A Study on How Empathy Works

[IBNobody]

I needed a break from my item compendium, so I started looking at Empathy.

I haven't broken the formula, but I've found some things that made it significantly easier to reverse engineer.

Here's what I found:

* Empathy, rank 6, always has a spread of 6. The description says it "heals for 30-35", which is an average of 32.5. Even when it scales, it's always a spread of 6. 45-50! 52-57! 56-61! Because of this, it was very easy to test weapons.

Just get a weapon, start getting hit, cast Empathy, and keep doing this until you get a spread of 6. Record your min/max heals and your str/dex/int/min dam/max dam/weapon stats. No averaging or keeping track of individual samples is needed.

* Assuming the description is correct, Empathy's formula is:

<AVG EMPATHY HEAL TICK> = <BASE EMPATHY AVG> + <SOME UNKNOWN QTY>

* Increasing your overall min/max damage by 1 for the same weapons increased your healing by 1. I only tested Rifles and unarmed, though.

* It looks like item level in addition to min/max damage contributes. If I matched up min damage or max damage values, my lv18 rare rifle always healed for less than my lv 20 common rifle. Likewise, my lv 15 common rifle didn't fit on the healing curve of my lv 20 rifle either. If it were just min/max/avg damage, I should've hit the curve.

However... I did have a lv 19 common rifle with almost exactly the same stats as my lv 20 rifle (+1 min/max dam). When I used it, I WAS able to hit the curve.

Here is my data:

https://docs.google.com/spreadsheet/pub?hl=en_US&hl=en_US&key=0AnQBvFf7X3CndGxCeEUwc2VleFlNS1c1bjZmQU45dmc&single=true&gid=1&output=html

-----

That's all the research I can do for the next couple of days. I hope I've planted the seed of curiosity. It should be very easy to track Empathy, given what I found out about fixed spreads.

Things to try:
Do different weapon types have different curves?
Does the damage implant increases your empathy ticks by 5?

Figuring out how Empathy works is critical for us to understand how skills work in general. I bet Leech, Wither, and Transference all have fixed damage spreads.

[WhoIsThis]

If empathy is anything like heal, then we would expect that adding more points to int would increase the empathy average heal.

[IBNobody]

It is not directly based on INT. INT contributes indirectly to your damage, as does STR and DEX.

[Hullukko]

Firstly, wonderful that you've taken the time to do that. Could you please include the weapon min/max damages there, too?

I think what WhoIsThis was saying just that. In PL int directly contributes to the base damage and thus also has a direct effect on spells like heal, the relation is not 1:1, but it's there. And it does so in SL, too. (whether the effect is seen as direct or indirect via base damage is inconsequential here, the coefficient is the one that we're after)

I'm currently doing similar analysis myself (from a slightly different point of view and slightly different reason), I'll post that when I get it done. Although my data collection is still a work in process, I can confirm that the spread of 5 remains in empathy regardless of int and weapon choice.

And for the record when I did similar analysis in PL, there were few anomalies in how the weapon base damage got included to the spells between various weapon types. Iirc, the 55 axe should've contributed more than the sword when calculated using weapon base damage alone, but that was not the case in practise. There are many possible explanations and it was never fully broken down afaik (or at least not before gcd got introduced and I quit). There could be hidden attributes, type based factors, time-to-hit relation, etc.

[noneo]

I was literally working on the same exact thing a few days ago... Strange...

Anyway, here is the data that I collected with multiple different weapons:
https://docs.google.com/spreadsheet/ccc?key=0AqePsv2s_tg8dFZWSGlsbHg1RHhvSWRGWkd6SE9MS Gc&hl=en_US

I saw the same general spread as you did, i.e. the "6 point" spread. However, when I had nothing equipped I got a span of 40 - 48. now this might have been from my skill not having a full third tick (because my HP was already too high), but I payed very close attention to try and avoid this..

Here are my findings:



Base Damage

14 - 15 (14.5)

45.4117647058824

35 - 58 (46.5)

36 - 63 (49.5)

40 - 73 (56.5)



Weapon Used

No Equips

"Pioneer's Pain
Grippers"

Rusted Pistol

"Veteran's Beta and
Beyond Sniper Rifle"

Scavenger's Rifle



Skill

lvl 6 Empathy

Empathy - Lvl 6

Empathy - Lvl 6

Empathy - Lvl 6

Empathy - Lvl 6



Description

Heal 30 - 35 every 3 seconds

Heal 30 - 35 every 3 seconds

Heal 30 - 35 every 3 seconds

Heal 30 - 35 every 3 seconds

Heal 30 - 35 every 3 seconds



Estimated Amount

34 - 50 (42)

58 - 81 (69.5)

65 - 93 (79)

66 - 98 (82)

70 - 108 (89)



Real AVG

45.4117647058824

57.5818181818182

64.1777777777778

64.1818181818182

66.8863636363636



Real Max

48

60

67

66

69



Real Min

40

55

62

62

64




The 4 tests of which I equipped a weapon all seemed to follow a good general curve. However, once I incorporated the "no-equip" option, the curve got destroyed. My theory is that there is a base empathy heal with no equips (45 point average or so), and then everything after that (when you are equipping more base damage) follows a set curve.

Also, pertaining to your “six point span theory” I do think this is a good guestimate for what the base damage will give (granted you find the min and max) but keep in mind that weapons are not a whole integer number all the time. You could have a weapon with 42.25 – 51.75 damage and it might show up as 42 – 52 on your screen. That is why on occasion when you go to assign another weapon to your character it will show up a “+0” in green (this just means it is not adding a full point more).

[Hullukko]

I've done about half a dozen unequiped data points at different int values. And then the same thing with claws and repulsor. At max int I've done the same test with pretty much every rifle available, a couple claws and few other random engineer weapons. Not once did I see the empathy spread go beyond the aforementioned delta 5. On few insignificant data points I got tired of waiting for the 6th as I was able determine which value was missing anyway.

I strongly believe you've made a mistake in your 40-48 span. You've must have hit max health or just looked it up wrong, or something. If you are able reproduce it, I would really like to hear how.

My analysis is yet to come though. I'm just as perplexed as I was when I did this same thing with PL (where it was a hell of a lot easier, because the skill page showed the spell damages and they simply changed when you changed weapons).

What is certain thus far, is that the base damage, i.e. the unequiped base damage, deduced from int (roughly 15.86 int for 1 point of damage) has a 1:1 additive relation with the empathy healing power. Another additive relation comes from the weapon induced base damage (the minimum especially), but that comes with a smaller coefficient, which somewhere along the lines of 0.9. Higher max damage, and this it what puzzles me, would appear to have a very slight yet non-negligible negative coefficient.
But roughly speaking the healing power (and other skills with their own coeffs, too) are affected by the stat determined naked base damage directly and on top of that most of the weapon minimum damage is added. The repulsor heals about 3 points more than salvage claws (and the claws +7 int is there included).

I'm going to try to do this without looking into pistols and repeaters and what not, because data collection is just too big a burden when a daughter is climbing all over me while I'm standing naked in from of some mobs spamming healing and trying to see all the little green numbers that fly from my toon's head.

[shannong]

So the takeaway here is:

A) Higher INT means bigger heals
B) Rifles provide bigger heals than the class-specific Engineer gloves, because the rifles always have the highest min/max out of all weapon types.

I mean, we all knew this intuitively, but here's the testing data to back it up and explain it.

[IBNobody]

So the takeaway here is:

A) Higher INT means bigger heals
B) Rifles provide bigger heals than the class-specific Engineer gloves, because the rifles always have the highest min/max out of all weapon types.

I mean, we all knew this intuitively, but here's the testing data to back it up and explain it.

Actually,

A) Higher STR/DEX/INT means more damage and bigger heals, with INT providing the most bang for the buck.
B) Rifles provide bigger heals than the class-specific Engineer gloves, because the rifles always have the highest min/max out of all weapon types.
C) Higher weapon levels provide more healing than lower weapon levels of the same weapon type, assuming equal min/max damage.

[roguedubb]

Just did some testing with the damage implant.
I tested each setup for 8 minutes each.
[level 30, all VI Shocktite gear, rev buff on, using Atomic Dispenser]

DAMAGE IMPLANT
86 to 90

WITHOUT IMPLANT
80 to 85

[Kahlua]

Just did some testing with the damage implant.
I tested each setup for 8 minutes each.
[level 30, all VI Shocktite gear, rev buff on, using Atomic Dispenser]

DAMAGE IMPLANT
86 to 90

WITHOUT IMPLANT
80 to 85

Maybe I'm not understanding the thread, and I didn't read everyone's post for content, but isn't this thread about healing efficiency and how different weapons/stats effect it?
Or maybe I'm just not understanding your post, Rogue.

Are these numbers the heal range you received with the implant? That would make sense to me because as with the rifle, higher base damage = higher heals.

[IBNobody]

Maybe I'm not understanding the thread, and I didn't read everyone's post for content, but isn't this thread about healing efficiency and how different weapons/stats effect it?
Or maybe I'm just not understanding your post, Rogue.

Are these numbers the heal range you received with the implant? That would make sense to me because as with the rifle, higher base damage = higher heals.

He was just proving out that the damage implant really does boost heals.

BTW: Thanks, Roguedubb!

[Gazroast]

High INT = bigger heals (Thats why I dress My Engi like this: http://www.spacetimestudios.com/showthread.php?39880-My-Lvl-30-Engi.&highlight=engi )

I have found that I can get 111DPS and still average 83 heal per tick(Incinerator)
With Black Widow and its higher damage, but much lower DPS and INT it only heals on average 2.5 more per tick..

My Engi has only 15% lower dps than an Op with 50% more Defence.
25% less armour than a comm. and depending on weapon either 30% more DPS or the same (If they are using Incinerator)
However I can HEAL my self, quickly and for free in the middle of a battle.. PRICELESS!!

*Engi Love*

[Gazroast]

Question:

If the description for every level of Empathy is the same, does it even make a difference making it or just leaving it as a 1 like revive?

Has anyone tested this?

[Kahlua]

He was just proving out that the damage implant really does boost heals.

BTW: Thanks, Roguedubb!
ahhh gotcha. Seemed like it could have been damage he did with the implant; in which case I'd say "well DUH"
Good info to know though.

[Hullukko]

I have found that I can get 111DPS and still average 83 heal per tick(Incinerator)
With Black Widow and its higher damage, but much lower DPS and INT it only heals on average 2.5 more per tick..


You make sound like the incinerator would be the winner, but it's not exactly so. You have to factor in the fact that the higher base damage from the widow is carried over to every skill. An engineer will do a lot more with the widow than with the incinerator. Just stop staring at the stats page theoretical dps, which is faulty in many ways (skill damage, enemy armor, "first hit effect").

If you want the numbers in practice: here (http://www.spacetimestudios.com/showthread.php?37663-Analysis-on-engineer-weapons-and-skills-in-practice)
(and again remember that the difference is carried to every healing and damaging spell. You don't see many 20k+ kills engineers running about with them whimpery incinerators. Here's why.

[Hullukko]

Question:

If the description for every level of Empathy is the same, does it even make a difference making it or just leaving it as a 1 like revive?

Has anyone tested this?

Odd question, because the description for every level of Empaty is not the same. The healing base points increase on every level.

But an interesting question nonetheless, because level 6 heals 30-35, but that ticks the well known 80something in practice. Could those points be sacrificed to an offensive skill. At least in sloucho the empathy certainly doesn't need to be maxed out (in groups of properly dressed 30/31 players).