My 2-gate Nonsense (with funny pictures!)

WinterNightz

United States111 Posts

February 22 2011 00:57 GMT

I was super pumped by the amount of responses I got with my previous blog-post, so now that I've thought about things more, I'm going to be doing a bit more of an analytical level of 2-gate-core vs. 1-gate-core.

Here is what we know to be true:

In order to fully chrono-boost warp-tech, you need approximately 75 energy on your nexus when research begins.
With a basic 1-gate-core (13-gate, 15-gas, core a.s.a.p.), constantly chronoboosted warp-gate completes at 5:15~5:25.
Warp-tech decreases production time (for non-templar units) by approximately 25%, increasing production capacity (units produced/time) by a third.
Warp-tech itself requires 140 seconds to complete, or ~94 seconds when fully chrono-boosted. This means that for a perfectly executed 4-gate, with all the final gateways completing the instant warp-tech completes, you need to place the gateways when warp-tech is 30%~55% complete (depending on how much you're chronoboosting warp-tech.

The build I'm working on: 2-Gate Core
(definitely not-yet-optimized)
+ Show Spoiler +

9 -- Pylon
12 -- Gateway
14 -- Gateway
15 -- Pylon (2 chronoboosts on probes)
16 -- Gas
17 -- Cybernetics Core

Features of 2-gate Core:

Warp-tech completed at 5:40. 15~25 seconds later than a normal 4-gate.
Earlier 2nd gateway means stronger early army production.
In order to get out the second gateway so early, probes are NOT chronoboosted until the second pylon begins (which means nexus energy will get prettttty damn high).
Don't worry, probes are still constantly produced. They're just not super-boosted from the beginning.

So here's the crux of it all. This is what I want to be true: it is my hope that the early 2nd gateway before the core provides enough early additional production capacity that the 15~25 seconds of time where 1GC has warp-tech faster is overcome. This is also the part that requires much more work/testing. But for now, we theorycraft!

Let's say our difference in tech timing (15~25 game seconds) is t.

So (after everyone's finished warp-tech) 1-gate-core has t more seconds of 4-warp-gate production capacity than 2-gate-core.
Since warp-gates increase production capacity by 1/3, that means it's 4/3 * s * 4 "gate-seconds" that 2-gate core loses out on. Using 15~25 seconds as s, that is equivalent to 80~133 seconds of a normal gateway cranking away at making dudes.

This is probably the point where people eyes start rolling back into their heads, so here's some amusing pictures!
+ Show Spoiler +

So basically, if 2-gate-core can run with 80~133 more gate-seconds (number of gateways at work * time) before warp-gate finishes, then it can have equivalent units to a 4-gate! (and thus be able to fend it off, theoretically.)

This is basically where the rubber hits the road and some more field-research needs to be done. Neither build has gateways constantly producing, 'cause you need to save up money to throw down the final warp gates.

RESEARCH PLAN:

Run through each build, constantly producing probes, up until the point when they're normally cut (24 mining minerals? Analysis of Macro. I have no idea! I may play protoss, but I've never actually 4-gated. XD)
Chronoboost warp-tech 100%, throwing down remaining gateways in time to complete as soon as warp-tech is done.
Produce NO ARMY UNITS. Obviously the money will end up being pretty high, but we will be able to assume after the fact that all of this money would have been spent on units and the pylons to support them.
Compare the amount of money assumed to go to army units with the amount of time we actually could have been making units. Obviously a single gateway cannot spend 8million minerals and 4million gas over the course of 2 minutes. So we find which of these numbers (gateways*time, or minerals+gas) are our limiting factor, and say "that is how much stuff this build can make before warp-gates get out!)

So my only real question right now is, when do people generally cut probes with a 4-gate? Also, I fully realize that this might end up being completely fruitless, but it's SCIENCE! >:D

WinterNightz

United States111 Posts

February 22 2011 00:58 GMT

1-gate-core:

24 probes total (stopped production at 22 in order to get up 3 remaining gates in time). 3 on gas.
Upon warping in 4 stalkers produced as soon as warp-gates are transformed, we have:
415 minerals
94 gas
32/42 food. (i.e. could have one of those pylons be added to the mineral count)
First point at which the there's actually enough money to build anything is 3:10.
That means from 3:10 to 5:25 (135 game seconds), we have 1 gateway to use up ~515/100 worth of resources.

250/100 into 2 stalkers = 84 seconds.
200/0 into 2 zealots = 76 seconds.
Total: 450/100 resources (and 8 food) and 160 game seconds, meaning we can really only get out 3 units in time while still executing a smooth 4-gate. NOTE: this is only an upper bound on what is possible. I don't know that it's actually possible. The army investment might interfere with putting down gateways, but probably not.

2-gate-core:

24 probes total. never stopped production. 3 on gas.
Upon warping in 4 stalkers produced as soon as warp-gates are transformed, we have:
610 minerals
38 gas
32/34 food.
First point at which the there's actually enough money to build anything is 3:25.
That means from 3:25 to 5:25 (120 game seconds), we have 2 gateways to use up ~600/50 worth of resources.

125/50 into 1 stalker = 42 seconds.
300/0 into 3 zealots = 114 seconds.
100 into 1 pylon
Total: 525/50 resources and 156 "gateway seconds", to be spent over 120 game seconds. Since we have 2 gateways, that's 240 possible gateway-seconds, meaning we only need our gateways producing 65% of the time in order to spend all that money and get every single unit out.

SCIENCE: Apparently I seriously screwed up chrono-boosting in the first example, because both replays finish warp-tech at 5:25. But one of them required me to cut probes. This is interesting, but probably due to the fact that I suck at this game. D: It also really shows how similar the end-result can be, if you chrono-boost properly with 2GC, or if you DON'T chronoboost properly with 1GC.

Obviously all match-ups work differently, but the reason I compared to 4-gate is because that's just the giant hammer falling on your head. "If the build can't defend a 4-gate, then it sucks!". The thing about this though is that it should theoretically have enough early army production to fend off massive barracks pressure (without the need for sentries), and since it's basically a 2-gate, it should be able to put down the hurt on a zerg who goes hatch-first. This is all just theory, though. Hopefully players better than me can actually put it into practice, but for now I'll just have fun in my own little custom games with friends. :X

RezChi

Canada2368 Posts

February 22 2011 01:14 GMT

I usually use 2-gate core to be honest, seems like i can get stalkers out much faster.

9-pylon
13-gate
14gas
15-gate
16-core
16-pylon (might have to cut probe production for a few secs not 100% sure because i tend to fuck up)

WinterNightz

United States111 Posts

February 22 2011 03:25 GMT

One last number play:

Let's pretend that with 2-gate core, we only did a "3-gate rush", and upon completion of warp-tech, we only had 3 warp-gates and 3-stalkers as the first warp-in. That means we would have an additional:

275 minerals
50 gas

to use in the 84 "gateway-seconds" that are would be otherwise unused. That makes time the limiting factor, so we can have 2 extra units: either 1 zealot and 1 stalker, or 2 zealots. So at the instant that warp gate finishes, we'll have:
2 stalkers, 4 zealots, and 3 stalkers warping in.

This means we can theoretically squeeze out a 4-food higher initial army while still having a fair amount of resources available to use to invest into later-game prospects (it's also important to realize that the army size in my 1-gate core example is actually a slight over-estimate based on available resources, and the 2-gate army size is a slight under-estimate.)

Please or register to reply.

My 2-gate Nonsense (with funny pictures!)

Completed

Ongoing

Upcoming