|
Poll: Which zerg's standard build gives more larvae?
(Vote): 13 pool with queen, hatch later
(Vote): 14 hatch 14 pool, queen later
02.04.2010
UPDATED: SCROLL DOWN FOR THE ANALYSIS
Im seeing that zerg players are divided into two:
Some zergs are using 13 pool queen first openings, the others are using 14 hatch.
The idea is that when u get an early queen it gives 4 larva in every 40 seconds (production rate: 1/10=0,1 larvae per second), compared to hatch that gives 1 larvae in every 15 seconds (production rate: 1/15=0,067 larvae per second) you can see that when also considering building times and orders, its quite a complex equasion.
So ill throw those builds in
queen expand:
13 pool
13-16 gas(or after queen)
15 queen
18 hatch
hatch expand:
14 hatch
14 pool
13 gas
queen after pool
Note that those build may vary when to take gas etc, but in the big picture they are similar to those what players are using.
So now we need an instant, when production rates of those both builds are the same.
And this is a moment when queen and hatchery are both finished on both builds.
(2 hatcheries and a queen will give production rate: 2 x 1/15 + 1/10 = 7/30 = 0,233)
Given this instant, which build is more efficient in terms of larvae production?
EDIT:
So here it is.
INTRODUCTION
When I originally thought of this issue it was tailored towards zvz, because in zvz considering recent trends (by that i mean masslings) production capacity is very important maybe even the key component.. So I read this thread here and compared that buildorder(without the upgrade) described there to a buildorder that I was using and found very good. Yes it was 13 pool speedlings. But as i analyzed the buildorders, the knowledge became more general, so that info can be used in other matchups as well.
ANALYSIS
These calculations are based on practical values(timings found on replays) and theoretical calculations. By that I mean calculated using exact given build times and not timings found on replays. To make things clear, I add an example:
From a replay we get a time value for start of a spawning pool lets say t(A), and need to know timing when the first 4 larvae will pop t(L). Then t(L) will be calculated so: t(L) = t(A) + 65 + 50 + 40 .
You will also notice that I assume that there are no delayes, other than waiting for mineral for pool and a hatch without sacrificing larvae.
This method is good because it will minimize the error limits when comparing those two buildorders.
These are the exact builds I used to get the timings:
13pool:
10 overlord
13 pool
16 overlord
16 queen
17 hatch
14hatch:
10 overlord
14 hatch
13 drone
14 pool
We only need timings for a pool and a hatch, the others can be calculated based on them.
Note that 13 pool doesnt contain gas, because it will delay the hatch considerably.
For 14 hatch the gas is taken after the hatch and the pool, so must be done for 13 pool also.
Why 13pool and not 14,15,16?
Because it will delay spawning pool and with that the queen. And will result in lesser larvae production.
Still you can argue that those 1-3 drone delayed from mining will compensate faster hatch.
This can be easily calculated, in fact you can do it yourself playing with this toy that I made for you.
So enjoy  .
Why 14hatch and not 13 or 15?
Because with 13hatch you will sacrifice a larva and in terms of optimizing its a huge loss(15 seconds). 15 and later hatch will delay increase of your production capacity again.
The main Formula
Larvae count = Production rate X Time interval
Assumption: Production rate is constant during selected time interval.
RESULTS
I made a toy to help with the calculations, you can play with it yourself here!
Yellow cells are only meant to be modified. This tool only compares different queen and hatch timings and calculates the larvae count based on that. It cannot be used for anything else.
Input data I used, based on my test replays:
Starting time for spawning pool (in seconds)
13pool: 110
14hatch: 155
Starting time for hatchery (in seconds)
13pool: 190
14hatch: 130
Note that these are estimated values, and may deviate in a small margin.
This chart shows larvae production rate characteristics.
As you can see there are small windows of time where the production rates deviate.
The areas under the lines are in fact larvae count.
The time when both production rates go even is @290 second mark.
The areas before that time line are the larvae count we are looking for.
Calculations indicate that the builds are practically DEAD EVEN in terms of larvae production:
25,83 larvae for 13pool
25,33 larvae for 14hatch
The difference is so small that it will certainly fit into error range.
CONCLUSION
As seen in the chart there are 2 bigger windows where the production rates deviate:
The first window will cause 13pool build to have earlier extra larvae (It will be about @265 second mark). In economic point of view its better to have drones earlier, but in the same time you will have already full saturation(2drones per patch) at your main. Luckily 2nd hatch will complete in about 10 seconds (265+17 compared to 290) when those 4 drones pop up.
The second window appears when 14hatch completes the queen and starts to catch on 13pool's larvae count, in practice this will not happen sooner than 270+40=310 second mark.
Although builds are equal in terms of production rates 13pool will give more options early on and is alot safer and because of that it is more superior in my eyes.
Only thing where 14hatch has advantage over 13pool is earlier gas.(assuming that 13 pool gets gas after hatch) and also maybe some sort of tactical value of hatch placement.
FINAL WORDS
That said I have a suggestion for optimizing the 13pool:
In that shape the buildorder has very late gas. If you need earlier gas the best time to take it
is around 16 supply just right after the overlord and before start of the hatch, ideally hatch will start at the same time when extractor finishes.
You only take 1 worker off the minerals(and not 4), so that the hatch delay is minimal.
It will be about @160 second mark: exactly at the same point when 14hatch takes gas!
So the ideal opening should look like this:
13 pool
16 overlord
16 gas
15 queen
17 zerglings
18 hatch
I havent tried this exact variation myself, but it looks good on paper.
Try it and tell me what you think!
03.04.2010
EDIT#2
I made another graph:
This shows produced larvae count.
Look at the time between 235 - 275 more closely.
14hatch will produce between this time slightly more larvae.
So if you make drones that means you can put them to work sooner than with 13pool.
That gives a slight economic edge (about 30-50 minerals).
Cheers!
|
Considering that more larvae is the only benefit of going 14 hatch, and that many pro players do it, I'm pretty sure 14 hatch gives more larvaes (and it make sense, too).
That said, I'd be very interested in the calculations and results no matter what.
|
I'd go with early queen because if you get rushed/proxied, the 2nd hatch goes down and you're left with no queen.
|
You can also consider that with a faster 2nd hatch you can get a faster 2nd queen as well.
|
Osaka27149 Posts
There are a lot of intangibles when it comes to actual gameplay (especially for defense and scout prevention) but in terms of straight larvae I guess the hatch first build is better?
|
When a hatch pops, you get 1 larvae to start off with no? When a queen pops, it has enough energy to spew larvae. queens take less time to make, and cost less money. So i would say if you used the extra larvae toward drone production, you'd actually get a faster economy in the time between when the queen and hatch get started and when they both pop. The advantage to having two hatcheries is that you have more minerals/gas locations to mine, and therefore i higher saturation yield. Since saturation doesn't come into play significantly in the time between laying the second hatchery in either of these builds, i would have to say going queen first is not only safer, but more economically viable and leaves the zerg with more options, and gives more opportunity to hide tech, due to having a semi-ranged defense unit right off the bat. So yeah, fast queen ftw in my opinion
|
People are forgetting that by going 14 hatch you get a double hatch/double queen economy MUCH faster. While there is indeed a timing window where you have more larvae as pool first (hence why its safer), the hatch build will not long after surpass the pool build assuming no very early aggression.
|
On March 31 2010 09:42 Thamoo wrote:
People are forgetting that by going 14 hatch you get a double hatch/double queen economy MUCH faster. While there is indeed a timing window where you have more larvae as pool first (hence why its safer), the hatch build will not long after surpass the pool build assuming no very early aggression.
I like your approuch.
You are right about that timing window if you are talking about zvz, with those both builds.
But that window opens up the instant when 13 pool finishes queen and closes when 14 hatch finishes queen.
In fact double hatch/double queen have practically same timing on both builds.
|
On March 31 2010 09:56 scAre wrote:
In fact double hatch/double queen have practically same timing on both builds.
In the scenario I'm playing in my head I find this hard to belive. I'm eager to see the calculations behind this =)
|
On March 31 2010 09:42 Thamoo wrote:
People are forgetting that by going 14 hatch you get a double hatch/double queen economy MUCH faster. While there is indeed a timing window where you have more larvae as pool first (hence why its safer), the hatch build will not long after surpass the pool build assuming no very early aggression.
That window opens up the instant when 13 pool finishes queen and closes when 14 hatch finishes queen(Im comparing production rates). After that 14 hatch will surpass 13 pool production rate, until 13 pool second hatchery finshes and rates go even. So u see this is a tricky one.
|
I've personally been favouring 9ovie, 13pool, 17ovie, 17extractor, 16 queen. The timing of those last three happens at almost the same moment - from there you have enough resources to throw down a second hatch or roach den or whichever path seems best - it's also roughly when you'll be scouting on Steppes of War [and other longer distance maps].
I've personally been favouring a greedy double expand build. It's a bit weak against an early timing attack but once your econ kicks in, you can macro extremely hard.
|
What we need is a graph where Y is the total number of larvaes produced total and X is time. Any volounters? ^^
Klimp : 13 pool is pretty much 14 pool, I don't think the differance between the two is that noticable compared to 14 hatch or 14 pool.
Double expend seems way too vulnerable to me, as well as unneccessary as 2 hatch/2 queen is already hard to support ressource wise in the early game.
|
Osaka27149 Posts
On March 31 2010 09:36 Stropheum wrote:
When a hatch pops, you get 1 larvae to start off with no?
No, you get three.
|
More larvae directly equals better economy, right? Since you won't reach full saturation with 1 gas before your second hatch is up as 13 pool variation.
Also, is 15 pool worse economy wise than 13 pool? 2 drones faster vs about 1 less drone later. (Can the 2 drones make up 50 minerals though?)
|
More larvaes only means better economy if you have the mineral/gaz required to make good uses of those larvaes. Idle larvaes aren't good economy.
15 pool is a better economy than 13 pool, else why on earth would anyone make a 15 pool? (And pros occasionally 15 pool).
|
15 pool vs 13 pool is also tricky comparison
With 13 pool you can get a bit faster queen and that means more larvae, but in the same time those 2 drones will be delayed a bit from mining minerals.
We can say that 15 pool is more economical compared to 13 pool when you are not planning to make drones from those extra 4 larvae (make lings instead)
But may or may not be economical when you are planning make drones from those 4 extra larvae.
I personally think that in that case they are practically same economy-wise.
So the question is: why make pool later and risk more when you get the same result from earlier pool?
These are my thoughts.
|
This should be very easy to test. Just get a partner and make a game. Player 1 does one build, Player 2 does the other. Only pump drones, and don't bother with gas (irrelevant for testing purposes). If each player focuses only on building drones and never supply caps themselves (make sure both players are decent at macroing), you can stop the reply every 15 seconds and get a supply count from each player. Write these values down and create a graph of time vs. supply for both players. Put them in excel and chart it over like 5 minutes. It should be pretty easy to tell exactly at what points each build excel.
|
I think the big benefit of early hatch is the ground control and economy it provides once the expansion gets going. While pool first orders give more larva, FE gives you more optimum workers saturations.
Zerg are finding that early game they can FE without too much trouble, and the gain they get once workers start pumping more than makes up for the slightest larva they get early game.
|
I think some people are assuming that 13pool queen forgoes a second hatch altogether. The fact is that this presents a larger scouting window on your opponent where you can either do a push to deny a fast expand, or drop a second hatch quickly, in which case i believe the timings are relatively the same to 14hatch, if not slightly quicker. The reason i think 13pool queen is faster is because you get more drones coming in while the second hatch is morphing, therefore you have a slightly later hatch, but you can theoretically instantly saturate the expo, while as the 14hatch will only have however many drones you maynard when it completes. Therefore, i think 13pool queen is better overall economy, bigger windows for choices, and the queen for defense.
|
On March 31 2010 10:14 Thamoo wrote:
Klimp : 13 pool is pretty much 14 pool, I don't think the differance between the two is that noticable compared to 14 hatch or 14 pool.
Double expend seems way too vulnerable to me, as well as unneccessary as 2 hatch/2 queen is already hard to support ressource wise in the early game.
13 pool has some really nice timings with expanding on 18 and a heap of flexibility with what your opponent does. You also have enough resources to get a Queen AND a Roach AND a Extractor as soon as the Spawning Pool finishes.
With my double expand build, I've been expanding to my natural and a hidden [pref high yield]. People generally don't look for a hidden expansion super early in the game - I've had it pay off far more often than not.
|
|
|
|
|
|
EPT
