Tower Defense Evolution - Rewrite


Pure tower defense: progress by killing biters, dense nest placement, periodic waves with bosses, Master Ammo Chest system. Rewritten with prototype-based enemy scaling for better performance and control. Discord of the server where im more active: https://discord.gg/93Vcq8PPSY

Tweaks
6 months ago
2.0
615
Combat Enemies Environment

g Issues

a month ago

Vague name, I know. But there is quite a bit that needs to be addressed here.

  1. Enemy spawns. I tried multiple runs and kept seeing behemoths as early as wave 5. With the standard 10 minute waves, that is 50 minutes into the game. That would be pretty crazy even if science wasn't gated to enemy kills. I thought that just can't be intended behavior, and I assume it isn't.
    In the wave system script, you have a section for tier progression, with this function:
    if wave_num >= 3 then
    add_next_tier(base_tier)
    end
    I understand the purpose of it. But as written, it increases tier by 1 not once upon reaching wave 3, but EVERY wave starting with 3. So you get medium at wave 3, big at wave 4, behemoth at wave 5. And the code uses math.random to pick one - meaning they are all equally likely. On my last run before looking at the code I lucked out with 0 behemoths at wave 5 so I managed, and then at least 15 on wave 6.
    This function cannot be fixed as is, it needs to be remade from scratch to function properly. Furthermore, randomly picking between 4 biter sizes throws any semblance of balance out of the window regardless of wave number. I would suggest a larger range of values and if-then cutoffs for specific size used, so that larger biters have a lower chance of spawning after they are unlocked based on tier, or a point pool per wave that is reduced for each biter picked (perhaps based on health).

  2. Science. Now that science uses labs, most techs take much longer to research than in vanilla. Assuming you have enough science, in base game needing more types of science does not affect time. This mod uses a single item and higher values for later sciences, which makes the duration skyrocket. For example, Physical projectile damage 3 needs 18000 lab-seconds in vanilla, but 42600 here.
    It is possible for labs to consume multiple units of the same science type "per cycle", so instead of calculating science costs as 1 token with (vanilla*multiplier) cycles of 1 second, I would suggest (multiplier) tokens being used per cycle. This would prevent the need for more labs (unless that is intended instead of the production chains for science) and, in my opinion, fit the style of the mod better with research being completed rapidly as long as you have the tokens.

New response