How "Life per Hit" could have worked better in D3

General Discussion
In D3 vanilla there was Life Leech, but it was removed because it was too powerful and also because it lead to counter intuitive playstyles, like that when your Barbarian was low on health, he jumped into a big group of enemies and healed up instead of moving away from a dangerous situation.

That was replaced with Life per Hit, which first was based on the proc-coefficient and the amount of enemies hit, but later it was simplified and every attack gave the same amount of Life per Hit no matter which skill you used and how many enemies you hit. The reason for the simplification was that LpH was too unreliable when you fought against smaller groups of enemies or just single targets, so it was basically useless against bosses.

I wanna make a different proposal (although I am totally aware that this will not be implemented, because they are not working that much anymore on D3, so this is more just for fun), which sort of is between the first version and the version of Life per Hit we have nowadays.

Here is the idea:

  • (1) Instead of making LpH based on the proc-coefficient, let each skill have a Life-per-Hit-coefficient, which basically is a multiplier for Life per Hit.
  • (2) For the first enemy hit, the LpH-multiplier is noticeably higher than for the next few enemies you hit, for example, the first enemy you hit with Seismic Slam gives you 250% of your LpH and the next 4 or 5 enemies you hit with Seismic Slam give you 30-50% of your LpH. Than if necessary, the next enemies you hit with Seismic Slam either give you 0% of maybe something like 5% of your LpH.
  • Keep in ind that the numbers that I use here are just examples for the purpose of illustration!

    The big difference between how LpH worked in the first version of LpH and my proposal is, that unlike every enemy you hit with an attack would give you the same amount of LpH as it worked in the first version of LpH, in my proposal the first enemy you hit would give you a substantially larger amount of LpH than the next few enemies you hit, which makes LpH much more reliable against single targets or smaller groups of enemies.

    Now, it still has to be debated if all enemies you hit after the first enemy shall give you the same amount of LpH, although it would be a smaller amount of LpH than the first target, or if just the next ~4-5 enemies you hit with an attack would allow you to benefit from LpH, so that it would be capped at ~4-5 targets max or if any enemy beyond these ~4-5 targets shall still give a very small amount of LpH. Also, the amount of maximum targets can vary from skill to skill.

    So what do you think of this concept?
    Would this make the mechanics of Life per Hit more interesting and deep without being too imbalanced? Or would still prefer Life Leech or something else?
    Yeah. I really missed my rend barb. I thought it was a fun play style that had your barb thrive off of large combat situations.

    After the change I kind of fell in love with the quake barb set when I decided to try it. I'd already fell into that play style with out thinking about it using legendaries. The set bonus was just damage gravy.
    I look to it more as the numbers generally being too small to matter, while stacking it further goes on to compromise other needed stats. Some abilities went on to instead gauge LoH based on your max HP, which I think is probably the more elegant solution than mucking up calculations with further post-proc checks. Even something like DH's Leech passive is a hard sell for the prior reasons, as it's both fixed numbers and means likely giving up some offense.

    But there's also more the general point that eventually only ZDPS or heavily buffed characters are really capable of taking hits, nevermind being able to live long enough to actually put LoH to use. For a while, the former also relied way more on various Regen sources which were both more potent and didn't actually need to have you whack something to get the HP back.

    In the end, I'm inclined to agree with the dev stance that damage%-based LoH and relying on it was not good design, in both D3 or even D2. However, we're also looking at the situation where its replacement is still lacking in functionality and perhaps creativity. From the group perspective, LoH could become a shared buff where after striking a foe, party members could also leech for 3s if the target isn't struck again. If they went the route of converting all the fixed numbers to MHP%, they could also extend it to whatever percentage lowers a mob's outgoing damage while increasing the damage they take as a more generalized leech or sap mechanic. This would also improve the values of the earlier noted skills since they gain a debuff component. But I don't expect any of that to happen, so I'll just shut up now.
    01/13/2019 07:33 PMPosted by Pestilence
    Yeah. I really missed my rend barb. I thought it was a fun play style that had your barb thrive off of large combat situations.

    After the change I kind of fell in love with the quake barb set when I decided to try it. I'd already fell into that play style with out thinking about it using legendaries. The set bonus was just damage gravy.

    Did you post in the wrong thread?
    Just asking, I currently do not see a correlation to what you wrote and the topic of this thread...
    01/14/2019 12:21 AMPosted by Saidosha
    But there's also more the general point that eventually only ZDPS or heavily buffed characters are really capable of taking hits, nevermind being able to live long enough to actually put LoH to use. For a while, the former also relied way more on various Regen sources which were both more potent and didn't actually need to have you whack something to get the HP back.

    I think to a large degree that also has to do with how the difficulty system works in D3. It just raises enemies damage and their life further and further and further.

    I think it would have been better if there would only be like 10 different difficulties, and at around difficulty 4, 5, or 6 the increase in enemy health and damage would stagnate or just increase by a minor amount and htat the enemies would be made more difficult in other ways, like increasing their movement speed, increasing their CC resistance, giving them improved AI, giving them new abilities, letting more elites spawn and giving elites new and more affixes, etc.

    01/14/2019 12:21 AMPosted by Saidosha
    However, we're also looking at the situation where its replacement is still lacking in functionality and perhaps creativity.

    That is how I see it as well and that is why I suggested to let the first enemy hit by a skill give you more LpH than the next few enemies you hit with it.

    That way LpH wouldn't be useless against bosses and it also wouldn't lead to a counter-intuitive playstyle were you jump into a large crowd of enemies to heal up instead of moving away from them.
    So with your idea, if my barb is low on health, I would still choose to run into a group of enemies rather than run away from danger??

    If we place too much emphasis on self heals, we would always choose to run into a group of enemies to heal rather than run away from danger. If we put more emphasis on EHP (or the number of hits we can survive in high GR), then we would actually run away from danger.
    I'm not convinced that % life steal cannot work well in conjunction with flat (or %hp) regen per second. I think that with the right balancing of offensive and defensive character stats it would be much more effective. When damage scales so much more than defensive does it becomes hard to balance and life steal can start to out perform additional defensive stats resulting in very choppy health gains which is not desirable. It would also be important to have a good balance between single target and AoE skills and that is something that would be good for the game in other ways.

    That could also make LpH have worked better in the past implementation but problem with that is attack speed becomes a very important factor which is not the case with life steal.

    My only real problem with your proposal of a hybrid system is that the mechanics would then be very murky. I like for stats to function in a way that is conveyed be their name.
    I like to think that the vanilla had so many good things but everything was just misplaced.
    That Life Leach and LpH were such good things that I really missed when they simplified everything.
    Maybe that was the mistake... simplify instead of put things on the right place.

    Btw, good point OP. Very nice done.
    01/13/2019 01:39 PMPosted by clueso
    (2) For the first enemy hit, the LpH-multiplier is noticeably higher than for the next few enemies you hit,


    Good approach to the problem. I actually had a similar idea but that require to add a dynamically switched 7th utility slot to the game. Your idea is more applicable to the situation.

    I also prefer "max life% gained per hit" to be more abundant between skillrunes or a rework of "lifesteal%" affix even if it meant to work at a fraction of effectiveness.
    Lifesteal was actually a fun mechanic but it was better restricted to fragile dodge based classes like DH only. It was no surprise a tank like Barbarian abused the crap outta it back at D3 vanilla.
    01/13/2019 01:39 PMPosted by clueso
    Life Leech
    Lifesteal.

    How can you make a post like this when you don't even know what stuff was called back then, much less how it actually worked...

    But as to the actual topic, Life per Hit is a nonsensical skill or effect, but since it exists, might as well keep it as simple as possible, using the least system resources as possible.

    Your ideas, while certainly unique, are needlessly complicated, and would additionally bog the system down with unnecessary math.

    Considering the fact that the best characters get barely 2.5 APS, even 20 K LPH is an adequate to affect any actual healing in higher difficulty situations.
    01/14/2019 12:30 PMPosted by SteLitY
    So with your idea, if my barb is low on health, I would still choose to run into a group of enemies rather than run away from danger??

    No, because in my concept the amount of times LpH can trigger would be capped at a certain point or the effectiveness on additional enemies you hit would significantly decrease beyond a certain amount of targets hit.

    Like this for example:
    First enemy hit: 250% LpH
    2nd to 5th enemy hit: ~30-50% LpH
    6th enemy hit and beyond: either 0% LpH or just a very small amount like 5%

    Numbers used here are just examples to illustrate the concept.

    In vanilla the problem was that life leech was not capped on the amount of targets you were hitting with a skill, therefore making it significantly more powerful on large groups.

    In the first version of LpH, it was useless against single enemies, because the amount of enemies LpH could trigger on with a single skill was also not capped, therefore making it much more useful against large groups of enemies, but useless against bosses.

    My proposal making LpH still very useful against single enemies and a bit more useful against groups, but not totally overpowered when you fight against larger groups of enemies.
    01/14/2019 01:33 PMPosted by naksiloth
    I also prefer "max life% gained per hit" to be more abundant between skillrunes

    This system indeed could also work with a percentage based recovery of maximum life instead of a flat life recovery, or maybe with a mix of both.
    Thanks for the input.
    I personally would not mind life steal % making a comeback. Part of what added to the thrill of combat in D2 was knowing sometimes you need to run in and just keep fighting literally for your life instead of running away and potting. The sense of danger was still there despite how obviously good life steal was.

    IIRC, the most leech I had was probably my Zealot wayyyy back in the day before runewords became insane, who had a dual leech amulet (Crescent Moon), life leech on both rings (BK and Carrion Wind), a helmet that had life steal (perfect upgraded Vampire Gaze with Um rune), and a belt (perfect String of Ears). That came out to something like 30%+ and he therefore didn't have much trouble staying alive when combined with upgraded ethereal Shaftstop or Guardian Angel or 1.07 Arkaine's Valor and an upgraded Herald of Zakarum or a Stormshield (running around with 75% damage reduce was fun but cost me some skill levels). My weapon didn't even have lifesteal, it was something like a 40/15 Schaefer's Hammer. I eventually replaced my amulet because the mana steal from Vampire Gaze was more than enough (I was only using Zeal and Charge with auras) and the extra life steal wasn't doing much offensively. I think I switched to Highlord's Wrath or Cat's Eye, I don't remember which.

    I think to make % lifesteal viable, all that's needed is some number tweaking. This can be done in a few ways or by combining methods: 1) Reduce % life steal available on items, 2) Reduce the number of items that can roll with % life steal, 3) Make lifesteal % diminish as it's stacked above....say 20%, 4) Instead of giving all the stolen life instantly, lifesteal life generation works over time like over 2 to 4 seconds or something, with this as a rollable stat.

    For instance, you find a unique ring that gives you 2% to 4% life steal over 2 to 4 seconds, you could roll one that has 4% life steal but takes 4 seconds to restore the life, but you want 4%/2 LS since that's the best roll. Life steal can keep its role in letting you survive and top off your health while not drastically giving as much of a combat advantage, yet still making it feel like you sometimes have to engage even at lower health (and can successfully do so against mobs that aren't too big/strong). This is a novel fix that I think would work to balance life steal and will also increase the rarity of top-of-the-line lifesteal items.

    The problem is that life on hit doesn't work well with the massive numbers being casually tossed about in D3 and that life on hit favors rapidly doing multiple weaker hits over big, slow ones. Combat is also way faster in D3, so if you're going to die, there's a good chance you won't be able to avoid it, anyway. And at high rift levels, a lot of stuff oneshots or twoshots you so the life on hit doesn't do much to heal you.
    01/14/2019 12:04 PMPosted by clueso
    01/13/2019 07:33 PMPosted by Pestilence
    Yeah. I really missed my rend barb. I thought it was a fun play style that had your barb thrive off of large combat situations.

    After the change I kind of fell in love with the quake barb set when I decided to try it. I'd already fell into that play style with out thinking about it using legendaries. The set bonus was just damage gravy.

    Did you post in the wrong thread?
    Just asking, I currently do not see a correlation to what you wrote and the topic of this thread...


    01/13/2019 01:39 PMPosted by clueso
    when your Barbarian was low on health, he jumped into a big group of enemies and healed up instead of moving away from a dangerous situation.
    We still have % based life leech. They're called simplicity and mirinae gems.
    01/14/2019 05:24 PMPosted by Seamonster
    We still have % based life leech. They're called simplicity and mirinae gems.

    Mirinae is 3% of max health while there is a cooldown on the proc effect and Simplicity is just bad. Compared to going full set and optimizing that kind of build, Simplicity offers nothing, especially since on high difficulties you get oneshot or twoshot anyway, even 20% max health wouldn't keep you alive. This hardly contributes to life per hit being viable or contributes to % lifesteal being as big as it was in D2.
    01/14/2019 04:23 PMPosted by Pestilence
    01/14/2019 12:04 PMPosted by clueso
    ...
    Did you post in the wrong thread?
    Just asking, I currently do not see a correlation to what you wrote and the topic of this thread...


    01/13/2019 01:39 PMPosted by clueso
    when your Barbarian was low on health, he jumped into a big group of enemies and healed up instead of moving away from a dangerous situation.

    Yeah the original Blood Lust rune was a life steal rune. It was something really high like 9%. Granted, rend only did like 300% weapon damage back then but as it got buffed and 2h weapons got buffed and general damage got buffed and then monster density became insane in some maps you were essentially invincible as a rend barb before RoS came out. Good times.
    01/13/2019 01:39 PMPosted by clueso
    In D3 vanilla there was Life Leech, but it was removed because it was too powerful and also because it lead to counter intuitive playstyles, like that when your Barbarian was low on health, he jumped into a big group of enemies and healed up instead of moving away from a dangerous situation.

    That was replaced with Life per Hit, which first was based on the proc-coefficient and the amount of enemies hit, but later it was simplified and every attack gave the same amount of Life per Hit no matter which skill you used and how many enemies you hit. The reason for the simplification was that LpH was too unreliable when you fought against smaller groups of enemies or just single targets, so it was basically useless against bosses.

    I wanna make a different proposal (although I am totally aware that this will not be implemented, because they are not working that much anymore on D3, so this is more just for fun), which sort of is between the first version and the version of Life per Hit we have nowadays.

    Here is the idea:

  • (1) Instead of making LpH based on the proc-coefficient, let each skill have a Life-per-Hit-coefficient, which basically is a multiplier for Life per Hit.
  • (2) For the first enemy hit, the LpH-multiplier is noticeably higher than for the next few enemies you hit, for example, the first enemy you hit with Seismic Slam gives you 250% of your LpH and the next 4 or 5 enemies you hit with Seismic Slam give you 30-50% of your LpH. Than if necessary, the next enemies you hit with Seismic Slam either give you 0% of maybe something like 5% of your LpH.
  • Keep in ind that the numbers that I use here are just examples for the purpose of illustration!

    The big difference between how LpH worked in the first version of LpH and my proposal is, that unlike every enemy you hit with an attack would give you the same amount of LpH as it worked in the first version of LpH, in my proposal the first enemy you hit would give you a substantially larger amount of LpH than the next few enemies you hit, which makes LpH much more reliable against single targets or smaller groups of enemies.

    Now, it still has to be debated if all enemies you hit after the first enemy shall give you the same amount of LpH, although it would be a smaller amount of LpH than the first target, or if just the next ~4-5 enemies you hit with an attack would allow you to benefit from LpH, so that it would be capped at ~4-5 targets max or if any enemy beyond these ~4-5 targets shall still give a very small amount of LpH. Also, the amount of maximum targets can vary from skill to skill.

    So what do you think of this concept?
    Would this make the mechanics of Life per Hit more interesting and deep without being too imbalanced? Or would still prefer Life Leech or something else?


    in real a real rpg lph is obtained by using magical objects or temporary powers. See vampiric touch ... Life drain etc. Its the same with the reflect damage. See spell reflection please. Damage reflection is a power of magic shield or armor and only affects melee damage . Not a permanent power that affects distant targets.
    01/14/2019 03:18 PMPosted by Auragami
    I personally would not mind life steal % making a comeback. Part of what added to the thrill of combat in D2 was knowing sometimes you need to run in and just keep fighting literally for your life instead of running away and potting. The sense of danger was still there despite how obviously good life steal was.

    IIRC, the most leech I had was probably my Zealot wayyyy back in the day before runewords became insane

    I think another reason why % based life leech worked in D2 was because most physical based skills were not AoE skills or simply did not have a large AoE. The physical melee skills with the largest AoE were probably Whirlwind and Zeal and the largest physical ranged skill was probably Multishot and ranged skills had their life leech reduced even further iirc (or was it just things like crushing blow???), while in D3 you have skills that have a much larger AoE.

    Also, % based life leech probably also should effect spells imo, so that would have to be taken into account as well.

    01/14/2019 03:18 PMPosted by Auragami
    I think to make % lifesteal viable, all that's needed is some number tweaking. This can be done in a few ways or by combining methods: 1) Reduce % life steal available on items, 2) Reduce the number of items that can roll with % life steal, 3) Make lifesteal % diminish as it's stacked above....say 20%, 4) Instead of giving all the stolen life instantly, lifesteal life generation works over time like over 2 to 4 seconds or something, with this as a rollable stat.

    I like the idea that it can give you back the amount of life you leech over a short duration instead of immediately.
    01/14/2019 04:23 PMPosted by Pestilence
    01/14/2019 12:04 PMPosted by clueso
    ...
    Did you post in the wrong thread?
    Just asking, I currently do not see a correlation to what you wrote and the topic of this thread...


    01/13/2019 01:39 PMPosted by clueso
    when your Barbarian was low on health, he jumped into a big group of enemies and healed up instead of moving away from a dangerous situation.

    Ah, okay, I didn't see it at first...

    Join the Conversation

    Return to Forum