Spiral Knights - Monster Damage

Monster Damage

25 replies [Last post]

Sun, 04/28/2013 - 12:03

Hexzyle

As most people are already doing stuff like figuring out defense, resistances, and monster health (cough Zeddy cough) I thought I'd work on something that I don't think anyone else is doing: Monster Damage. It's shown to be relatively simple from the tests that I'm doing (starting at Tier 1) but I might encounter complexities later on with split damages and advanced rules in later tiers.

If you have a Drake Scale Set or a Divine Set, you can help me out. Either of these can be used to test damage values in tier 1, the former can be used to test Undead and Fiend monster damages in tier 2 and 3, and the Divine Set can be used to test Beast and Slime damages in tier 2 and 3.

From what I've observed, Strata 1 monsters don't suffer any damage penalty for having a status attached to their attacks, and Strata 5 "pure specialized" attacks still may deal partial normal. (but I'll need to look into that further)

I haven't got a google doc for this yet (will be writing up one tomorrow if I can figure out how the darned thing works ^_^) but I'll post what little results I have here anyway.

On Depth 1
Damage distribution is categorized into groups: 70%, 90%, 100%, 125%, 150%.
Minor attacks deal 70% (14 on depth 1)
Ranged attacks deal 90% (18 on depth 1)
Melee attacks deal 100% (20 on depth 1)
Elite attacks deal 125% (25 on depth 1)
Heavy attacks deal 150% (30 on depth 1)

So far my current results for each family:

Gun Puppy: 90%
Rocket Puppy: ?
Mecha Knight (Basic): 100%
Mecha Knight (Charge): 100%
Mecha Knight (Bullet): ?
Mecha Knight (Bomb): ?
Scuttlebot: 90%
Lumber: 150%
Retrode (Swipe): ?
Retrode (Beam): ?

Thwacker (Basic1): 100%
Thwacker (Basic2): ?
Thwacker (Ranged): ?
Thwacker (Spin): ?
Knocker (Basic): 100%
Knocker (Charge): ?
Demo: ?

Wolver: ?
Wolver (Boosted): ?
Alpha Wolver: 125%

Jelly Cube (Basic): 100%
Jelly Cube (Spire): ?
Ice Cube (frozen): 125%
Ice Cube (melted): ?
Toxigel (Basic): 100%
Toxigel (Shoot): ?
Quicksilver: ?
Quicksilver (Charge): ?
Lichen: ?
Lichen Colony (Basic): ?
Lichen Colony (Shoot): ?
Giant Lichen Colony (Basic): ?
Giant Lichen Colony (Shoot): ?
Giant Lichen Colony (Thorns): ?
Polyp: 90%

Howlitzer (Shoot): 90%
Howlitzer (Kamikaze): 125%
Zombie (Swipe): 100%
Zombie (Leap): 70%
Spookat (Bite): 90%
Spookat (Shoot): 70%

Devilite: 90%
Overtimer (Melee): 125%
Overtimer (Ranged): 125%
Greaver (wing): 100%
Greaver (bullet): ?
Trojan (Slam): 150%
Trojan (Slam, Boosted): 250%
Trojan (Dash): 125%
Trojan (Dash, Boosted): 190%
Trojan (Boost): ?

Blast Box: 125%
Shankle: 100%
Spikes: 90%
Grimalkins: 150%
Mini Creep Cake: 70%
Shufflebot: 70%

I'm trying to figure out how much damage respawning blocks deal. I got smacked by one for 50 damage in a Candlestick Keep (it was one of those onese that respawn after 4 seconds) and that didn't seem right, so I'm thinking that the damage a respawning block causes might be inversely proportionate to how quickly it respawns.

You can determine how much damage you take from a certain attack by getting hit by it and counting how much percentage of your maximum health you lose. Getting hit by the same attack multiple times in a row and recording each percentage decrease will further increase the accuracy of the results. Submit the string of numbers here if you don't want to do the math yourself. If you do want to do the math, you find the average percentage that you got hit by (don't worry if it's a decimal) and multiply it by your maximum health (in bars) x 40.
For this method, it's better to start at 100% health, but it isn't entirely necessary.

Zeddy said that I could probably use the pixels of the health bar to determine how much health is lost from each attack but I'll have to look into that. If you want to send me screenshots of your health before and after attacks instead of recording percentage drops, that'll be useful too.

Sun, 04/28/2013 - 12:37

Bopp

great; keep it coming

Great. Please keep the data coming. In particular, damage at depth 19 or depth 24 is probably more relevant to your audience, than damage at depth 1.

Sun, 04/28/2013 - 12:54

Zeddy

How much damage do the shufflebots at GTH do? I'm betting those are 70%. Their health is 75%.

Mon, 04/29/2013 - 00:32

Hexzyle

@BoppI only have Drake

@Bopp
I only have Drake Scale, so I can only gather data from Undead and Fiend in later tiers, while in tier 1 I can test all monsters. I tried using Ancient Plate Mail, but for some reason I took less damage from a Kat bullet at depth 19 than I did with Drake Scale (197 instead of 274), leading me to believe that monsters in Strata 5 still do partial normal. (going on a rough estimate from what I've seen that APM resists 75% of normal damage, I'd say that "pure specialized" attacks at depth 19 still deal about 40% normal, which I assume gradually reaches 0% at depth 24)

Anyway, calculating at depth 1 is easy, gives accurate results, which can be scaled up to tier 2 and 3 for base approximations of damage.

@Zeddy
They deal 14 damage (70%)

Mon, 04/29/2013 - 04:49

Bopp

how?

Anyway, calculating at depth 1 is easy, gives accurate results, which can be scaled up to tier 2 and 3 for base approximations of damage.

I must have missed this. How does the damage scale up on deeper tiers?

Mon, 04/29/2013 - 05:46

Hexzyle

@Bopp

Well, for example, on depth 1 a Spookat's bite deals 18 damage, a Zombie's swipe deals 20, and a overtimer's attack deals 25.

On depth 19, a Spookat's bite deals roughly 272 damage, a Zombie's swipe deals roughly 296, and using the same ratio we can assume an overtimer's attack deals roughly 375.

Mon, 04/29/2013 - 06:08

Bopp

damage scales, but changes type?

Okay. So we need to measure at least a few things at deeper levels, and then make some inferences. You're saying that all monsters scale up by the same factor. That seems possible. But the content of their attacks (in terms of damage type) changes? So the normal part of their attack scales up very differently from the shadow part (say)?

By the way, I'm not arguing with you. I'm just trying to clarify things, on the path to figuring out how this calculation might be designed and implemented.

Mon, 04/29/2013 - 07:37

Hexzyle

Well of course, the

Well of course, the normal/specialized construction of attacks would change as you get deeper. I'm fairly sure it is a smooth transition, and not in large jumps. Example:

Monster in question is a construct.
A is the monster's melee attack.
B is the monster's ranged attack.

At depth 1-7:
both A and B deal 100% normal

At depth 9:
A deals 90% normal, 10% elemental
B deals 80% normal, 20% elemental

At depth 17:
A deals 75% normal, 25% elemental
B deals 50% normal, 50% elemental

At depth 19:
A deals 70% normal, 30% elemental
B deals 40% normal, 60% elemental

At depth 24:
A deals 50% normal, 50% elemental
B deals 0% normal, 100% elemental

This is just an estimate though. Getting accurate details on the split damage would be very difficult: As you go deeper, monster damage scales, armor defense scales, and damage split scales. All three variables are dependant on eachother, and seeing as we only know one, we're kinda stuck. If x + y + z = 10, we can't figure out both y and z when we only know what x is.

Mon, 04/29/2013 - 17:23

Zeddy

@Hexzyle

Excellent. That means you can find out all the damage amounts at depths 13 and 23.

Mon, 04/29/2013 - 21:19

Hexzyle

@Zeddy

Say what?

Mon, 04/29/2013 - 22:29

#10

Zeddy

@Hexzyle

That's the depths for T2 and T3 Guild Training Halls. Observe the damage the shuffles do there and simply scale it to figure the other values out.

By the way, are you certain that it's 70% damage as opposed to 75% damage, rounded down? With numbers that low, it could really be either of those.

Tue, 04/30/2013 - 20:36

#11

Hexzyle

@Zeddy

Could be. I'll run a few values through. Although 75% of 20 is neatly 15.

Tue, 09/17/2013 - 02:59

#12

Zeddy

I use a Spark of Life on this thread

I want to figure out how much defence a given piece of armour gives. We know monster defence, we know monster health, we know our damage, we know our health, but our defence remains a mystery.

Specficially, will the same amount of armour give the same amount of defence for two different depths? I'm certain that by using this method with dragon scale and with a single piece of defensive armour, we'll be able to figure this out somehow.

I'm hoping we can get a clear picture of how much defence is provided by a single 2* trinket and see if that is the smallest interval of defence possible, kind of like how it is with status trinkets. Probably not, but it would be really nice and easy if that was the case.

Tue, 10/01/2013 - 02:53

#13

Zeddy

Normal and Elite

On depth 1, full Dragon Scale, penta-heart trinket and heart perk for my sprite. A total of 19 pips.

On Normal, Spookat bite left me at 26% health after 40 bites. This indicated 14 damage, which is consistent with a 20% decrease. (20 * 80% * 90% = 14.4)
On Elite, Spookat bite left me at 34% health after 20 bites. This indicated 25 damage, which is consistent with a 40% increase. (20 * 140% * 90% = 25.2)

On depth 19, full Dragon Scale, penta-heart trinket and heart perk for my sprite. A total of 27 of pips.

On Normal, Devilite chair left me at 16% health after four hits.
On Advanced, Devilite chair left me at 24% health after three hits.
On Elite, Devilite chair left me at 22% health after three hits.

On depth 24, full Dragon Scale, penta-heart trinket and heart perk for my sprite. A total of 27 of pips.

On Normal, poison Devilite canister left me at 34% health after two hits.
On Advanced, poison Devilite canister left me at 23% health after two hits.
On Elite, poison Devilite canister left me at 23% health after two hits.

Yes, I made sure I wasn't poisoned or anything when I got hit.

Tue, 10/01/2013 - 19:37

#14

Coelydragon

@Zeddy

As you analyze this, are you assuming dragon scale has no hidden normal/shadow defense?

Tue, 10/01/2013 - 23:50

#15

Zeddy

@Coelydragon

If dragon scale has any of those, then the amount is not affected by scorching barrier's 75% defence boost, nor getting poisoned.

Sat, 10/19/2013 - 05:34

#16

Zeddy

I've discovered that the health percentage is rounded down, once again thanks to Lockdown. That means we can narrow down damage ranges by getting attacked with various amounts of health amounts and adding up the results.

For example, D24 gremlin thwackers inflict 400-401 damage to someone wearing Chaos.

I've calculated the raw damage of D24 zombies to be 525-528 by taking hits with permutations of Scarlet Shield and the sprite perk for +6 health. Any submissions to this topic with varying degrees of health are welcome.

Tue, 11/12/2013 - 07:41

#17

Kathrine-Dragon

@zeddy

Random thought: would it help to use Seerus ninjas to find pure damage values?

Tue, 11/12/2013 - 21:32

#18

Zeddy

Slightly. However, none that I know of appear on depth 24. For every enemy except those dealing elemental damage, pure damage can be found simply by wearing Dragon Scale or Divine.

Sun, 12/08/2013 - 08:52

#19

Skepticraven

@Zeddy

I'm interested in how you are doing the actual math behind the narrowing of the armor calculations. The only method that I could back calculate what math you seemed to be doing is...

Record pips of health (A) and % display when damaged (B).
Damage range is:
(A*40)-ceil(A*40*B) to (A*40)-ceil(A*40*(B-0.01))-1
Edit: Added the rounding corrections as mentioned in post #22

If I understood your calculations correctly (assumption being % shown always rounds down as noted with having 1% at almost no hp and 99% at almost max).
Reason being... I've started recording some data in order to attempt to better understand armor. Stupidly, the divine set I recently crafted got a UV med normal (which may turn out useful in testing half armors with UV bonuses).

I may also make another thread to split off the discussion about armor and the last piece of the puzzle: shields.

Sat, 12/07/2013 - 17:58

#20

Zeddy

@Skepticraven

Like I said in the post, I take hits while having various amounts of max HP. To calculate the gremlin damage, I used these numbers:
With a max health of 15 pips, I had 33% health left. (33% of 15 pips = 198 to 34% of 15 pips - 1 = 203. Invert to get damage range 397-402)
With a max health of 17 pips, I had 41% health left. (278.8 rounded up to 279 to 285.6, rounded down to 285. Invert to get damage range 395-401)
With a max health of 27 pips, I had 62% health left. (670 to 679, invert to get damage range 401-410)

Maximum minimum: 401.
Minimum maximum: 401.

Gremlins do 401 damage.

It seems I had a more accurate number than I first estimated. I'll have to search for bugs in the program I wrote to crunch these numbers for me.

Sat, 12/07/2013 - 21:55

#21

Skepticraven

@Zeddy

Thanks. I was missing the -1 for the high range (which makes sense). Although I'm kinda confused with the rounding. Initially it makes sense, but when looking at the preliminary data that I showed... damage from a shufflebot in D23 against a divine veil and vitasuit gives me a range of 218.4(high)-218.6(low). I'll probably have to double check these numbers, but it's rather odd because they appeared multiple times.

Sun, 12/08/2013 - 02:01

#22

Zeddy

@Skepticraven

That does appear odd. Let me know what comes out of your doublecheck. It's possible the percentage makes weird rounding at the border numbers.

If you could supply the raw percentages and health pips, I could have my program crunch through them for you.

Also, I miscalculated the 27 pips, 62% health thing. The actual range is 669.4 to 680.4, not sure how that ended at 679 when I did the maths last night.

So gremlin damage is still 400-401.

EDIT: I plucked the health/percentage data from your chart and shoved it into my program, and it spat back out 218-218. One of us have been doing the rounding wrong and I'm not sure which one it is. Did you perform the rounding before or after subtracting remaining health from total health? You round down the max health, not the max damage.

The maths part of my program looks something like this:
lowHealth = math.ceil(lowHealth)
highHealth = math.ceil(highHealth) - 1

my.lowest = max(my.lowest, fullHealth - highHealth)
my.highest = min(my.highest, fullHealth - lowHealth)

Sun, 12/08/2013 - 07:38

#23

Skepticraven

@Zeddy

That 218.4 and 218.6 are without rounding. If I tack in your roundings, I get the same answer. I guess it makes sense.

Thu, 12/26/2013 - 15:53

#24

Skepticraven

Sorry to revive this discussion again but...

I just tested the unarmored (divine/drake) vs deathmarked theory, since it is currently winterfest.
In the "Save Winterfest" mission, I wore drake scale and compared damages for a layoafer (ice devilite).
While not deathmarked by the grinchlin stalker, a layoafer snowball deals 412 damage on depth 24 on elite.
While deathmarked by the grinchlin stalker, a layoafer snowball deals 412 damage on depth 24 on elite. [I swapped to divine to make sure I was deathmarked for the damage.]

This confirms that there is no inherent PvE armor and the values observed in this thread are true damage.
If you wish to replicate this data while you still can, it can be found here.

This same method can be used to find Torcher and Grinchlin Thwacker true damage while the event is still taking place.

Sat, 01/25/2014 - 14:03

#25

Skepticraven

Shankle Damage

I'm unsure if this research ever reached a conclusion, but I have been recording pure damage types as they change by depth.
Specifically Shankles.

Can see how it changes on almost every depth here.

Search

Monster Damage