EPTWatch the replay, it just feels weird. Maybe if that problem was fixed, it would catch up.
Putting an end to the Zerg Opening Economy debate. - Page 11
| Forum Index > StarCraft 2 Strategy |
|
Lomilar
United States130 Posts
Watch the replay, it just feels weird. Maybe if that problem was fixed, it would catch up. | ||
|
Lomilar
United States130 Posts
A larva pop from a queen resets the spawn larva timer on the hatchery. So... say you have 0 larva, and are at 11/15 seconds on the invisible larva timer. When the spawn larva pops, you go to 1 larva, 2 larva, 3 larva (spawn larva timer turns off), 4 larva... then you spend them back to 0, and your larva timer is back at 0/15. Holy crap. That is kinda... not huge, but 1 larva per spawn larva big. That means that with epic hand speed, you can potentially save your half built larva, by spending them as they pop off one at a time and never letting the larva count get to 3. No wonder spawn larva is 4 larva. If it was 3 larva, it would not be as good as a hatchery, because it would reset the spawn larva timer without giving you that extra larva in the invisible build queue. I wonder if this is actually accurate. It sounds plausible. That would mean that watching your larva spawn and spending them down can net you an extra part of a larva. It would also mean that at super slow speed doing these precise builds, there is a chance of losing a larva, making it much much harder to test. | ||
|
Douillos
France3195 Posts
Incredible work by OP & Gratz to Lomilar for your build, i really want to try it in zvz! (mine was somewhat logical, i didnt really need to think that much :p ) | ||
|
Douillos
France3195 Posts
On November 30 2010 16:47 Hurkyl wrote: On a slight tangent, I admit it would be nice to have an optimized triple hatch build prepared for dealing with 15 Nexus openings on Shakuras Plateau.  OFF TOPIC: Use 16hatch / 15 pool Just get no gas after the 2 queens and put down your 3rd hatch when you have no larva left and 300 minerals (produce only drones and overlord when your 2 under supply cap). It actually goes ridiculously fast. But really the idea is no gas, and drone pumping till your at 300 mins. | ||
|
Hurkyl
304 Posts
On November 30 2010 18:21 Lomilar wrote: I just thought of something else, so I will chain post. A larva pop from a queen resets the spawn larva timer on the hatchery. Have you tested this in game? My casual observation suggests that it merely pauses the timer, but I haven't really tried to test it directly. | ||
|
Cambam
United States360 Posts
However, I would love to see some TLer more creative than me figure out a way to factor in things like worker/ebay/pylon blocks, as well as cannon rushes and 2 rax marines+scvs. One thing I thought of is maybe someone could make BO tester map that has a probe come over and build a pylon at a certain time. Then this map could be distributed on this thread and everyone could compare results of different builds against it. It'd be tough though because the timing is map dependent... But that's where many heads are better than one. My goal is to find two builds: the absolute most economical build (for when I know I'm not going to be attacked for a while [e.g. 15 nexus]), and then the absolute most economical build while still being safe from early aggression (from 2 marines to 4 gate allin). Pretty ambitious, but let's work towards it! | ||
|
Skrag
United States643 Posts
On November 30 2010 18:21 Lomilar wrote: I just thought of something else, so I will chain post. A larva pop from a queen resets the spawn larva timer on the hatchery. Are you absolutely 100% sure about this? I never managed to figure out anything conclusive, but I did some testing that seemed to indicate that larvae spawned from the queen might *pause* the timer without resetting it. Sounds weird, and eventually I gave up trying to figure out what was going on, but that seemed like the most likely explanation for what I was seeing. That would mean that watching your larva spawn and spending them down can net you an extra part of a larva. I *believe* this is true, and that if you spend the first couple larva as they're coming in, you can avoid any negative effect on the hatchery's normal larva spawn. It's pretty hard to do reliably and consistently, though, especially since you need to be spitting a new round at about the same time. | ||
|
Lomilar
United States130 Posts
A larva takes 15 seconds to spawn. Check. 3:11 I spend 3/3 larva 3:13 I spew forth larva upon thine hatchery 3:26 a larva spawns, I spend him. 3:41 a larva spawns. This one is a girl. I spend her. 3:53, larva pops. I spend them as they fall (in super slow mo) 3:57, a larva spawns. Mystery? Are you a second late? Why larva! Why! Second act! 4:10, I spew forth. 4:12, a larva spawns. Eggified. 4:27, a larva spawned. It is no more. 4:42, a larva spawned. Drone. 4:50, the larva pop. I wait for 3 to drop, then spend them all, perhaps even before the 4th one. 4:59, a larva spawned. Did it not stop the timer? Why 17 seconds? It is 2am. The mystery continues. <edit> Huh. Pause. At what point does it resume, and what point does it say hey, you, no more larva for 15 seconds. Does, in fact, spawning larva somehow hold a hatchery in stasis? Pausing the previous larva, as if the larva pylon became unpowered? </edit> | ||
|
Hurkyl
304 Posts
Also, note this hypothesis is consistent with ordinary Hatchery operation -- when the third Larvae pops, the timer is reset to 15 seconds and pauses. When you spend one, the timer starts ticking again, down from 15. I suspect you can't avoid a one second delay, since I think the Hatchery icon immediately shows the number 4, but you can't actually start selecting the larvae until they land. | ||
|
Lomilar
United States130 Posts
| ||
|
Skrag
United States643 Posts
On November 30 2010 18:54 Hurkyl wrote: I suspect you can't avoid a one second delay, since I think the Hatchery icon immediately shows the number 4, but you can't actually start selecting the larvae until they land. This appears to be true if you're looking at multiple hatcheries in a single group, but if you have the hatchery selected alone, you'll see the number go up incrementally, and you can start using larvae as soon as the number goes up. And like I said, my (very inconclusive) testing seemed to indicate that you could avoid any delay if you were *very* quick about using the first two larvae, and the hatchery itself didn't spawn one of its own at a bad time. Also, Lomilar, I *really* wouldn't worry about a 1 second timing difference unless that difference showed up consistently over a number of tests. Unless you're working on a custom map that's printing out exact timings (I think the highest resolution you can get is 1/256th of a second), using the game timer with a resolution of 1 second means that your margin of error is pretty big. | ||
|
Skrag
United States643 Posts
| ||
|
blackodd
Sweden451 Posts
edit: Lominar -> Lomilar Regarding OPs question of worker worth to minerals, I'd say we all agree on that 1 drone is worth more than 50 minerals. There is larva cost and spawning time, also the ability to make a building and attack, defend, which 50 minerals do not have! (lol) Set 1 drone = 75 minerals and see what happens? | ||
|
Lomilar
United States130 Posts
And I heart overpool. I want to see aggressive zergs. | ||
|
Douillos
France3195 Posts
On November 30 2010 18:39 Cambam wrote: My goal is to find two builds: the absolute most economical build (for when I know I'm not going to be attacked for a while [e.g. 15 nexus]), and then the absolute most economical build while still being safe from early aggression (from 2 marines to 4 gate allin). Pretty ambitious, but let's work towards it! Check OP spoilers: you have the 2 builds already... | ||
|
puissance
97 Posts
Concerning both leading builds, when would you put down your gas? I guess it is pretty variable, depending on the followup you are aiming for. My general rule is that I want my gas very close before or after I put down the pool if I want to go for speedlings and in case of Roaches I want to put it down after my hatch (or pool if I go hatch first). Some input would be appreciated | ||
|
blackodd
Sweden451 Posts
On November 30 2010 19:47 puissance wrote: I relly like this thread - great job! Concerning both leading builds, when would you put down your gas? I guess it is pretty variable, depending on the followup you are aiming for. My general rule is that I want my gas very close before or after I put down the pool if I want to go for speedlings and in case of Roaches I want to put it down after my hatch (or pool if I go hatch first). Some input would be appreciated In a real game I generally put my gas after the 18 hatch, if I play against protoss or a hellion terran I do 11 pool then extractor | ||
|
[F_]aths
Germany3947 Posts
| ||
|
michaelhasanalias
Korea (South)1231 Posts
On November 30 2010 19:05 Skrag wrote: btw, aren't you all glad we're finally putting and end to the zerg opening economy debate, once and for all? I feel like I should quote my own post from 2 pages ago that no one responded to  .There's a big question on how to define the question, but a simple graph of these openings and total resource count would aid in both asking and answering the ultimate question here: What is the most economical build until time T=t, the transition point to an actual build? (2-base saturation is about 6:30, so the longest this should go is ~6:30, not the arbitrarily defined 6 minutes.) If at any time any ONE build has the most minerals at any point between 0 and 6 minutes, that build should be counted. Because eventually you transition beyond pool, queen and workers.... and at THAT point, whichever opener is in the lead is the one you need to use for your build. If you make nothing until 6 minutes, then whichever has the most at 6 minutes is the most economical. What's being done right now isn't practical whatsoever because no reasonable build (at least in one player) only makes workers for the first 6 minutes. And insofar as there is one, it is certainly very rarely executed. If you drone until 3:30 then drop a warren, you need to know what's going to set you up for the most possible minerals until 3:30, and then THAT is the most economical build. There is no ultimate opening because the terminal transition time varies depending on what you decide to make. However, there IS an ultimate answer in that at any point in time T, there is some build X that is definitively most efficient. THIS question is easily answerable with the current replay archive that is available. In other words, the data already collected should be enough to answer this question. One graph with these builds and their total mineral count as a function of time would actually, definitively end this debate. And I think out to ~6:30 is sufficient (The time it takes to make 60 drones on 2 bases with 2 queens according to evo chamber.) AAAAnyway... carry on. | ||
|
k43r
Poland112 Posts
And when is the earliest time when you can make zergling speed not cutting drones. Or function when do you cut drones - how many minerals you loose - time of finishing speed. Or when you cna build economicly Roach Den, and graph how many minerals you loose if you build it safe. <mind blown> | ||
| ||
