25 May 2018

Balancing in Terraria: The Is-State (Part 2)

A few days ago I made a blog post where I went through the mind and thoughts that go and might have gone through the discussion of balancing. We came up with how we wanna put our classes. Now it would be interesting how far we are off. To determine this we have to analyze Terraria's balance.

Analyzing the Balancing

The factors we looked at when balancing were
  • How is the class played
  • How much defense should it have
  • How much damage should it have
So let's look at the defense and damage that Terraria provides us.

PyPlot of Terraria's defense per tier per class
To do this I looked at the defense of each armor for each of the four implemented classes and plotted these numbers using Python and the MatPlotLib module. It does not seem to plot but the x-axis shows the tiers I used. These tiers are: 0 - fresh game; 1 - brain of Cthulu / eater of worlds defeated; 2 - Hell reached; 3 - cobalt mined; 4 - Mythril anvil; 5 - Adamantite forge; 6 - mechanical bosses defeated; 7 - chlorophyte farmed; 8 - Plantera defeated; 9 - golem defeated; 10 - lunar event + Moonlord defeated. To run a little over the numbers the end-game armors that are craftable using Luminate once the end boss the "Moonlord" has been defeated have 78 defense on the melee armor, 68 defense on the range armor, 46 defense on the magic armor and 38 defense on the summon armor. generally, you can see the trend to be that the relation melee > range > magic > summon can be found here again. Which goes perfectly with what we assumed to be the case.


The next information we need to look into would be the damage. Now calculation the damage is a bit tough due to the following reason:
  • Weapon damage numbers are affected by weapons bonuses
  • We have to calculate in critical strike damage and critical strike chance
  • Some items hit multiple times
  • There are buffs provided by the armor set bonus that affects the values
  • Items may deal a lot of damage but have a huge use time or cooldown
Due to all these factors, we need to sit down and actually test the items. Though this has proven to be difficult too due to these reasons:
  • Attacking a target dummy may not provide buffs
  • Target dummies are not affected by homing attacks
  • Target dummies are not attacked by minions
And yet this is our best bet.

Preparing and Running the DPS Test

To prepare the DPS tests I downloaded a mod called CheatSheet, to allow me to get any item in the game without farming it first. I've created a new character and a new map and gave myself the best items available at that moment to my knowledge. I must also admit that I used the star cannon and the coin gun (with gold ammunition). Tests that can be done on a target dummy were done so. I'm assuming only one target dummy that is hit with no rebound. We're assuming a single target flying boss for these. I know technically there should be extra tests for multiple target and etc. but this should do for a simple comparison since single target DPS is required for the most bosses. any other tests for minion masters and magic buffs were done on bosses with only one hitbox. Additionally, I only changed armor and weapons as well as specific ammunition and necessary accessories like the mana flower or multiple worm scarfs (to negate boss damage). The damage number was guessed given the number shown by the in-game item "DPS meter". An approximate number should not do well then comparing the exact math but for a graph to look at it's enough.

PyPlot of Terraria's damage done by classes per tier
Thus due to the circumstances, the damage numbers may run wild if you try it yourself but it should be near with a few exceptions. Anyways as you can see in the graph we have a relation of magic > range > melee = summon.
In comparison our idea of balancing was magic > summon > range > melee. Now we were right with magic being top, though summon damage is much lower than range or melee in the actual game. I must say that I did not calculate any other items. This was just the plain minion damage. If the minion master would use the other classes items his damage would increase but even then it would not be higher than range damage.

Generally, though all the classes seem to be very similar during the progress for the tiers only splitting significantly after the Moonlord. The only difference is the range class which has a huge damage boost when using the star cannon, which requires farming stars during the night and the coin gun which in the damage tests consumes your gold. both these resources should be easily achievable when the time is put into it. Honestly, I am disappointed by the early mage damage and the overall summoner damage in comparison to the survivability.

Conclusion

So, in the end, it turns out we were right with the defense but wrong with the damage.
Now due to the lack of defense on some classes and the low damage output of them, it doesn't surprise me anymore that I preferred to play tanky melee all the time. You deal almost the same amount of damage compared to the others. Of course, there's no 20% more damage potion for melee or there is less damage increasing accessories for melee, but with that defense difference, I don't know.

I guess in our modding plan we need to mush the damage of a few classes. :P


Liked the post? Noticed an error? Wanna discuss the content or leave a comment*? You can join or check into the discord to do so! (*Note: Comments are disabled to avoid saving user data on this website.)
>> Join Discord

About Me

My photo
I'm a junior game developer and software developer and I created this blog to share my ideas, thoughts and whatever I'm working on or doing.