EPT
[G] TheStaircase - An Alternative Improvement Method - Pag…
| Forum Index > StarCraft 2 Strategy |
|
Orbry
Russian Federation3 Posts
| ||
|
JaKaTaKSc2
United States2787 Posts
 | ||
|
fengshaun
149 Posts
----------- NOTE: this method of keeping track of supply requires you to check your larvae constantly and react as soon as they pop. Please see the thread on TheCore (hotkey layout by TheJak himself) where spaceboy talked about a method to keep track of larva inject timing. This method can also be used to keep track of your supply. http://www.teamliquid.net/forum/viewmessage.php?topic_id=341878¤tpage=155#3097 read until the topic changes!  ------------ 1 hatch produces 1 larva per 15s. 1 queen produces 4 larvae per 40s. 1 hatch can produce larvae until 3 or more larvae are present. Let N be the number of hatcheries you have, assuming every hatchery has an injecting queen. Let L be the total number of larvae. Let x be the number of new ovies needed. Let t be the time from the last inject. Assuming you use all of your larvae and produce on average 2-supply units, you'll need about 2*L supply. During 1 production cycle: at t=0s: 0 larvae (you have just used up all your larvae. at t=15s: N larvae at t=30s: 2N larvae at t=40s: 2N + 4N larvae = 6N larvae I'm thinking based on a 2-supply army, you'll need about N+(N/2) ovies giving you 8*(N+(N/2)) = 8 * 3/2N = 12N supply. If you use N+(N/2) larvae, you're left with about 6N-(N+N/2)=4.5N larvae making a supply of 9N. Okay, that's too much supply. Let's try N/3: 8 * (N+N/3) = 8 * 4/3N = 10.666N = 10.7N supply, leaving 6N-(N+N/3)=4.666N=4.7N larvae which makes a supply of 2*4.7N=9.4N supply. That's *almost* perfect. You make 10.7N supply right before a production cycle and then make 9.4N supply, giving you some wiggle room. How can you use this information? Overlord production time: 25s at t=0: used all larvae and no additional supply. at t=15: N larvae, produce N ovies = 8N supply. at t=30: use about 1/3 of larvae to produce ovies = 8/3N supply at t=40: larvae pop: you have 8N additional supply to make units (1st set of ovies pop) at t=55: additional 8/3N supply (2nd set of ovies pop). This can be used for the *next* production cycle. If you continue this path, you should, in thought, never get supply blocked, although you will almost always have 8/3N additional supply that you can't use because you don't have enough larvae. I haven't thought a way around that. There simply isn't enough time for all ovies to pop in 1 production cycle, so you have use some larvae from 1 production cycle to build ovies for the next cycle. This is not optimized, of course. I've been focusing exclusively on supply effectiveness for the past week and this seems to be a good pattern to get used to (something like the idle SCV trick that terrans use, someone might be able to write the details). note: if you really want to get better as you're getting used to this pattern, try forcing yourself to look at your supply whenever you make the ovies. This way, you'll know when you don't need any more ovies because you just 1A'd lings into a siege line! :D | ||
|
Ninjury_J
Canada408 Posts
| ||
|
fengshaun
149 Posts
| ||
|
fengshaun
149 Posts
In other words, is like this? pop larva -> 15s -> pop larva -> 15s -> pop larva -> limit reached, no more larvae production -> use 1 larva -> 15s -> pop larva. Or like this? pop larva 3 times -> 15s -> limit reached, no more larvae production -> use 1 larva -> pop larva (the one that was waiting) I assumed the former, but if it's the latter, that might actually make it easier to keep track of ovies! (I would have checked it, but I don't have access to SC2 right now. I'll report back in a few hours) | ||
|
Ninjury_J
Canada408 Posts
To be more explicit, every 15 seconds, one larva is generated. To generate 3, it takes 45 seconds. An inject takes 40 seconds, so in that cycle, two larva are spawned (at t=15 and t=30). By injecting on the second cycle, does that reset the natural regeneration to 0? I'm pretty sure it works like the first way you suggest above. | ||
|
JDub
United States976 Posts
Edit: to summarize: On March 25 2012 14:25 Jonoman92 wrote: It took me a really long time to figure out what the OP was saying. I think an easier way of explaining it might be: 1. hatchery larvae generation is on a time counter 2. when the hatchery has 3 or more larvae this timer is paused 3. for some odd reason if the spawn larvae pops at the same time as the larvae generation then the timer is reset, when it should actually pop the larvae as soon and there are fewer than 3 larvae remaining, hence robbing the z player of a larvae they should have | ||
|
Ninjury_J
Canada408 Posts
| ||
|
fengshaun
149 Posts
On March 29 2013 07:15 Ninjury_J wrote: Wait, so by injecting consistently, you "lose" the third larva generated naturally by the hatch? Potentially, at least according to http://wiki.teamliquid.net/starcraft2/Spawn_Larva Oh JDub, I've been looking for something like that for so long! Thanks a ton! I'll revise my post above after I learn a bit more! | ||
|
Ninjury_J
Canada408 Posts
For example, I inject at 0. L1 spawns at 15. L2 spawns at 30. At 40 my inject pops and I reinject. Will my next larva spawn in 5 seconds or in 15? | ||
|
fengshaun
149 Posts
t=0s inject t=15s N larvae pop t=30s N larvae pop t=40s 4N new larvae pop (6N larvae total), timer pauses, 5s remaining. Use larvae, Inject. t=45s N larvae pop (7N larvae total) So the next cycle and susequent ones will be: t=0s inject (previous t=40s) t=5s N larvae pop t=20s N larvae pop t=35s N larvae pop t=40s 4N new larvae pop, timer pauses, 10s remaining. Use larvae, inject. next cycle: t=0s inject (previous 40s) t=10s N larvae pop t=25s N larvae pop t=40s 4N new larvae pop + N larvae pop. timer pauses, 15s remaining. Use larvae, inject. next cycle: same as the first cycle. total: 7N larvae per cycle So, you can use the third and first set of popped larvae to make ovies, and they'll be ready for the corresponding production cycle. Also, you won't lose any larva production. *The timer is approximate. According to that post, there is a 1-1.5s animation delay when larvae pop, but that shouldn't change the overall layout. The point is only to make a general layout of when to make ovies until you get used to looking at supply so you can go on autopilot! EDIT: fixed second cycle timings, who would have thought I couldn't add 20 and 15! Thanks Ninjury! | ||
|
Ninjury_J
Canada408 Posts
So inject one: 15, 30 Inject two: 5, 20, 35 Inject three: 10, 25, 40 Full reset? | ||
|
fengshaun
149 Posts
| ||
|
Ninjury_J
Canada408 Posts
| ||
|
fengshaun
149 Posts
| ||
|
Ninjury_J
Canada408 Posts
What is the Terran trick? I am unfamiliar. | ||
|
fengshaun
149 Posts
Probes and pylons is too general, there is still a very good chance one might forget the pylon part! It's not as autopilot as the other two races! | ||
|
JaKaTaKSc2
United States2787 Posts
| ||
|
fengshaun
149 Posts
A method to keep track of supply: I believe the replay has contains the information about supply, so it should be possible to list when supply = max supply and say how many seconds the player was supply blocked. If 25s or more, then it's a hard supply block. If less, then it's a soft supply block. Then, we can easily go back to that time in replay and see why we were supply blocked! For inject timings, maybe add a graph for all Queens' energy that goes up to 200, and make it "hoverable" (is that even a word?) to say what time the player missed an inject and the queen went above 25. The current block graph for inject timing is really brief, you can't tell when exactly you missed the inject! The reason I want there to be these time information is that with this extra information, you can just go to that time in the top big graph and see exactly how the map looked like. There is a very good chance you'll remember what you were doing at that time. The inject timing modification is not really necessary. I look at each replay right after playing a game, so I'll just check my inject timings then. But checking supply blockage is really hard, and the "time supply blocked" is of no use (although I've recently been hovering at around 1:00 rather than 2:00 or even 3:00, which is a big improvement!). Right now, it's really easy to keep track of both inject timings (pretty obvious from watching replay at 8x speed) and spending skill (ggtracker). In the same vein, I would really like an efficient an easy way to keep track of supply blockage so I have a better idea of how to make it better! | ||
| ||
