Removing % LS/MS from the game?

[zjat]

I have debated on removing entirely the passive effect of lifesteal and manasteal percents from the game. Using instead flat amounts for passive effects. While percentage effects may still exist on specific skills.
My research question is less on the side of how to do this, but instead on how many things this would affect. Like for example would this in any way break the game or gameplay?

If you have the time and the care to respond can you list things in pro/con format? Thx.

[Desocupado]

(I necro-ed this because it deserved an answer and the user is still active)

%Life steal
Pro: Proportional to dps
Con: Make healing too strong for high dps characters
Con: Make any monster with low dps meaningless (forcing you to use high dps monsters instead)

Flat Life steal:
Pro: Much easier to keep the healing at lower values
Con: Scales better with faster/AoE attacks, regardless of DPS
Con: Useless for spike damage (high dps low amount of attacks - which isn't too terrible for a flaw)

% Mana steal:
Pro: Scales with dps instead of attack speed
Con: Makes mana regeneration useless (mana is a very poorly implemented mechanic in vanilla)

Flat mana steal:
Pro: Much easier to balance
Con: Scales better with faster/AoE attacks, regardless of DPS
Con: Useless for spike damage (high dps low ammount of attacks)


The problem with life steal % is that player DPS is several times higher than their HP.
Also you don't need to have more than a flat 5% life steal mod.

[zjat]

Very insightful points! I will add my own insights to this for others that may look into this avenue of editing:

I ended up going ahead and testing out flat leach, using dodmg events and life on kill eventfunctions. I had a few problems and kept some aspects and left others.

Flat Life steal:
Pro: Much easier to keep the healing at lower values
Con: Scales better with faster/AoE attacks, regardless of DPS
Con: Useless for spike damage (high dps low amount of attacks - which isn't too terrible for a flaw)

Flat life steal immediately became overpowered on someone like the paladin. I had a hardcore lan session with a roommate and I could literally tank diablo's lightning (almost). However I had somewhere around 20 life per hit and quite a bit of hpregen and used zeal/fanat. I did however end up removing all but a few life% modifiers, all item affixes as % are now gone. But a Pskull for example only gives 2 life per hit instead of the originally tested 10. In regards to the style of D2 being a lot of getting hit and potion chugging, I don't see flat life steal as a major issue in the long run especially due to how characters were rebuilt in my mod.

% Mana steal:
Pro: Scales with dps instead of attack speed
Con: Makes mana regeneration useless (mana is a very poorly implemented mechanic in vanilla)

Flat mana steal:
Pro: Much easier to balance
Con: Scales better with faster/AoE attacks, regardless of DPS
Con: Useless for spike damage (high dps low ammount of attacks)

I also ended up testing mana per hit. This went horrible. Due to how mana regeneration is already a terribly designed system (vanilla), mana recovery on a late game caster is generally very high (iirc 30 per second or more), while an on-hit character requires not only mana leech but can often sustain full mana from it due to the huge damage spikes most characters can do. However, in testing this I found a way to do something I wanted to do at a previous point in my mod: remake how mana works. I tested a few concepts for mana regen, first on the barb and second on the druid. The amazon and assassin also had changes due to this failure.

The Barb: At the time I was a new player to league of legends and had just unlocked a character "Tryndamere" who gains fury on attacks and kills, double on crits, and gains fury when using his ultimate. I used this concept to rebuild the barbarian resource, first limiting the mana pool so that energy granted 1/2 as much mana, later to completely restrict mana on the barb to his base stat (50) while granting 0 with energy gains (there's no more +life/mana in my game due to this). Also, base mana regen is slightly negative. I made this new barb have a built in passive state from Bash, that grants him mana per kill (5 base), and mana on being hit (1 base), while making earlier skills free: Bash and Double Swing (and of course changing their strength and scale). This continues to work well, only tested in normal difficulty. Due to the barb not hitting fast, he is focused on hitting hard, or for that matter, being hit, to gain his strength. In the same functionality, all %dmgtomana effects were changed to flat effects to benefit a style for a tank focused character.

The Druid: Due to the lifesteal and manasteal changes I had played with, I used the druid as a secondary situation. I remade both wolf and bear forms to have further passives built into their states, causing the druid to have negative life on hit, regaining 2.5x on kill, with the cost based on strength. Hunger is one of the few spells left alone in the sense of converting % to flat, and thereby is a much stronger skill now. It also has a small passive that increases life per kill (1 per 4 skill point). Then comes the reverse "Fury" of the barb. I called it Nature's Fury. Instead of gaining mana on kill and having negative mana regen, the druid lost mana per kill as a punishment from "Nature" but was granted much higher mana regen also due to nature. This was a very fun thing, but after a few months of trying to balance it, I gave up. Losing 5-10 mana per kill left you constantly empty, and having 500% or more mana regen made late game mana regen crazy. (Stupid mana % regen)

The Amazon: Flat mana regen was atrocious. Strafe and multi shot skills could regen each hit as mana, making 1 mana per hit negate the entirety of the skill. "Double the mana costs!" then someone happens to get 2 mana per hit and make it free. "Triple them then!" Now it's too expensive to use unless you have 1-2 mana per hit and I still failed to solve the balance. AND THEN... Piercing arrows... Flat mana was not possible due to pierce and aoe/multi - arrow spells.

Assassin: While % mana leech was left in, the assassin is a hog for mana, similar to the barb, at early ranks. This makes a martial artist very weak early, but also makes a trapsin a potion popper. Remedy: I remade all the martial arts skills eventually, during this I gave all of them damage during the charges (similar to maul) and with serpent strike (% mana % life hardcoded) I granted 1 mana per kill per charge. This allows an assassin focus on death, you know cause "assassin." But only 1 MA skill was left as AoE effect, Phoenix Strike. However, in testing this I discovered the serpent strike ability was excellent for a trapper for mana purposes. (I also made earlier spells like firebomb and staticfield convert srcdmg). Thus, a combo of MA and Traps was now more feasible, while a full trapper would still most likely want to use serpent strike, the 3 mana per kill never scales upwards and can be easily gotten from 3 tirs.

Runes: Tir and (something else) were changed. Tir was reduced to 1 mana per kill due to some characters already having freebies, and others being more focused towards %leech or mana regen. Another rune was made to grant life per kill, originally a rune, later changed into a gem "Bloodstone" which gives a bonus for demon kills.

Tl;dr - There are many different approaches to resources. Blood (shapeshifting), "Fury"(killer barb), Charges (serpent assn), and nature's fury (elemental druid). Per kill ended up being better for mana resources for me than mana on hit due to AoE skills, Piercing effects, and general theme. While life per hit works fine due to character themes, such as multi shot healing massive amounts but costing enough mana to offset, As well as the general high damage hits and life trading associated with Diablo 2/lod vanilla.

tl;dr2 After having tested some things flat mana was really bad, while flat life was okay due to mana costs. Per kill works too.

[Desocupado]

Other systems for mana
Well using mana for something that isn't a resource works much better for a action game like diablo.
Diablo 3 had good examples of some workarounds - The regular skills cost 0 mana while powerful skills cost a lot.
You pointed using mana as synergy is quite cool.
Take a look at Whist Mod Mana Concept (It's a good thread)

When to use mana generation
If anything I'd say mana steal only makes sense if it's build in a skill, like concentrate/berserk. Or even cobra strike/hunger.
Heck firebolt could build up mana so you can use meteor more often.

Balancing High attack speed versus Low attack speed
Use:
(Magic) damage Reduced by XXX
Attack takes XXX damage

Using spike damage versus high attack speed is a strategy (sometimes based on class like the zealot - but a paladin could use charge as well).
Enemies need to have different strengths and weakness to each strategy otherwise all you will care is DPS. But if you apply a flat damage reduction multi-hit tactics get gets screwed.

Random related idea
Attacker loses X mana when monster is hit

[zjat]

High Damage vs High Speed Attacks
Per Skill Use
Something to theoretically alleviate the issue between these is to have the skill itself have a so to say limitation. For example multi-shot with 20 arrows can leech 20 life/mana if based on flat amounts. Instead that leech can be limited/categorized as simply 1 attack. (I chose not to mess with this though). Not sure if codable but could work for multi-hit skills like zeal as well, again not something I care to try.

Ratio on Skills
Another solution, again theorizing, is a ratio effect similarly taken in league of legends, specifically speaking of damage ratios. There "Ability Power" or AP increases spell damage. Not sure if the D2 engine could ever do this with leech, but the idea itself would be that skills or weapons could have a leech ratio, hard hitting skills having a higher ratio than lower skills. Reusing multishot for this example: 20 arrows, "1 life per hit" might have a 20% ratio per arrow, allowing up to 4 life leeched. Ceilings/floors being able to be engineered as wanted, such that a minimum of 100% (1 life) &or maximum of 200% (2 life).

These are just ideas I've thought of outside of the usage, or at least might current needs and capacities of d2. (I like to think and design a project similar to d2, "Clone", that has a 99.99~ chance of never going past paper, but is fun.)

Edit: Forgot to write this one:
Health/Stat based instead of Damage based
Aside from using a passive state, this is possible in d2. The logic behind this is that instead of leeching 5% or so from a target based on damage dealt, you gain a % of your max health, enemy health, or your missing health - not sure if the 2nd is possible in d2 engine. Each health gain/leech has a different situation for usefulness. Other stat-based functions would be within this idea, such as strength instead of max health. - This could also function (outside of d2) with a type of cooldown, such that higher attack speeds cannot proc more life regen

[Desocupado]

Ratio modification
You'd need some code editing

Life steal based on missing health
A skill could easily grant Life Leech per life missing using a calc field. At least such formula would make life leech ineffective to recover full health.
Unfortunately there's nothing that can be done with item statcost.txt

Well, instead of life steal mod on items - use a CTC when you hit.
The skill cast a quickness (burst of speed) clone that recovers a % of lost health during 2 seconds (let's say 10% of lost health). since each hit would just refresh the buff it wouldn't be more powerful with a high number of hits.

[zjat]

Well, instead of life steal mod on items - use a CTC when you hit.
The skill cast a quickness (burst of speed) clone that recovers a % of lost health during 2 seconds (let's say 10% of lost health). since each hit would just refresh the buff it wouldn't be more powerful with a high number of hits.

Personally, I would use something more in the spectrum of 1 second (due to how fast attacks are from early on). But I may actually use this now that I'm thinking about it.

[Desocupado]

Also, Life Leech % is calculated from the effective damage you dealt if this is accurate.

So having "(Magic) damage Reduced by XXX" would also reduce it's effectiveness.

Derail
(I'm using this damage reduced by XXX idea to sketch a way to balance monster's damage/health and player's health and mana)

[zjat]

Also, Life Leech % is calculated from the effective damage you dealt if this is accurate.

So having "(Magic) damage Reduced by XXX" would also reduce it's effectiveness.

Derail
(I'm using this damage reduced by XXX idea to sketch a way to balance monster's damage/health and player's health and mana)

I've always assumed LS was based on damage dealt due to the mechanics of Wc3 and games that evolved from it dota/lol. Wc3 borrowed heavily from d2, especially auras (thorns) and LS. Thorns in wc3 I'm unsure of, but in dota and league I know thorns damage is based on damage before armor, while lifesteal is based on damage dealt, thus lifesteal doesn't counter thorns. 30% LS granting 21 life vs a target with 30% damage reduction with 100 damage, while thorns would hit back at 30 damage for 30% returned. While this is usable for some games, I think that this hidden mechanic is tricky and deceptive and I hate deception in a video game mechanic.

-Sidenote- Due to changing LS properties in my game I went through and looked at all thorns as well, changing it to 10% +X/lvl, as well as updating the magic modifier to be higher and more useful. 300% will instakill any player in hell (Iron maiden in act4...), but 300 damage won't so it even makes pvp better (though I've never pvp tested my mod at all, nor plan to). This thorns/ls change that I did theoretically makes the game more transparent... I may remove the difficulty level changes to % leech since they are all skill based.

I still would like to redesign how mana % works somehow, after the frustration of doing flat mana leech failing miserably, I reduced all MS to 1-3% in the game and "rebalanced" mana regen % on a few skills and items. Mana is also 10-20% lower base in my game, but all skills are rebalanced to be usable early game or late game, such as firebolt being 1 mana with .1 or .2 increases per level, but fireball is 1 per level or so and meteor is expensive idr how much. It makes leveling and early gaming much faster/easier while your mana pool and itemization is usually trashy.

But uh, any ideas of how to make MS% less "useful" but stay in the game? The biggest issue with MS% is the fact you get 0 mana from early leech pre-100 physical hits, and 30 or so late game per hit with 3% on 1000. While my mod has 3% as rare and maximum, it is easy to stack items to give 1-2% per weapon/glove/ring/ammy (manalds and such). % regen is similarly glitchy due to it being mathematically based on your max mana pool... *Sigh*

Edit: Never thought about "On attack" would that affect multi hit skills like zeal and multishot/strafe? If I could limit it to say 1-3 mana per skill usage, that might work. Such that a multishot of 20 arrows would leech 3 mana, so would zeal etc...

[Desocupado]

Options to Thorns%:
"Attacker takes X damage"
"Attacker suffers Crushing blow"

Idea 1
In theory, it might be possible to make a "thorns curse" by giving negative flat life leech to the target, no? (with the number also based on monster/character level instead of just skill level)

Idea 2
If we play around with some hidden CTC skill on hit, like Laz did his "just been hit state" we can make the curse activate a hidden aura causing degeneration if you have attacked recently by using a dummy stat in the formula defining the health degen stat.

[zjat]

Options to Thorns%:
"Attacker takes X damage"
"Attacker suffers Crushing blow"

Idea 1
In theory, it might be possible to make a "thorns curse" by giving negative flat life leech to the target, no? (with the number also based on monster/character level instead of just skill level)

Idea 2
If we play around with some hidden CTC skill on hit, like Laz did his "just been hit state" we can make the curse activate a hidden aura causing degeneration if you have attacked recently by using a dummy stat in the formula defining the health degen stat.

I'm not so much concerned about thorns, I like where it is as a 10% + X/lvl system. I am however still looking for options with alternative mechanics to mana regen and mana steal. Mana per kill has been extremely useful for some classes such as the assassin and the barb, but the other classes feel a bit empty still, due to % regen being poorly designed and % leech being less divisible/manipulate-able.

[Desocupado]

One route:
Skills without mana costs, but with mana affecting their power as a synergy.

Let's take your Druid's "Nature Fury" for example.
Step 1 - All Druid skills cost 0 mana
Step 2 - Make mana per level and per energy point = 0 - staring mana is 100 (for simplicity sake)
Step 3 - Make a new passive increasing mana regen for each point in energy
Step 4 - Use your lose mana on kill idea
Step 5 - Put a synergy in each skill, like +2% damage for each mana % (and cut base damage by 50%) - this translates in 50% without mana 100 with half bulb and 150% with full mana

Mana gain mechanics
As far as mana gains go we can:
--Gain flat mana when 1 - strike, 2 - do damage or 3 - kills
--Gain flat mana struck or is damaged
--Gain % of withstand damage as mana

Other options:
Skill 1 - On hit, 2- when hit, 3 - after killing skills changing mana gain using some formula as on hit
Example
1 - quickness clone granting mana regen after hiting something
2 - being able to get mana on hit after being hit (rage)

Sugestion about mana costs
Assuming Vanilla skills - only the following skill should have a mana cost

Amazon - Maybe just the highest area spells (Immolation, Freezing, Plague, Lightning Fury and Multiply shot)
Assassin - Mind Blast - Shadow clones - True Traps (sentry but not blade)
Barbarian - Just active Warcies and Whirlwind
Druid - Summons and perhaps Hurricane and Armagedon (if their damage is upgraded enough)
Necromancer - Revive, All golems (but Iron) and all curses
Paladin - Most combat skills but Zeal - Maybe holy Fire/Freeze/Shock could use some mana cost as well
Sorcerer - High utility/inherent power spells (Teleport, Static field, Nova) and possibly Thunder Storm and Hydra (If you alter them a bit)

[Aught]

Personally I removed flat LS ans MS from ranged attacks, it can't be balansed Also there is diffuculty.txt that you can use for balansing % LS and MS. So no need to be afraid because of amount % MS and LS. I made this properties divided by 7 on the hell, and there is no reason to worry about stacking them. Good luck, bro!

[nightshade00123]

I believe there is a place for vanilla %LS and %MS in the game it just needs to be monitored a little closer as well as the damage output/attack speed of the characters that will likely be using it.

Also like Aught said, DifficultyLevel.txt combined with when the affixes can appear, or when recipes can be made, should be more than enough to do this adequately. Also monstats.txt has the 3 drain columns which effectively control exactly how much can be leeched per enemy and per difficulty. So with these alone you could easily monitor what enemies will give big gains and who won't.

I don't believe making the % steals just a "small" or "manageable" amount that is the same across the board for the sake of nerfing it is a good solution. But having some enemies be vulnerable to it and others immune to it or very strong against it can make a reason for stacking it in some cases be very helpful, you'll have to be careful of your skills still.........my rule of thumb is hard to use/slow to activate skills give big damage while AOE/fast skills very low damage, while on the flip side coming against a mob that can't be drained at all or very resistant to it would make you think now why did I stack all this % leech and not damage resist or damage output. And then mixing the mobs together can make you really have to put effort into who you drain for it to do anything at all.

Any how good luck!

[Desocupado]

Reducing LS
Indeed it can be balanced by combining both Difficulty divisor and specific monster divisor can be used to further nerf in nightmare and hell (when we assume monsters HP and players DPS rises)

Area versus single target
Use srcdmg to reduce aoe skill damage (multiply shot comes to mind).
Since flat dmg reduction make this kind of skill less powerful (both in dps and leech, so use it on some enemies) use it to control ranged leech.
Give melee classes aoe skills. (They deserve actually - examples include cleave, non-moving whirlwind, some sort of shockwave and explosions)
Don't forget you can balance Character's between Area Damage and Single target damage. so the ones good with area, would be terrible for boss killing (but never make a character with just single target skills)

General recommendation
% life steal mod should not come with enhanced damage % mod.
% Life steal is actually a good mod to be exclusive to Unique items (use life on kill instead)
Keep decent leech mods exclusive to one kind of gear (probably weapon or their gems/runes). all other gear should only give very minor amounts with high trade-offs.
Maybe just character skills have Leech / healing (abolish healing potions)

% Life regen
Create a new stat - Life regen per HP. Use this instead of the old regen.

Skill Ideas
Might be simpler to use Guild Wars approach - everyone class has a skill or two to heal themselves. Interesting options/suggestions:

• Timer based heal skill (the skill heals, and puts you in a state you can't use it again until it wears off) - (Laz skill timer mechanic)
• Mana based heal - the skill instantly heals but have a large tool in the mana supply (in can affect mana regen based on how much healthj was recovered)
• Prayer - Gradually heals, but limits your offensive potential (boring actually)
• Cobra Strike/Hunger/Feral Rage - Instead of using a high damage skill you waste time using low damage skill to heal yourself (quite fitting for diablo actually)
• "Strafe" Heal in a dangerous situation - The skill launches one missile per nearby enemy that produces a healing nova healing you - only effective with several enemies around
• Berserk-like - The healing skill makes your defense (or resistance) weak while you use it or if you used it recently (quite good actually)
• Inferno Heal - You need to click and hold the button to heal - it's similar to prayer in some ways, but more adequate for a ranged character with minions/traps/hydra
• Life steal based on missing health - Has the advantage we discussed earlier - It won't be effective at keeping you at full health
• Regen based on missing health - like the one above but makes a good adjustment for an aura
• Some skill heal you slightly when you cast them (for for fragile casters - since you won't heal when you are kiting/running - can be associated with a dummy stat)
• Healing based on current mana (if mana regenerates fast enough, the heal skill can be less effective with full mana and cost 0 mana)
• Heal when a summoned creature dies (healing nova missile on death )
• Damage resistance buff (Bone Armor/Ciclone Armor) - Whiel these skills don't directly heal, they serve a similar purpose
• Corpse eating skills - Fenris Rage/RegugitatorEat
• Non-moving healing minion/missile - The mana cost / low mobility is an interesting trade off for a heal skill
• Healing orbs - Diablo 3 - enough said
• Healing "Blaze" - Heal when you move
• Conditional heal skill - Only heals if you are under a specific HP threshold - very good if combined with a "strafe heal"
• "Just been hit" "Just killed" as a condition for Healing - If you use Laz "state/on hit/on death mechanics" you can put it's dummy stat in a formula for life steal/heal

[zjat]

Personally I removed flat LS ans MS from ranged attacks, it can't be balansed Also there is diffuculty.txt that you can use for balansing % LS and MS. So no need to be afraid because of amount % MS and LS. I made this properties divided by 7 on the hell, and there is no reason to worry about stacking them. Good luck, bro!

Been play testing my mod and mana is a huge issue for heroes that do not have a mana regen mechanic (specifically amazon in my mod). Mana% leech was never removed due to it's ridiculous issues in using flat, but due to your comment I have been reminded that I already made the divisor enough and made %MS weaker when spawned. My playtesting/lan with friends lately has shown that it's now super difficult to get MS% due to my gem changes. I now feel very comfortable with my more recent edits and have restored LS% to runes and increased MS% on skulls.

You're all very insightful people, I read all comments.