[EveVendetta]

Hi guys, I came across this thread on team liquid which gave a very good analysis of the macro abilities of different leagues.

I think the measuring of your Spending Quotient would be a very good tool in seeing what your standard is, and it can be used as well as a gauge for whether you are improving over the weeks. E.g. you can measure it once a month.

It's a long read but I think everyone will find it very interesting, so do check it out!

I'm a GM but my SQ is only 71.2, so I guess my macro is not really up to standard. Feel free to share your SQs!


The link can be found below!
http://www.teamliquid.net/forum/view...opic_id=266019

[crAzerk]

Man, the things some people do... I probably take as much time working on a school essay assignment as he took writing that. I wish there were Starcraft modules!
Great article btw

EDIT: Wow, just finished the article, all the math and stuff has me even more impressed. I wish I could do this for a living lol =X

[Welshy]

man if only i had this guys skills in writing lol was a good read non the less thanks for the link bro

[Uglymutt]

Good read. I'm impressed that all leagues seem to have a logarithmic band associated with that league.

[syfPhoenix]

haha yes i used his formula's to work out my macro is in the masters league yet i still cant win in masters

[Next_rim]

My macro is in masters range, but I'm still stuck in diamond.

[TragicHero]

Originally Posted by Next_rim

My macro is in masters range, but I'm still stuck in diamond.

I'm pretty sure most of us diamonds have a macro range of master league players. Frankly I don't see the difference between most master league and diamond league players, up till they're high masters (within GM MMR). So we should just win our matches and get promoted

[Next_rim]

Frankly I don't see the difference between most master league and diamond league players

In macro, there is little. execution and timings - that's what matters. Also, awareness and multitasking.

[RamboLicious]

Other than looking at the game graph right after finished it, is there a replay analyzer that can extract
1) Average Unspent Resources, and
2) Average Income?

Tried several one but won't help

[NDG.Stitch]

Originally Posted by Next_rim

My macro is in masters range, but I'm still stuck in diamond.

Macro is a big part in each game, but then along with Macro there is also reflexes and game sense. Just because you have a 200/200 pop army when your opponent has 150/200 doesn't mean you would win if you run all your units into tanks.
There are lots of players in Masters that have absolutely crazy macro but can't even deal with a well executed 2 rax or 4 gate!

[TragicHero]

Originally Posted by Next_rim

In macro, there is little. execution and timings - that's what matters. Also, awareness and multitasking.

Well, we ARE talking about macro right now aren't we


On topic:
Mine is around 60-80 over several games, and averaged out at 70.9. Need to get better at spending those minerals!

[johnnyli]

Wow that's a great write up. I really did notice that my worker macro was very sub par up until platinum.

[RamboLicious]

Some info gathered from my last 6 games:


Average SQ: ~ 98

However, I still suck cause my micro is so bad T_T

[syfChadMann]

yeah.. that thread is a load.. doesn't take into account that you DON'T make enough workers and not enough income.. it just balances your unspent verses your income.. if you have a really low income then it is easy to spend thus you get a good 'SQ'
its missing a third element of calculation which is so variable.

[RamboLicious]

Originally Posted by ChadMann

yeah.. that thread is a load.. doesn't take into account that you DON'T make enough workers and not enough income.. it just balances your unspent verses your income.. if you have a really low income then it is easy to spend thus you get a good 'SQ'
its missing a third element of calculation which is so variable.

Nah, I don't agree with you on this, bro.
SQ is calculated in a way that it's larger the more income you have, for example,

At #3, my AUS = 300 and my AI = 1035, my SQ is 90
but at #6, my AUS is tripled @ 929, while my AI is only doubled @ 2090, I still have a bigger SQ which is 101

[syfChadMann]

Originally Posted by RamboLicious

Nah, I don't agree with you on this, bro.
SQ is calculated in a way that it's larger the more income you have, for example,

At #3, my AUS = 300 and my AI = 1035, my SQ is 90
but at #6, my AUS is tripled @ 929, while my AI is only doubled @ 2090, I still have a bigger SQ which is 101

The goal of this calculator is to get a low number or a high number?

I dunno I ran it on a few 1v1 games I had and there was very different levels of macro being employed. The player that won out macro'd and stomped the other, however had more workers and more avg unspent... their SQ scores were similar.

The sum doesn't seem like it considers enough variables to be accurate.. like time and workers... but then you got to factor in harrasment causing lower worker counts at varied points in time. As for example, the first game with 501 unspent and 1731 avg income meals totally different things if the game length was 17 mins... or 37 mins. and will cause a significant difference to the macro in game.

That being said, I don't feel that I know enough about the formula to really make an assessment.. but i'll probably look into it soon..

[Next_rim]

As for example, the first game with 501 unspent and 1731 avg income meals totally different things if the game length was 17 mins... or 37 mins. and will cause a significant difference to the macro in game.

If it's averaged, it already is scaled for time. And if you look at average income, it already has worker # over time factored into it. Workers killed on harass = less income at some point in time = lower average.

[syfChadMann]

Originally Posted by Next_rim

If it's averaged, it already is scaled for time. And if you look at average income, it already has worker # over time factored into it. Workers killed on harass = less income at some point in time = lower average.

yes but as your income is an exponential growth scale your average income should increase as time goes on.. as you mine more at an increasing rate..

So if you avg income is 1000 minerals at 10 mins (for example, lets presume this value is really good)
you can also have a 1000 avg income at 25 mins.. which in this example would be bad and might present missed peaks of the exponential line... if you also had the same upsent reasources your SQ would be the same. and I know... 1000 at 10 mins = 10,000 and @ 25 mins = 25,000 which seems good.. but in ten mins.. that could be the best you can get.. but in 25, you could get better?...

basically it doesn't consider potential income in the game length... say, it is possible to have 2000 avg income at 25 mins? you see what i mean...

Its about the player cutting worker production and plateauing their economy at a particular income per minute. when does that happen, and how well do they spend that income.. is the real test of macro ability.

I hope I'm making sense.. :S

[RamboLicious]

Originally Posted by ChadMann

The goal of this calculator is to get a low number or a high number?

I dunno I ran it on a few 1v1 games I had and there was very different levels of macro being employed. The player that won out macro'd and stomped the other, however had more workers and more avg unspent... their SQ scores were similar.

The sum doesn't seem like it considers enough variables to be accurate.. like time and workers... but then you got to factor in harrasment causing lower worker counts at varied points in time. As for example, the first game with 501 unspent and 1731 avg income meals totally different things if the game length was 17 mins... or 37 mins. and will cause a significant difference to the macro in game.

That being said, I don't feel that I know enough about the formula to really make an assessment.. but i'll probably look into it soon..

The goal of this calculator is just for calculating your SQ. Your goal to "macro like a pro" can be translated as having a high SQ.

The length of your game wouldn't really matter, as long as it's "long enough" and it's a normal game (I wouldn't call it normal if you have 10k mineral and 200/200 food for like 5 minutes - if you know what I mean). What really matter is your AI (Average Income) and your AUR (Average Unspent Resource)

As the example I gave above, the more bases you have, the more AI and the harder it is to maintain a lower AUR.
Hence, while in the 2nd game, my AI is only twice as much, and my AUR is tripled, I still have a better - higher SQ.

[Next_rim]

you see what i mean...

I do see what you mean, but it's hardly applicable to starcraft for two reasons. First, probe count (and thus income) beyond 70 makes your army weak. Therefore, there is a theoretical "maximum" income you can and should have. Second, any macro game involves a 3rd base somewhere around 14 mins, if not earlier. Author recommends to use games that are 15+ minutes long for this very reason.

Now, if you look at difference between a 15 min game on 3 bases and 999 min game (still on 3 mining bases), the only factor that affects income is game actions - harass and more expansions. If you slip on expansions, or lose probes to harass, or lose a base, average income will plummit, signalling worse macro. Although it might not be directly connected to player macroing badly, maybe he was just out of position to defend a base, but a pro supposedly will get back on track very soon compared to lower league. That's why the author states that it's important to use longer games, since it is more likely the game will feature 3 mining bases.

[syfChadMann]

Originally Posted by RamboLicious

The goal of this calculator is just for calculating your SQ. Your goal to "macro like a pro" can be translated as having a high SQ.

The length of your game wouldn't really matter, as long as it's "long enough" and it's a normal game (I wouldn't call it normal if you have 10k mineral and 200/200 food for like 5 minutes - if you know what I mean). What really matter is your AI (Average Income) and your AUR (Average Unspent Resource)

As the example I gave above, the more bases you have, the more AI and the harder it is to maintain a lower AUR.
Hence, while in the 2nd game, my AI is only twice as much, and my AUR is tripled, I still have a better - higher SQ.

Refer to my previous post... but can i also add..
This also doesn't consider situations where you get a 200/200 army vs another 200/200 army.. obviously you don't want to just run in cause you need to keep macroing.. so it's a good idea to sack units... no.
you might sit on 200 looking for the right position or to catch your opponent out of position before you engage... This could cause you avg unspent to sky rocket... tho of course its a good idea to build more unit producing facilities so you can macro off the back of the big 200 vs 200 engagement.. which is going to spend some money..

anyway, this SQ calculator is an interesting concept and a good benchmarking tool.. but I think it has alot of flaws and lacks to consider some variables.. which SC2 is filled with...
Just my opinion.

[RamboLicious]

Originally Posted by ChadMann

yes but as your income is an exponential growth scale your average income should increase as time goes on.. as you mine more at an increasing rate..

That really depends, bro. But most of the time, your AI would be steady quite steady after, say 15 minutes mark, unless you are owning your opponent who doesn't harass at all, and have like 5 bases etc..

So if you avg income is 1000 minerals at 10 mins (for example, lets presume this value is really good)
you can also have a 1000 avg income at 25 mins.. which in this example would be bad and might present missed peaks of the exponential line... if you also had the same upsent reasources your SQ would be the same. and I know... 1000 at 10 mins = 10,000 and @ 25 mins = 25,000 which seems good.. but in ten mins.. that could be the best you can get.. but in 25, you could get better?...

basically it doesn't consider potential income in the game length... say, it is possible to have 2000 avg income at 25 mins? you see what i mean...

Its about the player cutting worker production and plateauing their economy at a particular income per minute. when does that happen, and how well do they spend that income.. is the real test of macro ability.

I hope I'm making sense.. :S

Understand your point, however, you didn't consider other possibilities.
For example, if you're being owned, harassed hard, and you still can keep your AI at 1000, even at 25 mins mark, and I would consider it good.

That being said, there are a lot of things needed to be considered in Macro Ability evaluation. However, those are really hard / impossible to do.
Taking an statistical approach by whatthefat @ http://www.teamliquid.net/forum/view...opic_id=266019 is the best way to do it . If you have time to read through the post, I think it would be much more helpful.

[RamboLicious]

Originally Posted by ChadMann

anyway, this SQ calculator is an interesting concept and a good benchmarking tool.. but I think it has alot of flaws and lacks to consider some variables.. which SC2 is filled with...
Just my opinion.

yep. that's what it really is. It's surely not 100% correct, that's why we need more than just a small number of examination.
There can be a lot of flaws, however, statistically, that still reflect the macro ability of one vs the other.
A better / higher class player will have a great chance of having higher SQ comparing to the other

[syfChadMann]

Originally Posted by Next_rim

I do see what you mean, but it's hardly applicable to starcraft for two reasons. First, probe count (and thus income) beyond 70 makes your army weak. Therefore, there is a theoretical "maximum" income you can and should have.

Okay, so thats exactly the point.. income is exponential until the point of selective plateau. Picture the income graph at the end of the game.. it goes up exponentially (well, it should) then plateaus... poor macro players might plateau the income at 45 or 50 works.. where as your suggesting 65-70 is better, right? so in varied game lengths, the two players can theoretically get the same SQ from similar Avg income and avg unspent. This is the fault of the SQ calculation.

[Next_rim]

poor macro players might plateau the income at 45 or 50 works.. where as your suggesting 65-70 is better, right? so in varied game lengths, the two players can theoretically get the same SQ from similar Avg income and avg unspent.

Not really, since average income of 45 workers in 45 mins will still be lower that average income of 70 workers in 15 mins. In addition, 70 workers in 15 mins and 70 workers in 45 mins will still have same average income, but varied unspent resources. And a better macro player will have lower average unspent resources, so you still get to measure macro penis length. Besides, it all doesn't matter, since you look at 25+ games of 15-25 mins, and those rare outlier differences (if possible, which I doubt), will be averaged out by a large number of normal expected results.

[Volition]

I am enjoying the opinions and reasoning in this thread. While ChadMann is correct, it isn't perfect, SQ still appears to be a fairly handy tool to evaluate macro. To not use this tool due to the slight imperfection is like throwing a book away because it has a scratch on the cover. The 200/200 point is correct - but we play to win not to get high SQs. Maybe it isnt your SQ overall, but compared to the other player in the game (there are of course exceptions, due to units that can vastly cover their costs, such as siege tanks).

I think SQ as a tool is fantastic, and the reason is because of the "rules" of macro:-
1) always build probes;
2) never get supply blocked; and
3) always spend all of your money.

If you are doing all three of these, your SQ can only be high. Some things will affect it, but if you spend 25 mins with >800 resources, and then when you are at 200/200 it shoots up to a few k, your average unspent resources will not be that bad - because you have macro;d like a boss for 25 mins!

[TABottles]

It's a very good thread, I was on it a couple of days ago. The OP put so much time and effort into it. Some good data and comparisons between leagues. I usually lurk around The Closed Thread Lounge (lol) so this post was a nice change.

[shmiman]

It's quite cool to have something to gauge your macro skills against others, but I agree it could probably do with some refinement like what if one cuts worker production and such. Otherwise, gj to the poster, really much effort put into it

[Evan]

this results not really accurate i found out. But you can use it as a reference but dun take this too seriously

[AxS.Sorathez]

My SQ in games of less than 10 minutes (read: against protoss and when I lose badly to zerg) Is around 73.6 which puts me at master level, although I'm in Gold.

In games LONGER than 10 minutes however is around 51.3 which puts me around Gold, where I should be.
Whereas my overall score (63.4) put me in platinum-near diamond.

(This was all based on the last 10 games I played with 600 average income and above and excluding some practise matches).

[TABottles]

Hey guys I don't know how to find my accumulated bonus pool, though it's written in the thread. Something about 5 placements? Where do i go from that?

[Volition]

Originally Posted by FaDeSorathez

My SQ in games of less than 10 minutes (read: against protoss and when I lose badly to zerg) Is around 73.6 which puts me at master level, although I'm in Gold.

In games LONGER than 10 minutes however is around 51.3 which puts me around Gold, where I should be.
Whereas my overall score (63.4) put me in platinum-near diamond.

(This was all based on the last 10 games I played with 600 average income and above and excluding some practise matches).

Sorathez, those numbers you need to work on your macro when you are on multiple bases (that is, games longer than 10 minutes). Check out a few of your replays, and if you still are unsure how many gateways/robos to build, ask a higher player what building compositions they have on 2/3/4 bases.

Beware of statements such as "my overall score put me in platinum-near diamond". The SQ doesnt indicate that at all - it is just saying that you are mining minerals and spending them at a rate consistent with a platinum player. But a platinum/diamond player may do things you do not, such at multitasking drops, micro, scouting and positioning.

At the end of the day SQ is a tool on one element of the game. But still very useful.

[inFeZa]

Maths... the bane of my existence.
Cool read none the less.

[AxS.Sorathez]

Originally Posted by Volition

Sorathez, those numbers you need to work on your macro when you are on multiple bases (that is, games longer than 10 minutes). Check out a few of your replays, and if you still are unsure how many gateways/robos to build, ask a higher player what building compositions they have on 2/3/4 bases.

Beware of statements such as "my overall score put me in platinum-near diamond". The SQ doesnt indicate that at all - it is just saying that you are mining minerals and spending them at a rate consistent with a platinum player. But a platinum/diamond player may do things you do not, such at multitasking drops, micro, scouting and positioning.

At the end of the day SQ is a tool on one element of the game. But still very useful.

I'm well aware of this, it was merely a statement based on the data I have collected.
The overall score thing, was based purely on the spending quotient, I am not nearly silly enough to believe it makes me a platinum/diamond level player :P StarCraft is an intricate game, and macro is but one part of it.

[c_kAelle]

good find, ill check that out now