|
I think the approach of using the stock value of resources in a given base is a poor way to consider what unit mix is reasonably affordable. I posit that a more appropriate approach would be to use the rate of collection (income) as this more accurately defines what resources are available to a player at a given point in time.
Now, a standard base has 8 mineral patches and 2 gas gysers. The marginal rate of collection for the first 2 gatherers on a resource patch is 40 per worker. For the third worker per patch that rate reduces to ~23. Therefore, if we 16 gatherers on 8 mineral patches the income would be 640, and if we had 24 gatherers the income would be 824. For gas geysers the first 3 gatherers all have a marginal rate of collection of 40...so each geyser with 3 workers nets us 120 vespene.
We can give these numbers actual value by comparing them to your normal harvesting strategy. Early game when you only have ~19 gatherers, 16 on minerals and 3 on gas your income is approximately (640,120), with a ratio of (5.3 : 1). If you proceed to double gas your income will be (640, 240) with a ratio of (2.7 : 1). A "fully saturated" base with 24 on minerals and 1 vespene gives an income of (824,120) (6.9 : 1), and 2 vespenes is (824,240) (3.4 : 1).
So how are these numbers useful? We can start to craft efficient builds or plan our harvesting strategy to maximize our resource use. An important step in this process is understanding the real cost of units. For example, a marauder does not cost (100,25). One must remember that a marauder has 2 supply and because 8 supply costs 100, the real cost of the marauder is (125,25) or (5 : 1)..assume the minerals lost in building time to be negligible as you can create an extra gatherer who is on build duty. In this manner we can properly allocate costs to units for more accurate builds...for example if you want to be able to do a specific timing push with a certain number of units at a certain time you can figure out what upgrades and how many production buildings you'll need and you can allocate these costs on a per unit basis. The end result is that you will get a ratio of minerals to gas needed. You then can try to match this "required" ratio to your actual "gathering ratio." This will help to keep your resources low and your output as efficient as possible.
Critiques?
*For those who are not mathematically inclined, this is not too pertinent as you will naturally figure out these ratios and get a feel for what you can afford, however this method could be useful for refining a build.
** Also do not assume you can simcity for 8 or 9 minutes without getting harassed/attacked...this does not mean this info is useless. Over time you can develop an idea of how many gatherers/units you lose to early game antics and can then allocate these additional costs to your resource ratios.
|
A nice way to continue the discussion in my gas = minerals thread, but since that one got locked I am pretty sure that JWD will appear and lock this one as well, on the grounds it's the same arguement reworded.
Still, if this lasts more than 5 minutes I will come and add my thoughts  .
|
he's in no way saying that gas = minerals, he's merely simplifying the "notation" a bit, it's a nice idea could be helfpul. but i'd say tha the actual numbers do matter, because you need to know how many units you can produce and that comes from the actual number of resources collected and not just from the minerals to gas ratio.
|
Hyrule19200 Posts
You say there's a mining rate of 40 per worker. You need a time to have a rate.
|
On November 12 2010 23:48 tofucake wrote:
You say there's a mining rate of 40 per worker. You need a time to have a rate.
It's per minute or whatever the blizzard rate is in.
|
What he's saying is that you need to consider the cost of supply as well and to factor in collection v production rate.
I too had considered something like this - but came across this snag; tech buildings which otherwise would not be built if that unit was not going to be built. Should this be factored into the cost of the unit? What happens when multiple units benefit from this tech -> reactor gets you marauders, reapers, tanks etc and what of upgrades?
|
On November 12 2010 23:30 jrdn wrote:
Early game when you only have ~19 gatherers, 16 on minerals and 3 on gas your income is approximately (640,120), with a ratio of (5.3 : 1). If you proceed to double gas your income will be (640, 240) with a ratio of (2.7 : 1).
When you double gas then your mineral income is lower due to 3 more worker on gas.
I thought its common sense to base your expenses on the amount of workers at a certain point in time. Subtract supply and worker cost and you got approximately the amount of minerals you can spent.
|
On November 12 2010 23:59 ezdez wrote:
I too had considered something like this - but came across this snag; tech buildings which otherwise would not be built if that unit was not going to be built. Should this be factored into the cost of the unit? What happens when multiple units benefit from this tech -> reactor gets you marauders, reapers, tanks etc and what of upgrades?
This is a good point. Perhaps when dealing with tech structures and upgrades it will be better to have a separate "account." For instance..if you are figuring a build with 2 marauders, 6 marines, concussive shell, and 2 siege tanks w/ siege mode the addition would be: 2 marauders (2 x 125,25) + 6 marines (6 x 62.5,0) + 2 siege tanks (2 x 187.5, 125) + upgrades (concussive shell, (50,50), tech lab (50,25), siege mode (100,100) + factory (150,100)) = (1350, 575) = 2.3 : 1. Therefore, we will need a third gas to keep up with mineral production to maximize (even 16 gatherers in minerals and 6 in gas gives a ratio of 2.7:1). Or perhaps we should consider a build with less gas demand. Perhaps more marines and no marauders.
|
The rates (per minute? could you please edit to clarify?) you mention are maximal values in ideal settings. How do you suggest one reasonably models the transition from the initial 6 miners to a fully saturated base (assuming no enemy harassment, as you mentioned, or other factors which prevent miners from following optimal paths, like turrets in the mineral line)?
Furthermore, your numbers hold on Blizzard maps at the moment, but disregarding the future directions mapmakers will follow, how does the following case fit into your model: A Zerg player who expands early (the now typical 14 hatch) while scouting her opponent to find out she can overrun him with Zerglings, and does so successfully without producing any units that require gas. How does the ratio of minerals/gas help the Zerg player make any decisions, if at all? My opinion currently is that it doesn't.
Any chance you could explain your theory better?
Cheers.
|
On November 13 2010 00:40 bashalisk wrote:
The rates (per minute? could you please edit to clarify?) you mention are maximal values in ideal settings. How do you suggest one reasonably models the transition from the initial 6 miners to a fully saturated base (assuming no enemy harassment, as you mentioned, or other factors which prevent miners from following optimal paths, like turrets in the mineral line)?
Furthermore, your numbers hold on Blizzard maps at the moment, but disregarding the future directions mapmakers will follow, how does the following case fit into your model: A Zerg player who expands early (the now typical 14 hatch) while scouting her opponent to find out she can overrun him with Zerglings, and does so successfully without producing any units that require gas. How does the ratio of minerals/gas help the Zerg player make any decisions, if at all? My opinion currently is that it doesn't.
Any chance you could explain your theory better?
Cheers.
As far as the rates per minute: when watching a replay if you check the income tab these are the numbers I am using. I assume they are on a x per game minute basis.....so 1 gas drone would harvest 40 vespene per game minute.
This model is not a growth model, however it can be used to help prepare a growth strategy. For instance, the model will not tell you when to make things, however, if you know you want to build a certain number of roaches at a certain time the model will be able to guide you in determining how to allocate your drones.
The model is more designed to maximize resources based on a unit mix you want. For instance, if you want 4 lings for every roach you have, this model could be helpful. 4 lings = (125,0); 1 roach = (100,25). The total equals (225,25) or 9:1. This means if your resource collection ratio is less than 9:1 (like 5:1) you will have excess gas and should plan on spending it somehow so it doesn't pile up.
However, the model is clearly not necessary if you are using a minerals only build as described above (less the 100 gas for the speed).
|
OK I see where you're coming from with this. It might be good for early game optimization, I suppose. I'll have to find some time to use it on an existing build by tomorrow morning, and maybe I'll post the results here.
|
|
|
|
|
|
EPT
