Custom Map #2373: [RUSH] Not My Mars. By: CS Z

[AutoPost]

This topic is for discussion of map #2373: [RUSH] Not My Mars


Author: CS Z
Size: 210x100

[CS Z]

I see an under 1-minute posted cheese time.  I would love to know how that was done.

Only thing I can think of is a zillion cannons + ERNs + liberal use of Teleports + leveraging the three UltraCaps (I just realized I should have removed those).

[Niflthaena]

Oh, that was me. I'm not sure I'd consider it "cheese", though? I mean, no more than these units are meant to be cheesy. Really, the only thing I do special is using paused mode and frame advance, as I can't keep up with the units normally.

Attaching a GIF of my replay, but it goes by in the blink of an eye, so lemme annotate my key points:

• One instant build unit
• Four instant build units
• Exponential spread of instant build units
• Cover a field in miners
• A whole bunch of ERN fabricators
• One frame later, delete all of them and replace, setting all the miners to want ERNs, repeating every couple frames

By the 25 second mark, I have over 90 energy production and 25 instant builders.
By the 26 second mark, I have over a dozen teleporters.
By the 30 second mark, I'm doing three-frame combos of place-instabuild-teleport a dozen units at a time - be they cannons or mortars.
By the 40 second mark I'm including instant-build units in the swarms to get nullifiers in place; that's also about the time I realized that the instant build units only target a specific cell once every n seconds, so I started offsetting my unit arrays for greater throughput. A lot of those units die in seconds as I teleport them straight into creeper a dozen at a time.

Nullifier completion: 51.4s
Item completion: 1m23.5s
Tower completion: 2m05.4s

A large proportion of the game time is spent waiting for towers to charge. For more accurate "rush" benchmarks, reducing the tower resources might help.

To be frank, the three UltraCaps basically don't matter. They cut down the first instant builder by about two seconds, but after the first one is up it bootstraps the rest of the base in a matter of frames. I accidentally broke one of them with my terraform experiments and didn't even notice its absence.
This is by no means optimized; I spent a while experimenting with the terrain modification units, and better use of instant builders for resupply, a better portal array, and properly offset unit arrays from the beginning would make a significant difference. I might give it another go at some point.

---- EDIT ----

Okay, I gave it another go. This time is much closer to being optimized, at:

Nullifier completion: 21.1s
Item completion: 31.7s
Tower completion: 54.0s

The main improvement here is switching from a "port and hold" strategy to a "mitosis" strategy once a toehold is established. Every inch of space that has no creeper on it is free real estate - one instant builder can create a cannon and another instant builder in the blink of an eye, as long as the ground isn't covered. The same goes for my own units; a cannon's already fired its shots? Delete it and use its space for instant builders. A builder made two units? Delete it and put another builder one square over. An area's clear enough to afford a single second of delay? Delete half the turrets, replace them with portals, port the existing turrets elsewhere and spawn new turrets.

In short, energy isn't real and everything is expendable.

There's still room to improve.  The main base would be much more useful with a couple more rounds of ERN creation - I was a little too eager to fill space with instant builders and ran out of space for ERNs, and my much more aggressive porting strategy meant 90 energy per second was not enough to keep the portals running. The easternmost towers could also be linked up earlier if I went a little more for connecting space and a little less hard on the emitters.
But I think the biggest improvement would actually come from going strictly west-to-east. The mitosis strategy is measured in frames - each layer can be built within two frames of its space being clear - which means the quicker the creeper disappears, the faster the base expands, and every cannon faces east when built. The impact of this is actually visible in the video, with several of the eastern fronts falling seconds quicker than their western counterparts.

Again, video is attached.
(This 54 second run, which can be played back in five seconds, took me over an hour to do. No savescumming - just plotting my moves carefully for each frame.)

[CS Z]

That's, um, more dedicated than I likely will ever be.  Very nice though.