|
On September 17 2011 18:06 skiptomylou1231 wrote:Show nested quote +On September 17 2011 18:04 straycat wrote:
How is the Average Unspent Resources calculated? It's available when you click Economy Breakdown
Yes, sorry I should have been more clear, I know that the game calculates it for you. What I mean is more: What does it represent? Is it like
(sum of: (r_i)) / length_of_game_in_seconds
where r_i is the resources held at second i. So that a three second game where you start with 1 mineral and gain 1 mineral per second and spend 1 mineral during second 2 get an AUR of:
(1+1+2)/3 = 1,33 minerals.
Or is it something else?
|
|
|
On September 17 2011 18:05 killerdog wrote:
might have been mentioned already but i dont think you took into account the fact that even at grandmaster level, protoss has to wait for rounds of warpins, or all races have to wait for enough resources to build an expo, or terran/zerg needs to wait for round of marines/larvae injects. there are lots of things stopping them keeping their recources constantly at zero, so mabye if you work out how much a round of production costs (on average) and different points in the game, and then subtract half of that from the average unspent rescource, (mabye modifyin based on game length) then you might be able to exclude the neccesary build up.
That's true, but it also doesn't matter. Like it says in the FAQ, he thought that would make a difference but it apparently doesn't make enough to matter, so the SQ of all 3 races stays even regardless of the specific mechanics.
Great work! This is what happens when a game has a huge fan base :D we end up with some smart, curious, or personally invested people who put a lot of work into something that benefits the whole community! Incredible system, I'll use it every now and then to get an objective view of my macro.
|
i dont see how you came up with the formula. it seems very random. what do you mean with emperic data? what exately did you do?
also this just has to be included in sc2gears! gj.
|
emperic data is data gathered over time and experience without measuring it excactly, for example fire=hot which you found out through empiric studies as child
|
You should exclude mirror matches from the data set otherwise the T workers created will be heavily skewed, as TvT lasts like 4x longer than PvP on average.
|
The most interesting and counter-intuitive piece of data for me was the graph displaying workers built over time per race. What seemed odd to me was the fact that terran players make more workers than protoss, despite protoss having chrono boost. What do you guys think is the reason for this? My guess would be that terrans typically expand sooner than protoss in the beginning but I still think it's odd.
|
On September 17 2011 18:39 oni_link wrote:
emperic data is data gathered over time and experience without measuring it excactly, for example fire=hot which you found out through empiric studies as child
lol yeah. but i dont understand how he creates a formula out of that.
|
On September 17 2011 18:41 harhar! wrote:Show nested quote +On September 17 2011 18:39 oni_link wrote:
emperic data is data gathered over time and experience without measuring it excactly, for example fire=hot which you found out through empiric studies as child lol yeah. but i dont understand how he creates a formula out of that.
must be a stroke of genius. how did newton find gravity? answer: he didnt find it, he measured it.
same happened here
|
nice numbers!
looks realistic, i am in diamond and have exactly 65 SQ
|
100 players of a random masters division makes the masters league graphs pretty inaccurate
|
On September 17 2011 15:15 Knuppe wrote:
Blizzard should add the IdrA-league :D
On a few leagues above him would be the Major league.
|
On September 17 2011 14:53 Azzur wrote:Show nested quote +On September 17 2011 14:29 Crypt wrote:
Sorry if this has already been posted, but there is an aspect that significantly exaggerates this data.
That is: lower skilled players stay in games longer after they've already lost. This goes beyond things like command center floating and building hiding, which I would believe happens more as the skill level is lower. The player in the position of winning stop or slows creation of workers and spending, since trying hard isn't as necessary any longer. This situation is much more common in lower league play and is not a determinant of their win/loss performance and should not be viewed as a skill-based cause of their league placement. Idra is the perfect example of the opposite extreme. There is very little "dead time" near the end of his games. If he's lost, he leaves immediately without slowing down macro rather than dragging out the game and lowering his macro/time "score."
Result: The gaps between leagues in all presented graphs is exaggerated. Nope, I don't buy this. I believe there are definite gaps between leagues. The empirical data supports what many higher level players know intuitively. Of course there are gaps. I was in no way denying that. I'm saying the gaps are not practically as large as they appear because lower league games tend to have more dead time after games have really already been decided. But, whatever...hard to feel that this will be read among 30+ pages of comments.
|
Awesome posting!
However, I think it would be even greater, someone would be able to derive some potential fields of improvement for their own playstyle.
Therefore I gathered also the data about my opponents and matched their SQ with my own and whether the game resultet in a win or loss.
(I know that having only 20 samples is a very weak point, I would like to discuss it nonetheless.)
Leaque Platinum (nearly all opponents as well)
Win ratio for the samples: 40%
Avg Game length: 21.85 min
My average SQ: 62.26 (+- 7.21)
Opponents avg SQ: 58.20 (+-12.73)
My SQ was higher in 65% of the games than my opponents one
In 60% of the games, whoever had the higher SQ, did win the game
My own conclusion from this data would be, that dispite often having a higher SQ than my opponents I lost 60% of the games. It seems, that I am often not able to make my makro advantage the game winning factor.
Therefore, I guess the conclusion can be that I loose to other factors like decision making, micro, harrassement.
Do you think it is possible to make these kind of conclusions from this kind of data? Or is it too much interpretation with too weak data (too small sample)?
Regards
|
Great work 
|
76.77 in the last 7 games, come at me brahs!
|
Isn't it interesting that in GM toss is the race that produces the lowest ammout of workers? Shouldn't terrans be there, considering they got MULEs?
|
On September 17 2011 19:31 Shewklad wrote:
Isn't it interesting that in GM toss is the race that produces the lowest ammout of workers? Shouldn't terrans be there, considering they got MULEs?
maybe thats the reason toss suck in winning rates nowadays :D
|
So... if Idra is higher than the average GM, does BM > GM?
|
One of the best posts this year. Can't wait to calculate how shitty my macro is!
|
|
|
|
|
|
EPT
