you think you can write something that will do that?
|
On July 09 2013 07:28 JaKaTaK wrote:
@teuthida
what is going to happen is we'll get a huge dump of replays, and then we will analyze them en mass. I'll give you parameters, and then we see what happens when we apply those parameters to the games. Namely which leagues get what scores on average and the separation between them. Do you think you can write something that will do that?
Jak, I think so, I'll give it a shot anyway.
|
Sweet, I hit up dsjoerg, send me your info via PM :D
|
|
|
|
|
|
EPT
