|
On November 29 2010 04:24 Guerrilla705 wrote:Show nested quote +I regard to many of the comments, I want to clarify that this is simply to PROVE which general order timing will yield best economy in the long run ok that is a VERY vague statement. Like others have said, the best economic build is to never pool, you can't say that we are doing this build independant of other pressure, but we still need a spawning pool. If you are going to put in a spawning pool then what the other player is doing must be factored in, because Zerg is a reactionary race. We don't get to just start every game however we please, we must respond to what the opponent is doing. Also, you discredit yourself as a scientist if you are saying this single experiment proves anything. Science never PROVES anything, only attempts to disprove. If you want to make a scientific argument you need to recognize your uncontrolled variables, the problems with this experiment, and suggest ways to improve the experiment. It needs to be repeatable for others and they need to be able to do it better
Ok thanks for splitting hairs in regard to semantics.
If you honestly think a no pool build is more economical, maybe you should submit it for testing... but I doubt you would like the results.
Let's try to keep the trolling to a minimum here please.
|
I think that you its worth noting that the ideal eco build, would obviously be unobtainable in a real game. Its a balancing act of risk and reward, there no point in seeing who can load the most on one side of the scale
|
Anyways, here's an 11 pool FE build. In my test, whenever I sent a drone to minerals, I always sent it to the same patch (the fourth one, clockwise). When using drones to build things, I tried to take them from the middle, or rally them from eggs.
10 Extractor trick. (rallied an egg)
11 Spawning Pool
10 Overlord
10 Extractor trick
14 Queen (note: I didn't have to wait for minerals)
17 Hatchery (note: rallied the egg made after queen, and it arrived just in time)
17 Overlord
Use spawn larvae when queen pops (same time as overlord is started)
20 Queen
Use spawn larvae again when available (~25 supply), then send queen to expansion
26 Overlord
@160 minerals (about) send 6 drones to expansion, and set both rally points to expansion
Use spawn larva twice, when available
33 Overlord
36 Overlord
Use spawn larva twice, when available
44 Overlord
Reset main rally point to main
My results:
Supply = 49/60: 41 drones, 6 overlords, 2 queens, 2 drone eggs at 16/17, and 2 drone eggs at 1/17
Current minerals = 343
Both hatcheries have about 13 seconds before making a larva
Both hatcheries are at 30/40 in the spawn larva cycle
Edit: I can't figure out how you're counting drones or minerals. The above number for drones is what's actually out and mining at that second.
Edit2: I slacked a bit on spawn larvae, and didn't optimize my starting drones like encRoach did
|
On November 29 2010 04:33 Hurkyl wrote:
Also, it is relevant to note where in the larvae cycle the hatcheries are, as well as the spawn larvae. There's a big difference between having the next spawned larvae drop at 6:01 and 6:39!
This is an important point that I'm glad someone mentioned...
However, I feel these numbers are less relevant toward overall economy than drone count (including production) and minerals mined. If two builds are very similar in these regards, then the "larva in production" data would also be factored in.
|
On November 29 2010 05:26 jdseemoreglass wrote:Show nested quote +On November 29 2010 04:33 Hurkyl wrote:
Also, it is relevant to note where in the larvae cycle the hatcheries are, as well as the spawn larvae. There's a big difference between having the next spawned larvae drop at 6:01 and 6:39! This is an important point that I'm glad someone mentioned... However, I feel these numbers are less relevant toward overall economy than drone count (including production) and minerals mined. If two builds are very similar in these regards, then the "larva in production" data would also be factored in.
The problem with ignoring this data is that an inferior build would look better than a superior build if you sample just as the inferior build gets its benefit from spawn larvae -- this can be up to an 8 drone difference!
If you're looking at supply, for example, the current leader gets a lot of benefit from this effect since a spawn larva finishes just before 6:00, giving it a temporarily inflated supply count.
|
On November 29 2010 00:52 falstag wrote:Okay, you want to stop at the 6 min mark. Thats enough time for a expo and 2nd queen. I ran evo chamber and told it ot build 40 drones, 2 queens, and 2 bases in 6:00. It was able to complete those requirements at 5:49. 40 drones will fully saturate two mineral lines, but you will have no gas. my evo chamber claims that 16 hatch 18 pool is the way to do it. It leaves you supply blocked at 40/38, and leaves you with 1146 mins. + Show Spoiler +@0:00 M:50 G:0 L:3 S:6/10 BuildDrone @0:15 Spawned: Larva+1 @0:17 Spawned: Drone+1 @0:17 M:54 G:0 L:3 S:7/10 BuildDrone @0:27 M:51 G:0 L:2 S:8/10 BuildDrone @0:32 Spawned: Larva+1 @0:34 Spawned: Drone+1 @0:44 Spawned: Drone+1 @0:47 Spawned: Larva+1 @0:47 M:104 G:0 L:3 S  /10 BuildOverlord @0:55 M:53 G:0 L:2 S /10 BuildDrone @1:02 Spawned: Larva+1 @1:12 Spawned: Overlord+1 @1:12 Spawned: Drone+1 @1:12 M:106 G:0 L:2 S:10/18 BuildDrone @1:12 M:56 G:0 L:1 S:11/18 BuildDrone @1:17 Spawned: Larva+1 @1:19 M:53 G:0 L:1 S:12/18 BuildDrone @1:29 Spawned: Drone+1 @1:29 Spawned: Drone+1 @1:32 Spawned: Larva+1 @1:32 M 5 G:0 L:1 S:13/18 BuildDrone @1:36 Spawned: Drone+1 @1:47 Spawned: Larva+1 @1:47 M:177 G:0 L:1 S:14/18 BuildDrone @1:49 Spawned: Drone+1 @2:02 Spawned: Larva+1 @2:02 M:270 G:0 L:1 S:15/18 BuildDrone @2:04 Spawned: Drone+1 @2:10 M:300 G:0 L:0 S:16/18 BuildHatchery @2:17 Spawned: Larva+1 @2:17 M:68 G:0 L:1 S:15/18 BuildDrone @2:19 Spawned: Drone+1 @2:32 Spawned: Larva+1 @2:32 M:171 G:0 L:1 S:16/18 BuildOverlord @2:34 Spawned: Drone+1 @2:47 Spawned: Larva+1 @2:47 M:234 G:0 L:1 S:16/18 BuildDrone @2:57 Spawned: Overlord+1 @3:02 Spawned: Larva+1 @3:02 M:349 G:0 L:1 S:17/26 BuildDrone @3:03 M:310 G:0 L:0 S:18/26 BuildHatchery @3:04 Spawned: Drone+1 @3:17 Spawned: Larva+1 @3:17 M:163 G:0 L:1 S:17/26 BuildDrone @3:19 Spawned: Drone+1 @3:25 M:202 G:0 L:0 S:18/26 BuildSpawningPool @3:32 Spawned: Larva+1 @3:32 M:79 G:0 L:1 S:17/26 BuildDrone @3:34 Spawned: Drone+1 @3:47 Spawned: Larva+1 @3:47 M:196 G:0 L:1 S:18/26 BuildDrone @3:49 Spawned: Drone+1 @3:50 Spawned: Hatchery+1 @4:02 Spawned: Larva+1 @4:02 M:326 G:0 L:2 S:19/28 BuildDrone @4:02 M:276 G:0 L:1 S:20/28 BuildDrone @4:04 Spawned: Drone+1 @4:17 Spawned: Larva+1 @4:17 M:417 G:0 L:2 S:21/28 BuildOverlord @4:17 M:317 G:0 L:1 S:21/28 BuildDrone @4:19 Spawned: Drone+1 @4:19 Spawned: Drone+1 @4:30 Spawned: Spawning Pool+1 @4:32 Spawned: Larva+1 @4:32 M:479 G:0 L:2 S:22/28 BuildDrone @4:32 M:429 G:0 L:1 S:23/28 BuildDrone @4:32 M:379 G:0 L:0 S:24/28 BuildQueen @4:34 Spawned: Drone+1 @4:42 Spawned: Overlord+1 @4:43 Spawned: Hatchery+1 @4:47 Spawned: Larva+1 @4:47 M:453 G:0 L:3 S:26/38 BuildDrone @4:47 M:403 G:0 L:2 S:27/38 BuildQueen @4:47 M:253 G:0 L:2 S:29/38 BuildDrone @4:47 M:203 G:0 L:1 S:30/38 BuildDrone @4:49 Spawned: Drone+1 @4:49 Spawned: Drone+1 @5:02 Spawned: Larva+1 @5:02 M:389 G:0 L:3 S:31/38 BuildDrone @5:02 M:339 G:0 L:2 S:32/38 BuildDrone @5:02 M:289 G:0 L:1 S:33/38 BuildDrone @5:04 Spawned: Drone+1 @5:04 Spawned: Drone+1 @5:04 Spawned: Drone+1 @5:17 Spawned: Larva+1 @5:17 M:506 G:0 L:3 S:34/38 BuildDrone @5:17 M:456 G:0 L:2 S:35/38 BuildDrone @5:17 M:406 G:0 L:1 S:36/38 BuildDrone @5:19 Spawned: Drone+1 @5:19 Spawned: Drone+1 @5:19 Spawned: Drone+1 @5:22 Spawned: Queen+1 @5:32 Spawned: Larva+1 @5:32 M:657 G:0 L:3 S:37/38 BuildDrone @5:32 M:607 G:0 L:2 S:38/38 ExtractorTrick @5:32 M:582 G:0 L:2 S:37/38 BuildDrone @5:32 M:532 G:0 L:1 S:38/38 ExtractorTrick @5:32 M:507 G:0 L:1 S:37/38 BuildDrone @5:34 Spawned: Drone+1 @5:34 Spawned: Drone+1 @5:34 Spawned: Drone+1 @5:34 Spawned: Extractor Trick Finished, Drone Restored @5:34 Spawned: Extractor Trick Finished, Drone Restored @5:37 Spawned: Queen+1 @5:47 Spawned: Larva+1 @5:49 Spawned: Drone+1 @5:49 Spawned: Drone+1 @5:49 Spawned: Drone+1 -------Goal------- Drones: 40 Queens: 2 Bases: 2 At time: 6:00 Minerals: 1146 Gas: 0 Supply: 40/38 Larva: 3 Drones: 36 Overlords: 4 Queens: 2 Bases: 1 Spawning Pools: 1 Sun Nov 28 09:45:54 CST 2010: 2111.146
The fact that this build ends with 4 overlords and leaves you supply blocked pretty thoroughly discredits it as the most economical build. Try this with a longer time frame than 6 minutes and then provide the order for a more reasonable assessment.
|
On November 29 2010 05:33 Hurkyl wrote:Show nested quote +On November 29 2010 05:26 jdseemoreglass wrote:On November 29 2010 04:33 Hurkyl wrote:
Also, it is relevant to note where in the larvae cycle the hatcheries are, as well as the spawn larvae. There's a big difference between having the next spawned larvae drop at 6:01 and 6:39! This is an important point that I'm glad someone mentioned... However, I feel these numbers are less relevant toward overall economy than drone count (including production) and minerals mined. If two builds are very similar in these regards, then the "larva in production" data would also be factored in. The problem with ignoring this data is that an inferior build would look better than a superior build if you sample just as the inferior build gets its benefit from spawn larvae -- this can be up to an 8 drone difference! If you're looking at supply, for example, the current leader gets a lot of benefit from this effect since a spawn larva finishes just before 6:00, giving it a temporarily inflated supply count.
Hurkyl wrote:
Edit: I can't figure out how you're counting drones or minerals. The above number for drones is what's actually out and mining at that second.
Note that I am not basing my data on supply or simple current drone count. I stated that all units in production will be prorated based upon their degree of completion. For example, if a drone is half done, it counts as .5 drones. If two drones are half done, it counts as 1 drone.
This would prevent the recently created drones from adding 8 supply, because the drones would only be about 5% complete and therefore would only equate to (5% * 8) = 40% of a drone.
Hope this clarifies things.
|
On November 29 2010 05:33 Hurkyl wrote:
The problem with ignoring this data is that an inferior build would look better than a superior build if you sample just as the inferior build gets its benefit from spawn larvae -- this can be up to an 8 drone difference!
If you're looking at supply, for example, the current leader gets a lot of benefit from this effect since a spawn larva finishes just before 6:00, giving it a temporarily inflated supply count.
This is of course true and a big problem.
A question - do you actually know how the replay 'spendings'(spent? I have a different language version) tab works? In the winners replay it shows 4100 total resources spent, which is bollocks, since if you count them yourself you get 3400 minerals spent. -> the tab shouldn't be trusted unless someone knows whats up.
Also, the 'leader' has 40 drones at 6 minutes, not 41.
-edit- Ahhh, I see. It counts the starting hatch+6xdrones+overlord in. Anyway, it's important to state whether someone uses the tab values or real ones.
|
Using this methodology rather than optimizers (what you dismiss as "theorycrafting") introduces human error into the equation. Your initiative to employ the scientific method is admirable, but the reality is that reductionistic approaches (i.e., the genetic algorithm that has become so popular these days) are going to be far more accurate in this very limited sphere than any method whose results depend on perfectly consistent performance from both the player as well as the network. Since neither is possible, you are in effect introducing hundreds, if not thousands, of variables in each game that you play.
|
On November 29 2010 05:46 FlayedOne wrote:Show nested quote +On November 29 2010 05:33 Hurkyl wrote:
The problem with ignoring this data is that an inferior build would look better than a superior build if you sample just as the inferior build gets its benefit from spawn larvae -- this can be up to an 8 drone difference!
If you're looking at supply, for example, the current leader gets a lot of benefit from this effect since a spawn larva finishes just before 6:00, giving it a temporarily inflated supply count. This is of course true and a big problem. A question - do you actually know how the replay 'spendings'(spent? I have a different language version) tab works? In the winners replay it shows 4100 total resources spent, which is bollocks, since if you count them yourself you get 3400 minerals spent. -> the tab shouldn't be trusted unless someone knows whats up. Also, the 'leader' has 40 drones at 6 minutes, not 41.
Note that at replay start, the game counts the initial starting units at 750 minerals spent. Perhaps I should subtract this from the totals, but it will be the same for every game and will only change the relative ratios of one spending to another instead of total spending.
Also note I included drones in production to reach 41 in the replay provided.
|
On November 29 2010 05:43 jdseemoreglass wrote:Show nested quote +On November 29 2010 05:33 Hurkyl wrote:On November 29 2010 05:26 jdseemoreglass wrote:On November 29 2010 04:33 Hurkyl wrote:
Also, it is relevant to note where in the larvae cycle the hatcheries are, as well as the spawn larvae. There's a big difference between having the next spawned larvae drop at 6:01 and 6:39! This is an important point that I'm glad someone mentioned... However, I feel these numbers are less relevant toward overall economy than drone count (including production) and minerals mined. If two builds are very similar in these regards, then the "larva in production" data would also be factored in. The problem with ignoring this data is that an inferior build would look better than a superior build if you sample just as the inferior build gets its benefit from spawn larvae -- this can be up to an 8 drone difference! If you're looking at supply, for example, the current leader gets a lot of benefit from this effect since a spawn larva finishes just before 6:00, giving it a temporarily inflated supply count. Show nested quote + Hurkyl wrote:
Edit: I can't figure out how you're counting drones or minerals. The above number for drones is what's actually out and mining at that second. Note that I am not basing my data on supply or simple current drone count. I stated that all units in production will be prorated based upon their degree of completion. For example, if a drone is half done, it counts as .5 drones. If two drones are half done, it counts as 1 drone. This would prevent the recently created drones from adding 8 supply, because the drones would only be about 5% complete and therefore would only equate to (5% * 8) = 40% of a drone. Hope this clarifies things.
This still works out to a little over a 4 drone error. While the difference between 20/40 and 0/40 is 2 larvae, you count it as 4 (because the 20/40 just had its 4 drones pop)
|
On November 29 2010 05:49 Voros wrote:
Using this methodology rather than optimizers (what you dismiss as "theorycrafting") introduces human error into the equation. Your initiative to employ the scientific method is admirable, but the reality is that reductionistic approaches (i.e., the genetic algorithm that has become so popular these days) are going to be far more accurate in this very limited sphere than any method whose results depend on perfectly consistent performance from both the player as well as the network. Since neither is possible, you are in effect introducing hundreds, if not thousands, of variables in each game that you play.
If the builds were close enough that only tens of minerals or fractions of a drone were the determining factor then I would agree with you. So far this has not been the case and the differences between builds have been wide enough to exclude the very minor effects of human error or variability.
Also, if anyone is able to beat the results I have provided on any build, I will be willing to provide them in place of my own.
|
I'm surprised the classic 14 hatch 15 pool isn't in the list of things to compare against. My results are:
39 drones
one done at 14/17
one drone at 0/17
one drone at 11/17
6 overlords
One hatchery at 40/40 in the spawn cycle
One hatchery at 0/40 in the spawn cycle, with 4 larvae in the air
764 minerals
Incidentally, anti-optimizing (rallying all drones to the same patch that's a bad one) drones feels like it makes a big difference. The build didn't flow as smoothly as when you do normal things with your drones.
|
Overpool, no extractor trick
Queen at 15
Overlord at 17
Hatch at 18
Queen at 20
After second queen spit, go to second hatch
Overlord at 23
Overlord at 30
At this point, I maynarded 5 drones over, probably should have been more.
Overlord at 32
Overlord at 35
At 6minutes
43 drones
6 overlords
2 queens
345 in the bank
2 drones in production, just started, putting me at 49/52 supply.
![[image loading]](http://www.gamereplays.org/community/uploads/repimgs/repimg-33-169752.jpg)
Edit: The retarded thing is, this is almost the 7RR opening, making drones/expansion instead of double-OL and roach warren around 18. Tricky!
|
12 pool FE:
10 Trick
11 Overlord
12 Spawning Pool
16 Queen, send drone to expo
18 Hatchery
18 Overlord
18 Trick
19 Queen
23 Overlord
(After spawn larvae, set rally to expo and transfer 4 drones. Also send second queen to expo)
29 Overlord
36 Overlord
39 Reset main rally point to main
This gets
41 drones
4 drone eggs at 8/17
1 drone egg at 6/17
1 drone egg at 3/17
6 overlords
Spawn cycles are at 27/40 and 9/40 (but I was slacking)
498 minerals in the bank
(I really should put another overlord in there, but it's not relevant by 6:00)
|
isnt this conclusion a bit obvious...? hatch first will mean better eco ofc... unless u dont use the extra larva for drones
|
On November 29 2010 06:54 donut boi wrote:
isnt this conclusion a bit obvious...?
Only if you forget about queens.
|
On November 29 2010 05:58 Hurkyl wrote:Show nested quote +On November 29 2010 05:43 jdseemoreglass wrote:On November 29 2010 05:33 Hurkyl wrote:On November 29 2010 05:26 jdseemoreglass wrote:On November 29 2010 04:33 Hurkyl wrote:
Also, it is relevant to note where in the larvae cycle the hatcheries are, as well as the spawn larvae. There's a big difference between having the next spawned larvae drop at 6:01 and 6:39! This is an important point that I'm glad someone mentioned... However, I feel these numbers are less relevant toward overall economy than drone count (including production) and minerals mined. If two builds are very similar in these regards, then the "larva in production" data would also be factored in. The problem with ignoring this data is that an inferior build would look better than a superior build if you sample just as the inferior build gets its benefit from spawn larvae -- this can be up to an 8 drone difference! If you're looking at supply, for example, the current leader gets a lot of benefit from this effect since a spawn larva finishes just before 6:00, giving it a temporarily inflated supply count. Hurkyl wrote:
Edit: I can't figure out how you're counting drones or minerals. The above number for drones is what's actually out and mining at that second. Note that I am not basing my data on supply or simple current drone count. I stated that all units in production will be prorated based upon their degree of completion. For example, if a drone is half done, it counts as .5 drones. If two drones are half done, it counts as 1 drone. This would prevent the recently created drones from adding 8 supply, because the drones would only be about 5% complete and therefore would only equate to (5% * 8) = 40% of a drone. Hope this clarifies things. This still works out to a little over a 4 drone error. While the difference between 20/40 and 0/40 is 2 larvae, you count it as 4 (because the 20/40 just had its 4 drones pop)
Also, even ignoring spawn larvae, your method of counting drones has the same flaw. After 17 seconds, there is no difference between a drone egg that has just started, and a drone egg that's almost hatched. With your reckoning, your drone count at 6:00 does not directly correlate to your drone count at 6:17!
Said differently, the only difference between a 1/17 drone egg and a 16/17 drone egg is a flat 5-10 minerals, depending on how saturated your base is. After 17 seconds, both are harvesting minerals at the same rate.
The difference between 1 drone and 0 drones is that the former harvests an extra 20-40 minerals every minute, depending on saturation.
The two situations are very different, yet your method of counting production by drones + fractional drones treats them as roughly the same.  In nearly every circumstance, I would rather have two drone eggs at 1/17 than 1 drone and no egg.
This is why I have always measured production via supply and position in the larva cycles.
|
While I think the purpose of this makes sense, and is clearly defined. Whether the most economical or not, is there any weight to whether the build is viable in game? IE in TvZ 16 hatch 15 pool sounds awfullly awfully risky.
so my question would be have you actually game tested doing this?
|
On November 29 2010 07:13 Hurkyl wrote:Show nested quote +On November 29 2010 05:58 Hurkyl wrote:On November 29 2010 05:43 jdseemoreglass wrote:On November 29 2010 05:33 Hurkyl wrote:On November 29 2010 05:26 jdseemoreglass wrote:On November 29 2010 04:33 Hurkyl wrote:
Also, it is relevant to note where in the larvae cycle the hatcheries are, as well as the spawn larvae. There's a big difference between having the next spawned larvae drop at 6:01 and 6:39! This is an important point that I'm glad someone mentioned... However, I feel these numbers are less relevant toward overall economy than drone count (including production) and minerals mined. If two builds are very similar in these regards, then the "larva in production" data would also be factored in. The problem with ignoring this data is that an inferior build would look better than a superior build if you sample just as the inferior build gets its benefit from spawn larvae -- this can be up to an 8 drone difference! If you're looking at supply, for example, the current leader gets a lot of benefit from this effect since a spawn larva finishes just before 6:00, giving it a temporarily inflated supply count. Hurkyl wrote:
Edit: I can't figure out how you're counting drones or minerals. The above number for drones is what's actually out and mining at that second. Note that I am not basing my data on supply or simple current drone count. I stated that all units in production will be prorated based upon their degree of completion. For example, if a drone is half done, it counts as .5 drones. If two drones are half done, it counts as 1 drone. This would prevent the recently created drones from adding 8 supply, because the drones would only be about 5% complete and therefore would only equate to (5% * 8) = 40% of a drone. Hope this clarifies things. This still works out to a little over a 4 drone error. While the difference between 20/40 and 0/40 is 2 larvae, you count it as 4 (because the 20/40 just had its 4 drones pop) Also, even ignoring spawn larvae, your method of counting drones has the same flaw. After 17 seconds, there is no difference between a drone egg that has just started, and a drone egg that's almost hatched. With your reckoning, your drone count at 6:00 does not directly correlate to your drone count at 6:17! Said differently, the only difference between a 1/17 drone egg and a 16/17 drone egg is a flat 5-10 minerals, depending on how saturated your base is. After 17 seconds, both are harvesting minerals at the same rate. The difference between 1 drone and 0 drones is that the former harvests an extra 20-40 minerals every minute, depending on saturation. The two situations are very different, yet your method of counting production by drones + fractional drones treats them as roughly the same. In nearly every circumstance, I would rather have two drone eggs at 1/17 than 1 drone and no egg. This is why I have always measured production via supply and position in the larva cycles.
The difference between a 1/17 drone and a 16/17 drone is clearly 15 SECONDS. Those have to be weighted more when measuring the prorated value of a drone in production.
And clearly 1 finished drone is superior to two drones at 1/17... I have no idea what you are trying to get at here. Please stop filling up the thread with this convoluted logic, as it is not even relevant to any stated build or situation yet encountered.
|
|
|
|
|
|
EPT
