life on hit question

[EnafAequi]

So I began a char, a warden, and used the skill that double the attack speed, I had 2 life on hit, and I was getting 2 life way less often than every hit. Something must be wrong.

[Squarebit]

There's actually a cooldown of 0.5 seconds on health/mana on hit. This is because for example the Warden could more or less become invincible with the high attack speeds otherwise.

[EnafAequi]

Hum.. I'm a bit puzzled about this.

It soundss like bad design, you have to implicitly restrain a feature (the player keeps thinking if he keeps increasing his attack speed LoH will follow, while this is not true), and it destroy interesting sinergies, and play style that differs from others classes.

It's like Diablo 3 turning dexterity from evasion provider to armor provider because they didn't wanted to go to the trouble of tunning their game around evasion.

I imagine the warden is mainly a ranged class, fast clearer, high mobility, squishy

Having a health per hit without cooldown does not make him invincible, on the contrary that's what the game is all about I think: he will be more or less in danger depending who he is facing, make sure that those dangers are present and effective.
Regardless of attack speed, he will still be increadibly weak versus thorn monsters higher than the life on hit, and since he is not supposed to have as much defense as melee, even if he regenerate fast, he will be weak against one shot crits. Or also against large packs or teleporting melee monsters. If these threats are present, limiting health per hit seems like limiting a fun feature.

Maybe instead change the skill with double attack speed saying that life on hit is halved with this skill for example?

Or even, what the hell, create a monster mod that make life on hit damage the player instead of healing him? Something called "wicked thorns"

I don't know, it's not a big deal, it's very specific, but I feel like it's not how it should be. Or at the very the player should be informed about this, to make enlighten choices.

(Great game btw)

[EnafAequi]

My bad !

I just saw that the 0.5 cooldown is stated under life on hit description of the help section.

[BEANS BEANS BEANS BEANS BEANS]

I'm somewhat inclined to agree with EnafAequi in terms of game design.

I think there should be no internal cooldown, but perhaps the damage scaling should be done more heavily. In this way, a heavy attack-speed build has a viable source of sustain by using up affixes (a trade-off). Having an internal cooldown as it is limits build diversity, but having LoH/MoH in itself requires sacrificing affixes, so it makes the stat much less appealing.

As an example, attacking 10 times with a 100% damage skill in 1 second with 10 LoH should regen 100 health, as should 5 times with a 200%, or once with a 1000% (or it could scale even harder, so with 200% it regens 150 health, and 1000% it regens 500 health, for example).

Fast attack builds tend to be less appealing than massive alpha builds in these sorts of games anyway, so I think having an ICD on on-hit effects just adds another nail in the coffin, as it were.

Regardless of what direction you take LoH/MoH, I look forward to it!

[EnafAequi]

What you discribing (regen scale with damage) is leeching not life on hit and it works very differently, there is no leeching in this game as far as I know.

But you have a point, if to make life on hit works, a cooldown is needed, then maybe leech instead would be better.

[BEANS BEANS BEANS BEANS BEANS]

Originally posted by EnafAequi:

What you discribing (regen scale with damage) is leeching not life on hit and it works very differently, there is no leeching in this game as far as I know.

But you have a point, if to make life on hit works, a cooldown is needed, then maybe leech instead would be better.

Life on hit scales with damage in this game - that's what the help info says. "The larger your hit, the more you will restore".

[Squarebit]

Hmm, currently the actual health and mana restored is also depending on how big your hit was in relation to damage (so for example crits restore more). The reason for the cooldown is not only high attack speed, but also the fact that skills hitting multiple enemies instantly counts each enemy hit as a hit, and would trigger the health/mana on hit effect. So putting a Firebomb on a group of 10x zombies would get you 10 procs instantly.

Reading this has made me reconsider a bit, though, so I will experiment with lowering the cooldown and some other ideas... it's possibly the most difficult stat to balance, I think!

[EnafAequi]

I don't understand how it scales with damage... it's a flat value on equipment, i'm confused.

"it's possibly the most difficult stat to balance, I think!"

sure is one of the toughest, even AAA games have trouble to balance sources of regen, it's a tricky spot

[Squarebit]

Originally posted by EnafAequi:

I don't understand how it scales with damage... it's a flat value on equipment, i'm confused.

Consider that value the minimum value that will be restored, and the larger your hit above your Damage stat, the more will be restored. For example striking for twice as hard as your Damage stat will get you about 140% of the Health/Mana on hit, but the scale isn't linear.

It looks something like this: http://i.imgur.com/rnpNLy6.png
The X is multiplication of damage (eg 10 means scoring a hit 10 times your damage stat) and the Y is multiplication of the Health/Mana on hit. So if you hit 10x as much as your damage, you'll get about 3x as much Health/Mana on hit as your stat indicates.

(It can't ge blow 100% though even though the curve looks like it.)

[EnafAequi]

thanks, got it, LoH affected by damage multiplier with diminishing returns

[Squarebit]

Yep that's it :)

I'm gonna set the CD much lower (probably around 0.1 sec) with the upcoming preview build of the new skills, and we'll see how it feels! Practical testing is often the best way to go about this stuff, I find :)

[EnafAequi]

You could use steam test build if you want the community opinion about how some quick changes feel
I don't know if there are enough players for it to be really useful tho

[Squarebit]

yeah I would but I'm like halfway through rewriting the entire skill tree stuff, so I'm setting up a separate branch for it with special builds as it would render existing saves incompatible with the current version if played (this build will use a separate save directory no worries there)