Tree of Savior Forum

Regarding the damage formula

Gameplay Feedback Features
#1 
  • 	I would like to see the damage formula for this game changed. It's far too simplistic in my opinion and right now, you usually hear beta testers talking about how they're pumping all of their points in a single stat, normally dex for physical attackers. It is fundamental for the development of the game's meta to have a bigger variety of possible builds.

  • 	First, I will explain my issues with this formula. From what I can see, it is basically (cpatk+skill-tpdef). αcrit+βcatk+(emod.eatk),* with cpatk being the caster's physical attack, skill being the skill damage number, tpdef being the target's physical defense, α being a binary variable assuming the values 1 or crit^(-1),**crit being the crit modifier of 1.5, β being a binary variable that can assume 0 if  α=crit^(-1) and 1 if α=1, catk being the caster's critical attack,emod being the target's weakness/resistance to the elemental attack and eatk being the elemental attack given from equipments such as Arde Dagger. 

  • 	My concerns with this formula mostly come from the fact that crits are too strong because it's the only multiplication done in the entire formula, besides damage modifiers such as weaknesses and resistances. Therefore, there's barely a trade off on using most of your points in dex, which also increases your evasion rate. I've seen builds of high level players where they used most of their points on dex for the evasion and critting, with a few in str and con. str and con don't do much in terms of damage when compared with critical hits, because they only increase your points for a bit more than 1. Another concern with this formula is how multi hit skills are disproportionately stronger than single hit skills, which tend to relatively fall off in damage the higher you go into the circles and do not carry the ability to multiply the bonuses from Arde Dagger and such equipments by the times the target is being hit. Personally, I don't feel that great when I level up because with this formula, I'm pretty much doing the same damage to monsters no matter if they are at the same level as me or 70 below. It feels like I'm going upstairs, waiting for the next tier of equipments and using all my points in dex since that's where most of your damage comes from, rather than climbing a mountain, looking doing and being impressed by how much progress I've done so far.

  •   In my opinion, the damage formula should  have more multiplications and to put it simply, be more complex and potentially have more variables. Make str and con more relevant, make the player feel like he's powerful when he levels up. I've made a rather simple damage formula myself just as an example of what I feel like it should have been: {[(1+clvl/5.maxlvl)(patk+skill)-(1+tlvl/5.maxlvl)(pdef)(1-diflvl/200)+βcatk+(emod.eatk)].αcrit.mod.patk/pdef}.{[1+λ(cr-sr)/100]^(-cr)}, with diflvl = (clvl-tlvl). 

  • 	Let me explain what I thought when I wrote this formula. I added variables regarding caster/target's level and an exponential decay on the efficiency of lower circle skills being used by higher circler characters. In the first part (1+clvl/5.maxlvl), I wanted to make so that the higher your level, the higher your damage would be, this is being adjusted by the coefficient 5 multiplying the maxlvl, meaning a max level char would multiply 1.2 by his (patk+skill). (patk+skill) is the caster attack plus the skill's damage, which could even be used as percentage if you make skill= [(per/100-1)patk+skilldmg], with per being the percentage number minus the sign and skilldmg being the flat number of damage a skill does. 

  •   With those two being multiplied, the target's defense would be subtracted from it. -(1+tlvl/5.maxlvl)(pdef) is pretty much like his caster counterpart, the higher is the level of the target, the more his defense gets boosted. (1-diflvl/200) is a multiplier that considers the difference between the difference of level between the caster and the target's  level; With this formula, if the target is 2 levels higher than the attacker, the former's defense gets boosted by 1%, this number can be higher if you change the 200 that is dividing -difflvl though.

  • 	Crit, damage modifiers and patk/pdef, the latter being extremely important in the balance of a game, are all multiplied to the difference between caster's offense and target's defense.

  •   Finally, the damage decay from skills, represented by{[1+λ(cr-sr)/100]^(-cr)}, which is pretty much a reverse compound interest. λ is a variable that can assume different values, 0 if no skill is used and something different from 0 if a skill is being used. That way, you can weigh by how much each skill can decay over time, the value doesn't have to necessarily be one for all skills, especially if you're a c3 class using a c1 skill of you own class. (cr-sr) is the difference between the caster's rank/circle and the skill's rank/circle, divided by 100 because otherwise the efficiency reduction curve would be too high. ^(-cr) means the higher your circle, the more skills you have and the bigger is the penalty for using a low circle skill, the only problem with this is the amount of circles we could potentially get in the full game, but the percent penalty could still be diminished by altering ^(-cr) to ^(-cr/2), for instance.

  • 	In the end, this is just my personal concern with the game's damage formula being too nice with critical and multi hits. It should be more complex and add more variables in order to make the experience more enjoyable. The formula I presented was just an example to show you from the IMC what I mean, adding more variety to the game. I know a change this huge is difficult to happen, as it would involve a lot of rebalancing in stats, skills and monsters this late into the development. However, the objective of this post is to show IMC that I and possibly others are not as happy with the damage formula as we could be in a game with so much potential.

** In those formulas, I’m using dots to denote multiplications and commas to separate the interger from the decimal part. I’m not a native English speaker, sorry for any spelling mistake.

3 Likes
#2

Complexer formulas while interesting are not always better.

There are a LOT more multipliers in the game than you’d think initially, besides critical rate, and even the current basic formula can get fairly complicated if accounting for -everything-
Not to mention critical attack is directly based on your strength and not much else, making it hard to get if your want a lot of crit rate from a ton of dex (bar highlander/barbarian, but that’s not the formula’s fault…).
Moreover, critical resistance becomes pretty high later on in the game and directly lowers your critical rate on those enemies, add to that that Magic attacks can’t even crit at all, and you find yourself with only two classes that can really make full Dex work.

Archers need STR, SPR and DEX. Going full on either of them is a bad idea unless you have some incredibly good gear, and even then…
Wizards need INT, SPR and CON. Unlike archers who usually wear leather for more evasion, and benefit from DEX more than just evasion, Wizard won’t be able to dodge much at all, so you’ll want some Con. Int is your main source of damage, but wizards are sp hungry so you’ll want some SPR too.

Stat differences become more and more apparent further into the game. Dex is no sole winner here

#3

Wouldn’t that mean you’re fishing for more dex then? x1.5 + crit attack is huge, not to mention the evasion bonus. I’m not saying that people shouldn’t invest in str, but that or the most of the part this investment seems rather small when compared with the investment on dex

I don’t get why would archers need much SPR, it is only really useful early on, buying mana potions is a better alternative. I agree that STR is useful for them, but that comes from crit rate buffs from both Rogue and Swift Step c2’s attribute, I feel like con is more useful for archers than spr, not by much though.
Don’t know much about wizards, but I doubt there’s any heavy investment into spr or con as it has for int, as they can’t crit.

#4

Early on, x1.5 damage is, well, a lot. Especially compared to the tiny attack gain one would gain from going STR instead.

But go towards late game a bit and…well:
Level 200 Archer, Rank 6, Full STR
587 Phys
387 Crit Attack
49 Crit Rate

Level 200 Archer, Rank 6, Full DEX
210 Phys
10 Crit Attack
301 Crit Rate

Notice already how 2101.5 + 10 = 325 is by far short of STR’s NON-crit 587 Attack? Not to mention; if that archer ever lands a lucky crit (or yknow, is a rogue, free 60~80% crit rate right there), that would be 5871.5 + 387 for a whopping 1267 attack.

I know I’m excluding skills here, but the logic holds. You’ll surely crit a lot with a full dex build, but keep in mind that the raw damage of your attack after a crit could very well be lower than that of a STR build without said build even landing a critical rate in the first place. Add in the fact you’re less likely to crit at higher levels EVEN with a lot of DEX, and the gap only gets bigger.

Now I’m not saying you should go full STR. Theres gems, skill choices, weapon stats and the like to consider too. Having a ton of rate but no crit attack or raw damage is bad. Having a ton of attack and raw damage but no rate is equally as bad. But when you consider all the classes, skills and equipments on top of this, you’ll start to realize just how versatile a stat build can be. Does my class require more Dex, or do I need Str?

A great example is wugushi. Their poison scales with Str, but unless they go rogue or invest in a lot of Dex, they’ll never have a lot of crit rate. Do you want more DoT damage, or rather do more damage on the rest of your skills?

Also, this one’s important; with the exception of Rogue’s Sneak Attack, most crit rate bonuses use your existing crit stat to calculate with. Swift Step C2 does not add a flat 25% crit rate. It adds (Raw Crit Value) * 0.25. Which means that if you have no dex to begin with, it wont do anything, but also by extension means that you need LESS dex than normal because this skill would make up for the difference.

Also, SPR is more valuable than you’d think. PvE it lets you spam skills more leisurely without having to pray you dont run out of pots anytime soon. In PvP it lets you bypass some of your opponent’s block stat, vital when fighting tanks. And it gives magic defense, which is absolutely life-saving in areas with high level magic monsters, as you CANNOT evade magic attacks.

To be entirely fair I’d say Archer needs everything but INT, haha.

#5

And I think this is the biggest issue here, the issue with Dex and Str at the moment is that +50% damage in your skills is just a lot more than you can get with Str once you consider not only your equips but the base damage and multi-hits, specially at lower levels - at the moment every build will just go dex until it gets 233~329 crit rate, and then go full Str, with some Con if you are an Archer or Cleric.

Your equipments and skills will indeed dictate what’s the better path while leveling, but the end result will be the same, even if they increase the crit resistance of monsters as the level cap increases, the fact is that you’ll be gaining a lot of damage and crit attack from your equips, so losing some damage in your stats will never be relevant enough that losing a semi-constant 50% damage increase will be a choice.

Although I’m using the world “never” here quite loosely, as at rank 11 we have an extra +100% Str bonus from ranks alone, so every point at str is actually 2 points and the figures there might be very different from what he have, for all that matters. (as extremely high crit resistances)

Only two skills scale with str afaik, and you would still go with plenty of dex since you’ll be spamming oblique shot anyway (and using multi shot too, and the evasion)

PvE you don’t really need spr at higher levels tho, alch pots are good enough and unless you are spamming a lv 15 oblique shot (which you shouldn’t be), you really won’t be out of SP, if anything some spr helps for the regen, but that’s kind of it.

And ignoring the PvP aspect as I have no idea how much block penetration you would need there, both def and m.def are flat, so you would need 5 points in spr to decrease 1 damage or you could put 5 points in con and get extra 425 health, spr makes very little difference in this regard.

#6

A skill needs to do at least about 540 Skill Damage Per Hit to make the critical damage of the DEX build outperform the raw damage of the STR build, and the skill damage needs to become exponentially higher the higher the level cap becomes, not counting lucky crits on the STR build.

It entirely depends on WHAT skills your class path has, and if you’re using equipment that compliments your stats properly. Meaning none of the stats are useless compared to DEX, so I’m not disagreeing with you here at all, it’s just that that was my point, and reply to this thread to begin with:

There is stat diversity, it’s just hard to see until later in the game. The formula given by the OP would mostly fix early game but change relatively little late game, as builds would be just as diverse.

#7

I’m not sure if you think about diversity the same way I do, but my point goes through the fact that bosses at the level 140~200 have around 60 crit resistance, with 293 crit rate you’ll crit with 90-100% chance against basically everything, with 233 60%-75% - there’s no point in going full dex, but there’s even less of a point going full str.

The only variety right now is “does you classe have a crit rate buff?”, and this will change the amount of str:dex you’ll get, but that’s it, the majority of players will just go until they get a decent crit rate and focus on str with some con/spr for comfort, there’s no reason to deviate from this considering the higher damages from weapons and skills as you increase on ranks and levels. Everyone does literally the same thing as there’s no real competition at this point in the game.

Regarding OP formula, I have my doubts about how good it would be, considering skill decay and all, but I honestly believe it’s better than what we have, although I don’t think anything is going to change at this point.