Spiral Knights - Analysis of Shield Defenses (and determining where the Grand Tortoise stands)

Analysis of Shield Defenses (and determining where the Grand Tortoise stands)

104 Réponses [Dernière contribution]

jeu, 06/11/2015 - 09:19

Hexzyle

I'm curious to how much of an effect a shield's level has on its defensive ability. Specifically, the high normal defense shields (and their ability to block the Trojan overhead swing) And I'd like to put to rest whether or not the Grand Tortoise really is a reskin of the Ancient Plate Shield.

Does anyone here have a Grand Tortoise, Ancient Plate Shield, or Volcanic Plate Shield and are willing to do some testing in Elite FSC?

IGN: Hex-Chan

Shield Bar is 360 pixels:
After taking a hit from an unboosted Trojan overhead slam, the following shields were left with:

Grand Tortoise (Level 5)
D24: 191 pixels remaining
D25: 191 pixels remaining
D26: 191 pixels remaining
D27: 191 pixels remaining
D27: 75 pixels remaining (White Trojan)

Grand Tortoise (level 10) [Credit to Flowchart for the data]
D27: 78 pixels remaining (White Trojan)

Ancient Plate Shield (Level 10 + Normal: Low)
D27: 89 pixels remaining (White Trojan)

Ancient Plate Shield (Level 10) [Credit to Jenovasynthese for the data]
D24: 260 pixels remaining
D25: 258 pixels remaining
D26: 255 pixels remaining
D27: 252 pixels remaining
D27: 88 pixels remaining (White Trojan)

Ironmight Plate Shield (Level 4) [Credit to Holy-Nightmare for the data, and Jenova for analyzing it]
D24: 190 pixels remaining
D25: 190 pixels remaining
D26: 190 pixels remaining
D27: 190 pixels remaining (±3 due to image imprecision)
D27: 75 pixels remaining (White Trojan)

If possible, I'd love to see a number of shields with different normal defenses tested, to really highlight the effectiveness of Plate Shields versus Trojans.

Heavy Plate Shield
No data

Omega Shell (Level ?) [Credit to Counttab for the data]
D26: 98 pixels remaining (±1 due to gui shake)
D27: 57 pixels remaining (White Trojan)

Aegis
No data

Crest of Almire (Level 10) [Credit to Jenovasynthese for the data]
D24: 30 pixels remaining
D25: 30 pixels remaining
D26: 31 pixels remaining
D27: 31 pixels remaining
D27: 43 pixels remaining (White Trojan)

Example screenshot
APS Level 10, Normal Low UV: D27 Elite

jeu, 06/11/2015 - 18:01

Bopp

interesting

I can't help you with this one. I just wanted to let you know that someone's reading the thread. Good luck.

jeu, 06/11/2015 - 20:37

Fe-Tarkus

If OOO removed my rank restrictions, I could upgrade my Heavy Plate Shield to a VPS or Ironmight, but nope. I can't help either.

ven, 06/12/2015 - 00:03

Exiled-Gremlin

I have an Ironmight Plate

I have an Ironmight Plate Shield, I think it has the same normal defensive abilities as Ancient Plate Shield, mail me if I can help you.

ven, 06/12/2015 - 01:41

Hexzyle

If OOO removed my rank restrictions, I could upgrade my Heavy Plate Shield to a VPS or Ironmight, but nope. I can't help either.

I would like to see Trojan slam's effects on Heavy Plate Shield too, this would be useful in determining the defense difference between 4 star and 5 star.

ven, 06/12/2015 - 04:05

Jenovasforumchar

i join in

I am fallowing this, too, but I was expecting that there are plenty of results by now.

I have myselfe fallowing shields between the Heavy Shields:
-APS, heat10
-Gorgomega, heat 5 or 6, had to Login to confirm
-Omegaward, heat (below 6, have to Login for exact Level)

Just tell what exactly you Need. I assume
-only FSC Trojan's Overhead-slamm to beginn with.
-While Shield has Max durability
-and playing on Elite Mode

I will do some testing later and give in each deapth 3 Screens per tested trojan.

sam, 06/13/2015 - 02:33

Hexzyle

Just tell what exactly you Need. I assume
-only FSC Trojan's Overhead-slamm to beginn with.
-While Shield has Max durability
-and playing on Elite Mode

Yes, this is correct. That is because:
1. Only the overhead slam deals primarily normal damage (as well as being the attack you want to induce the most: the dash puts you and the Trojan in a bad position, and the Attack Up! makes the next attack dangerous)
2. If the shield didn't have max durability, there's no way we'd be able to determine how much damage the Trojan's attack did
3. On Elite mode, monsters deal the most damage (presumably). Since I want to see how shields are affected when in the really high numbers (high normal defense) it's useful for monsters to deal the most normal damage possible (since apparently defense has a diminishing effect when the ratio of defense to attack is quite high)

sam, 06/13/2015 - 11:45

Flowchart

I have a fully heated Grand Tortoise just don't feel like doing FSC right now (if I'm going to grind, might as well be for kats). If nobody gets to it then I'll post it later.

sam, 06/13/2015 - 12:45

Hexzyle

I sort of... can't get to doing this at the moment, my computer is having problems.

dim, 06/14/2015 - 08:42

Skepticraven

↓

Be sure to reference my Shield HP Chart. D27 says that 100% of the shield bar is 825 hp. It may also be of interest to figure out a trojan's damage on that depth. My methods seem to stand pretty strong for doing that... they just take awhile to minimize down to the exact value.

dim, 06/14/2015 - 12:24

#10

Hexzyle

Thanks Skepti. I'm sure this information will come in handy.

lun, 06/15/2015 - 13:05

#11

Jenovasforumchar

'Data! Marvellous shiney data!' -by a certain someone

Here is my APS test as requested:
-> Heat level 10 APS, elite, sprite: gun damage, chaos set:

D24 - Bash:
http://fs2.directupload.net/images/150615/plklsr48.png
http://fs2.directupload.net/images/150615/435zuunj.png
http://fs2.directupload.net/images/150615/6sp6jiyl.png

D24 - Smash:
http://fs2.directupload.net/images/150615/5c2pnca2.png
http://fs1.directupload.net/images/150615/cx4dcjwr.png
http://fs2.directupload.net/images/150615/prgcaz4h.png

D25 - Bash:
http://fs1.directupload.net/images/150615/kr2xzpfk.png
http://fs1.directupload.net/images/150615/uxps8vr7.png
http://fs2.directupload.net/images/150615/svbmlh2k.png

D25 - Smash:
http://fs1.directupload.net/images/150615/c23gemge.png
http://fs1.directupload.net/images/150615/jcee3are.png
http://fs1.directupload.net/images/150615/s7u6v77t.png

D26 - Bash:
http://fs1.directupload.net/images/150615/44mixbdn.png
http://fs1.directupload.net/images/150615/d52dtoox.png

D26 - Smash:
http://fs2.directupload.net/images/150615/zy8s3bma.png
http://fs2.directupload.net/images/150615/k9yae9ix.png
http://fs2.directupload.net/images/150615/5bzr7vhd.png
http://fs2.directupload.net/images/150615/j7e6337f.png

D27 - Bash:
http://fs2.directupload.net/images/150615/f2bgplqn.png
http://fs2.directupload.net/images/150615/jhp6d5kp.png
http://fs1.directupload.net/images/150615/isuusd96.png

D27 - Smash:
http://fs2.directupload.net/images/150615/n4dh5pq6.png
http://fs1.directupload.net/images/150615/cata6g7c.png
http://fs1.directupload.net/images/150615/w4b5edmo.png

D28 - Vana's Smash:
http://fs1.directupload.net/images/150615/jupog7wo.png

I want to add that in D27 my team mates did got aggro and ran like headless chickens around that trojan. Therefore I believe that he was using almost each time his damage charge before I got damage by his bash.

In that bossfight I planned to messure the bash, too, but I lost interest after a few smashes. Exspecially since it wasn't asked by you.

-Jenovasynthese

lun, 06/15/2015 - 21:26

#12

Hexzyle

Thanks a bunch, Jenova! I'll get right on checking out these values

In that bossfight I planned to messure the bash, too, but I lost interest after a few smashes. Exspecially since it wasn't asked by you.

I already know that Vanaduke breaks the Ancient Plate Shield in one hit, but the damage overflow is exactly zero so only the statuses are applied.

Hmmm, well, I'm fairly sure my screenshot was taken before the Trojan cast any buff, yet my shield took the same damage as yours (Aside from the 0.3% damage reduction from my Normal Low UV)
Wierd: Trojans are doing ~209, 220, then 234 damage to the APS in D24, 25, and 26 respectively. Why the massive spike to 624 in D27?
Definitely needs to be triple checked.

Also, screenshots where the shield health bar is "shaking" from being hit, when it's larger than normal, are completely useless for data collection as I'm using the pixel size of the shield health bar to find its remaining health. I don't care how much of the purple bar is left, I just need images of the blue bar before it begins to recharge

mar, 06/16/2015 - 09:25

#13

Qwez

EDIT:Maybe because the white Trojan is a more powerful variant of the Trojan.
I did not intend for this to be a racist comment, but it sure did look like one :|
I was thinking in terms of elite monsters and they happen to be white and black...

I don't feel like playing right now, so I can't help you with testing Heavy Plate Shield.

mar, 06/16/2015 - 00:10

#14

Snarby-Slasher

@13

Lol dat racist comment

lun, 06/29/2015 - 14:22

#15

Jenovasforumchar

closed?

Will you continue with this project, hex or can it be counted as graveyarded?

lun, 07/13/2015 - 00:07

#16

Hexzyle

At the time of post #8, my GPU had kicked the bucket, and I am running my computer in safe mode with networking on my graphic card's other DVI port that isn't fried.

So I am unable to play Spiral Knights until my new card arrives in the mail, which should be very soon. I will resume this project ASAP.

EDIT: BUMP
Graphics card arrived, still looking for people to help me fill in missing data. Level of the shield doesn't matter, we can take data for multiple levels if needed, it'll help get an idea on how levels affect shields.

mar, 07/14/2015 - 03:54

#17

Jenovasforumchar

shield levels

when I will make a GT, I will take my time with heating it, so I cann add data to this. Ofc there is no way to undo a heat Level, so there will be much to messure which takes time. I also may add that GT is not at my #1 anymore since someone said it had same defense and health like APS (which still no one clearly tested or noted somewhere where i searched for it) - sooooo it will take a Long time until i get my fingers on that.

Ah just noted this: since ASP shields are pretty much for free, it is posible to test dmg with my heated 10 asp and with a newly bought asp as comparision.

dim, 07/19/2015 - 04:39

#18

Jenovasforumchar

tortoshield (yellow)

just wanted to drop that i got my yellow 4* tortoshield at heat 10 and ready to craft it to 5 stars. but before that i wanted to ask you if you need any tests with it?

Edit: your picture link in the opening doesnt work anymore.

Edit Edit: since i got no answer, this tortoshield is now 5* heat 01. i will keep it there to be able to make further testings since i am myselfe interested in the evolution. but where should a pierce shield best be used with these statuses?

dim, 07/19/2015 - 22:04

#19

Flowchart

only took a month but here is Grand Tortoise, level 10, no uv:
http://i.imgur.com/OXRjp2B.jpg

lun, 07/20/2015 - 14:04

#20

Jenovasforumchar

I recalculated life and confirmed it was indeed 42!

42. Definitely 42!

But before we come to the interesting part, just some other screens from a very time intense Vana-Run with my fully heated APS with no UVs while selecting the *shaking* screen shots out:
(Shaking is the effect when the shield energy bar exceeds the starting position to the left and thus will be wrongly displayed)

D24 - Bash:
http://fs1.directupload.net/images/150720/9jz4d9d2.png

D24 - Smash:
http://fs2.directupload.net/images/150720/q6m79r4e.png
http://fs1.directupload.net/images/150720/rl9ddi4k.png

D25 - Bash:
http://fs2.directupload.net/images/150720/imxj29y9.png
http://fs1.directupload.net/images/150720/syzxrb2p.png
http://fs2.directupload.net/images/150720/b3qik3r7.png

D25 - Smash:
http://fs2.directupload.net/images/150720/8wbjfqdu.png
http://fs1.directupload.net/images/150720/mqvh5b68.png

D26 - Bash:
http://fs1.directupload.net/images/150720/vjze8wat.png
http://fs1.directupload.net/images/150720/l5peldk6.png

D26 - Smash:
http://fs1.directupload.net/images/150720/w4uwvf6t.png
http://fs1.directupload.net/images/150720/iaxuorwd.png
http://fs2.directupload.net/images/150720/6mce2a6z.png

The answer to life, the universe and anything else:

D27 - Black Trojan's Bash:
http://fs2.directupload.net/images/150720/rx42ii32.png
http://fs2.directupload.net/images/150720/skvenu3a.png

D27 - Black Trojan's Smash:
http://fs2.directupload.net/images/150720/g79zzvxn.png
http://fs1.directupload.net/images/150720/w5g88h2v.png

D27 - White Trojan's Bash:
http://fs1.directupload.net/images/150720/ckgnilp7.png
http://fs1.directupload.net/images/150720/jqclqqlx.png
http://fs2.directupload.net/images/150720/rjpdmwon.png
http://fs1.directupload.net/images/150720/hnlg6ect.png
http://fs1.directupload.net/images/150720/y6eykzgj.png
http://fs2.directupload.net/images/150720/5sp4xxmg.png

D27 - White Trojan's Smash:
http://fs2.directupload.net/images/150720/wbopp3ml.png
http://fs2.directupload.net/images/150720/9yzpq5ji.png
http://fs1.directupload.net/images/150720/6c299tbw.png
http://fs2.directupload.net/images/150720/8ojanp6s.png
http://fs1.directupload.net/images/150720/fk9hsgst.png

So White Trojans are T4 Trojans? :3

One thought that is strengthening by adding fallowing element to the white trojan's hit: his charge attack forehand:

D27 - White Trojan's charge strengthened Bash:
http://fs2.directupload.net/images/150720/437hxdhg.png

D27 - White Trojan's charge strengthened Smash:
http://fs2.directupload.net/images/150720/lgqkvr2a.png

Note that I was using a fully heated APS with no UVs additionally to my chaos set with UV med fire res on each part and THAT I HAD FULL HEALTH before he attacked. My Sprite had no relevant perk epquipped (gun damage) and I was not using trinkets.
APS has the *perfect* status value vs this attack and STILL it broke and it even damaged my health.

Therefore I guess that this trojan data belongs to a tier below the current one that we actually have now as maximum. Similar to an arena inside the arcarde where there can deeper-tiered monsters appear.

Even if this might not be accurtae, then we have at least solved the mystery of that previous *common damage spike between all trojans* in D27.

jeu, 07/23/2015 - 08:12

#21

Hexzyle

Oh? I never noticed a

Oh? I never noticed a graphical difference with the Trojans. It's good to know this particular Trojan is using custom values rather than there being something wonky with how damage is calculated at this exact amount (what I was suspecting it was)

Thanks for all the data Jenova, I'll have time to process it tomorrow.

I don't have plans currently to extend testing to other tortodrone shields, as they have very little in common with plate shields and mostly share stats with their sidegrades. I've mistakenly put Heater Shield up there too.

jeu, 07/23/2015 - 09:13

#22

Parasthesia

So basically the data is

So basically the data is showing ancient plate as superior so far?

ven, 07/24/2015 - 08:50

#23

Hexzyle

@Parasthesia
So basically the data is showing ancient plate as superior so far?

With the small amount of data that I've obtained on the Grand Tortoise, that is what it's leaning towards.
If I can get some data on a Volcanic or Ironmight Plate Shield, and that data turns out to be identical to the Grand Tortoise, then yes, it will be confirmed.

@Jenovasforumchar
Your screenshots of the D27 "black" Trojan's overhead swings line up correctly with the rest of the data, this is great news.
It seems that only the overhead swing is buffed on the white Trojan? The Shield Bash seems to deal the same amount of damage between the two different types of Trojans. So this means either White Trojan overhead swings have a far higher ratio of Shadow vs Normal damage, or Trojan Shield Bashes deal untyped damage that isn't resisted by any shield (I'm pretty sure they don't though)
Either could be probably tested with a Gorgomega or a Crest of Almire or something

You know what the interesting thing is? These "White" Trojans share the appearance of Tier 1 Trojans.

sam, 07/25/2015 - 11:03

#24

Jenovasforumchar

@Flowchart

You made one screen yet. Please make multiple screens of the same trojan. in best case with the same resolution I did for better comparision (does higher solution changes things after all? After all your Emergency revive shows another place than mine, that makes it look like APS stands more damage or you simply made the screen in the wrong moment)

What you need to take care of: make the screen just in the moment when the purple damaged part of the shield starts to disappear. If you make your screen in the *damaging* moment, then the shield health moves around fastly and thus a screen in this moment will be corrupted. You can see many corrupted screens in my first testing series in post #11. In my second one in post #20 i filtered out all the corrupted ones and delivered bare data of bash and smash.

Question:
,,I did make a screenshot, how can i see if it is qualified or corrupted?"
A:
A corrupted screen is easily discovered when the first part of your shield's health exceed more to the left than it is doing in an idle condition. This will not be the case anymore after the purple shield's health started to decrease and exposing some missing shield's health.

Q: ,,So are screens that show missing purple health always qualified?''
A: As long as they had full health before the damage to be tested hit AND the screenshot shows the shield rised AND it did not start to regenerate: then yes.

TLDR (whatever these 4 letters mean): Flow, please make more qualified screens.

@Hex:
yes, I knew that these Trojan's skins were T1 skins and thus did not match their tier. Same goes for flame puppies. But with trojans I did not expect that they are actually different in strength, too.

sam, 07/25/2015 - 20:01

#25

Flowchart

I find your demanding tone hilarious. But I still have the source video, so I'll put up some images and you can pick the one you want.
(not censoring chat this time, not like people don't know who my alt is from wiki)
Shield smash
Shield bash

lun, 07/27/2015 - 09:43

#26

Hexzyle

Thanks Flowchart. I'll run my level 5 Grand Tortoise past the white Trojan later to see if it makes a difference.

dim, 08/02/2015 - 11:39

#27

Jenovasforumchar

CoA vs Trojans

Fresh new data - just picked in the morning and already eatable:

Crest of Almire, heat10, no UVs

D24 - Bash:
http://fs2.directupload.net/images/150802/gpijgytf.png
http://fs1.directupload.net/images/150802/du9zwixr.png

D24 - Smash:
http://fs2.directupload.net/images/150802/zdhzv7t2.png

D25 - Bash:
http://fs1.directupload.net/images/150802/rnler25y.png
http://fs1.directupload.net/images/150802/3ebhrjox.png
http://fs2.directupload.net/images/150802/dvotbwll.png

D25 - Smash:
http://fs2.directupload.net/images/150802/fulozdtz.png
http://fs2.directupload.net/images/150802/bpyznnio.png
http://fs1.directupload.net/images/150802/fjutysf7.png

D26 - Bash:
http://fs1.directupload.net/images/150802/hu4aglho.png
http://fs2.directupload.net/images/150802/2ikcgel9.png

D26 - Smash:
http://fs1.directupload.net/images/150802/kaxyaee4.png
http://fs2.directupload.net/images/150802/w9rvh434.png

And occassionally I got this while doing the tests:
D26 - Smash with an active defense orb (not that it would make any difference, but since I got it I am mentioning it):
http://fs2.directupload.net/images/150802/ny5fvyg2.png

D27 - Black Trojan's Bash:
http://fs1.directupload.net/images/150802/8zhfgwhd.png
http://fs2.directupload.net/images/150802/oxnj2vcv.png

D27 - Black Trojan's Smash:
http://fs2.directupload.net/images/150802/7joroogu.png
http://fs2.directupload.net/images/150802/4mge6uav.png

D27 - White Trojan's Bash:
http://fs1.directupload.net/images/150802/5gp3chdc.png
http://fs2.directupload.net/images/150802/7rln5jvy.png
http://fs2.directupload.net/images/150802/2hf6fh9n.png

D27 - White Trojan's Smash:
http://fs1.directupload.net/images/150802/vw8n4cdd.png
http://fs1.directupload.net/images/150802/ykiwbr7l.png
http://fs1.directupload.net/images/150802/b3h33aq8.png
http://fs2.directupload.net/images/150802/j5nxqj4d.png

jeu, 08/06/2015 - 05:27

#28

Hexzyle

Thanks Jenova, this indeed confirms that the White Trojan is doing a larger portion of shadow damage on its slam attack: the CoA defends against it better than Black Trojans, while the Plate Shields defend better against the Black Trojan than the White one.

sam, 08/08/2015 - 09:09

#29

Jenovasforumchar

Omegashell

It would be really interesting to see Omegashell added to the analysis.
http://wiki.spiralknights.com/Omega_Shell

I don't know if you think the same, but meanwhile I see in this topic more than just the question whether GT has or hasn't the same defense, but morely a whole analysis topic of shields - though if currently only vs FSC Trojans.

Let us get the numbers how this game is programmed hrhrhr!

lun, 08/10/2015 - 21:33

#30

Hexzyle

Alright, I changed the title of the thread.
Sorry that I'm rather slow with the adding of the data. Omega Shell and Ironmight Plate Shield are two shields I really want, but I'm waiting to obtain a decent UV on their predecessors before I craft them.

lun, 08/24/2015 - 04:36

#31

Holy-Nightmare

Question on using pixels

If you are using pixels as a measurement are you assuming that everyone is using the same screen resolution?

lun, 08/24/2015 - 12:12

#32

Jenovasforumchar

Answer on using pixels

It doesn't matter.

If you ever change the resolution of that game then you will see that the interface stays the same.

Just have a look on Flowchart's screen (who is using a higher resolution) and mine, then you will see.

By the way: I know that you are the tanky-type of player. Do you have any shields in stock that didn't get tested yet above? I bet you have some shields with 5 stars that are just waiting to get tested, it can be any shield as long as it has normal defense and as long as it stands at least one trojan's hit.

Contribution is always welcomed.

lun, 08/24/2015 - 12:20

#33

Holy-Nightmare

/shrug

I've got a VPS and IPS lying around

As a side note: Do you think that the GT should have the same defenses as the APS or not.

lun, 08/24/2015 - 14:00

#34

Jenovasforumchar

testing would be great!

It would be really awesome if you can contribute with your two shields by doing around two FSC runs with these shields epquipped and testing it against said trojan's hits.

To your question... uh I *thought* that both had the same defense, but in tradeoff for the shieldbash, the GT has its movement malus. Well it was proven different though.

lun, 08/24/2015 - 20:14

#35

Hexzyle

If you are using pixels as a measurement are you assuming that everyone is using the same screen resolution?

The GUI is always 360 pixels long for the shield bar.
This is why you get complaints of the GUI being way too large at lower resolutions.

As a side note: Do you think that the GT should have the same defenses as the APS or not.

Absolutely, and I will be very frustrated if it turns out it has identical defenses to the VPS/IPS
If it has some wacky defenses that seems like a mistyped number, that'd be okay because it'll likely get fixed because of it being bugged. But if something is intentionally crap/unbalanced, it'll likely stay that way for a long time >:(Iron Slug, Acheron, Chaos, Cautery Sword, Blitz, Winmillion, Neutralizer, Royal Jelly Mail, Vitasuit Deluxe, Cobalt)

mer, 08/26/2015 - 10:25

#36

Fehzor

Thank you guys for doing this by the way, I'm curious and would offer to help but have not fully heated my normal tortodrone shield + have unique variants on the majority of my shields that would bung up the numbers.

As a side note- I think it should be identical to omega shell in all stats sans shield bash/msi. This would make it extremely well balanced as it would be the best general use option for tortodrone shields.

sam, 08/29/2015 - 08:33

#37

Skepticraven

↓

Be careful if you include the stone tortoise/omega shell shields.
They have 25% more hp than the other shields.

Edit: direct link to raw data

dim, 08/30/2015 - 10:03

#38

Jenovasforumchar

@Skeptic

Isn't the displayed shield bar larger compared to other 5* shield bars?

So if it has 25% more health, it would mean that Omega got also 25% more pixels.

I've never wielded an omega shell, therefore I wonder why else you warn about this (since it is a common knowledge that Omega Shell got more).

dim, 08/30/2015 - 11:47

#39

Skepticraven

↓

No. It is possible it is a long standing overlooked bug.
Pixel count is determined by star count, not HP. These are the only two shields in which that distinction would actually make a difference...
If you swap from omega to another 5* shield, it will (incorrectly) show the "damage" wrapped around to the left hand side as silver hp would (if you were to unequip a health trinket). The damage resisted, however, begins on the far right as expected. It is a goofy phenomena.

I did a quick login for confirmation (in GH training).

mer, 09/02/2015 - 16:47

#40

Jenovasforumchar

Barbarous Thorn Shield vs Trojans

Before we start: I wonna suggest that we add the data from the trojan's bash, too. I just tested today my DSS, heat 6 (which has not any normal def at all) and it could easily take the bash. The smash was (as expected) devastating.

Now here are the BTS' results

BTS, heat 10:

D24 - Bash:
http://fs2.directupload.net/images/150902/lzu4pxi5.png
http://fs2.directupload.net/images/150902/t586p77n.png

D24 - Smash:
http://fs2.directupload.net/images/150902/ug8g4zbt.png
http://fs1.directupload.net/images/150902/hrauszyk.png

D25 - Bash:
http://fs2.directupload.net/images/150902/4ozmuify.png
http://fs1.directupload.net/images/150902/m4ozhz4h.png

D25 - Smash:
http://fs1.directupload.net/images/150902/f3aozame.png
http://fs1.directupload.net/images/150902/id2cijky.png

D26 - Bash:
http://fs2.directupload.net/images/150902/vvgd355d.png
http://fs1.directupload.net/images/150902/xtesty9e.png

D26 - Smash:
http://fs1.directupload.net/images/150902/y4ozstfc.png
http://fs1.directupload.net/images/150902/nvttw5gu.png

D27 - Black Trojan's Bash:
http://fs1.directupload.net/images/150902/tncvgcr2.png
http://fs1.directupload.net/images/150902/n3m939if.png

D27 - Black Trojan's Smash:
http://fs1.directupload.net/images/150902/nfz5z7ea.png
http://fs1.directupload.net/images/150902/3oqn3vjp.png

D27 - White Trojan's Bash:
http://fs1.directupload.net/images/150902/e4ful7ji.png
http://fs1.directupload.net/images/150902/rqgtxxco.png

D27 - White Trojan's Smash:
http://fs2.directupload.net/images/150902/o2gsi9tp.png
http://fs1.directupload.net/images/150902/ucxm2r8l.png

Note that the White Trojan's smash in D27 is able to destroy BTS with one attack, but he does not deal additional damage to my health in that process. Stun, however, can still get applied.

By comparing this data to the CoA's results, we should be able to determine the exact amount of shadow damage from trojans in pixels, can't we? The only thing that would remain in that thesis is actually how much the bare shield *health* reduces. Because of this once again: I recommend to add the weaker trojan's bash to allow the DSS to join the test.

mer, 09/02/2015 - 16:46

#41

Jenovasforumchar

Trivia

To determine the part of shadow damage, it would be necessary to confirm that CoA and BTS had the same amount of normal defense. Therefore I did a quick test on a void Lumber in D25 to confirm that the wiki bars are the very same:

BTS:
http://fs2.directupload.net/images/150902/wwlj87v2.png
http://fs2.directupload.net/images/150902/xs6rsuov.png

CoA:
http://fs1.directupload.net/images/150902/yyjbegfd.png
http://fs2.directupload.net/images/150902/h25qsadi.png

Result: Positive - bars are the same.

Edit:
By the way: I am currently concentrating my heating completely on Dragon Scale Shield - it reached heat level 8 meanwhile and I got the rads to heat it to 9 as soon as the heat bar is filled. So it wont take long till I got an exclusive shield which is so exclusive that it is total crap if the enemy dealsnormal and/or shadow damage.

So all damage caused by trojans will be only held by the bare DSS' health and nothing else. In other words... we can kinda say that this would be the 'base' damage.
http://wiki.spiralknights.com/Dragon_Scale_Shield

Well the other shield of question would be the Celestial Shield for things like Lumbers...
http://wiki.spiralknights.com/Celestial_Shield
Is anyone reading this who obtained a Celestial Shield?

@Fehzor:
Actually it would be good to see data with UVs, so we can see how different UVs are affecting the shields, too.

mer, 09/02/2015 - 19:52

#42

Skepticraven

↓

@Jenovasforumchar

Shield HP doesn't change with heating, only resistances do. If you're using the shield as "armorless", then lvl 1 heat will be exactly the same as lvl 10 heat.

jeu, 09/03/2015 - 12:11

#43

Jenovasforumchar

@Skeptic

yea... I was asking the same, too today while travelling home. Thanks for the info, well DSS heat level 8 will do it. I may upgrade it in future maybe to have a shield that can test e.g. elemental or piercing when there is a split normal/piercing or normal/elemental damage to test.

side note:
I did read about your method to get the shield health. The big list in the end is pretty chaotic or simply too stuffed to read something out as laie. But I snapped one thing that bothers me:
You mentioned 'known damage numbers'. Do you already know some exact numbers of damage from any monsters? If yes, can you link a place where the results are (if existing)?

jeu, 09/03/2015 - 15:19

#44

Skepticraven

↓

The first couple sheets of the massive data spreadsheet were meant to be summaries.
These are the ones I took the time to narrow down. They should be all linked to (viewable by clicking on the cell) the sheet and location of where the number was obtained.
All of them are single damage type values (no split damage).

Unfortunately, they're not useful for for calculating armor values (with reduction), because they are after rounding. (Example: D19 shankle deals 262.00000001 to 263.0000000000 damage.)
I haven't figured out an excellent method of minimizing monster damage below that 1 damage point range without figuring out the depth multipliers (for both monster damage and armor) first. If someone wants to continue towards this work, I have a good spread of data on a variety of weapons. I don't have the free time to give towards this stuff anymore.

ven, 09/04/2015 - 02:05

#45

Hexzyle

(Example: D19 shankle deals 262.00000001 to 263.0000000000 damage.)

Um, a better way to word it is:
262 < x <= 263

ven, 09/04/2015 - 12:57

#46

Jenovasforumchar

Lockdown

can't lockdown be used to determine the exact damage number? Like hitting an attack once - well then there might be some rounding in decimals. Well, but hitting with the same attack type 10 times gives the resulting number already an accuracy on 1 digit beyond the comma. If possible, 100 hits will rise the accuracy up to 2 digits beyond the comma.

Of course you need to kinda make sure that no body else would interrupt. So I guess it can be donly at certain times when there is like no LD at all, so you can start a run only with people who know about the testing.

I recall that such kind of testing has been done already. So this would actually solve the rounding error, if you simply take a certain damage more than once.... wouldn't it?

sam, 09/05/2015 - 04:44

#47

Krakob

It wouldn't. Damage is rounded when it's applied, not on the results screen. For example, if you deal 60.2 damage 100 times it won't be ceiling(60.2 * 100), it'd be ceiling(60.2) * 100

sam, 09/05/2015 - 05:34

#48

Skepticraven

↓

You may be able to minimize a variety of weapons (say in the arcade using my current methods), find a couple with very specific known damage values and then bring those into LD to test armor (varying with armor boosts)... It's a long process. I danced around using this method to find D23 neutral monster armor using only a proto sword here while trying to prove Gear-Storm's equations. I never got around to running the numbers with supernova (since it would be very close to the link between the linear/non-linear range of the equations). There was a method to my madness - trust me.

Zeddy has done the tests you have described already by just reading the damage number from the attacker, but your maximum resolution is +/- 1 unless you draw the conclusion that weapons have relational multipliers (this thread).

Anyways... Shields.
We are pretty much limited to monster damage for testing them (unless we can get a mod to test with us using their all-powerful crowbar).
In order to reduce monster damage below the +/- 1 range, we need pretty exact armor values and trustworthy damage equations.
The current methods above (in the OP) are best and will be useful in the future once these criteria are met.
We can already pull relational characteristics from this, but may miss the very fine details.

@Hexzyle

I'm aware that's a better wording, but from past experiences people (whom don't deal with mathematics often) tend to misunderstand that format. Otherwise I'd even go as far as using (262,263].

sam, 09/05/2015 - 12:29

#49

Jenovasforumchar

Trojan vs Dragon Scale Shield FINALLLYYY (god this took me long)

Woa, that was probably my most time intensive run I ever have done through FSC!

For this run I was using:
-a non-relevant sprite perk (gun damage)
-Dragon Scale Shield (DSS), heat level 8
-Wyvern Scale Mail (4*), heat 10
-Dragon Scale Helm (5*),Heat 10

So I have totally no normal or shadow defense but some very relevant fire def for zombies and other sources in FSC. But let us look at the trojans now.

The idea behind using neither shadow nor normal resistance on amor was to get that *pure* damage in case that the shield can not bear the trajan's attack. Well, surprisingly, only the smash of the white trojan was able to deal any damage at all if the shield had full health.

And while testing I encountered one very strange anormality: Trojans seem to have at least two different smashes. The weaker one can not even break the DSS with one hit. (Later about this after the screens.)

Here the data:

D24 - Bash:
http://fs2.directupload.net/images/150905/cuer8ix8.png
http://fs2.directupload.net/images/150905/xir58ia3.png
http://fs1.directupload.net/images/150905/utkhopkq.png
http://fs2.directupload.net/images/150905/bbykznjy.png
http://fs2.directupload.net/images/150905/nkheewxi.png
http://fs1.directupload.net/images/150905/fbm6juhx.png
http://fs1.directupload.net/images/150905/h7dadk77.png

D24 - Smash, shieldbreaking but zero damage to Knight's health, but stun inflicted:
http://fs1.directupload.net/images/150905/ih6vvqy3.png
http://fs2.directupload.net/images/150905/6d7w5mc4.png
http://fs1.directupload.net/images/150905/3goruu6x.png
http://fs1.directupload.net/images/150905/pb8niwhl.png
http://fs1.directupload.net/images/150905/zrqkjntx.png
http://fs2.directupload.net/images/150905/rknra2ka.png
http://fs2.directupload.net/images/150905/9wtf9669.png

D24 - Smash, shieldbreaking but zero damage to Knight's health, no stun inflicted:
http://fs2.directupload.net/images/150905/j2wv3ko2.png
http://fs2.directupload.net/images/150905/ccyk97sh.png
http://fs1.directupload.net/images/150905/y4u5ozv8.png

D24 - Smash, anormally weak smash that can not break the shield:
http://fs2.directupload.net/images/150905/vabaak24.png
http://fs1.directupload.net/images/150905/dlaohsaw.png
(The next one did probably hit the fire trap, not sure though, but keep in mind that this screen has to be seen with care)
http://fs1.directupload.net/images/150905/qoqloy54.png

D25 - Bash:
http://fs2.directupload.net/images/150905/d9r3o55d.png
http://fs1.directupload.net/images/150905/j7nyg2p3.png
http://fs1.directupload.net/images/150905/o9fjkycn.png
http://fs2.directupload.net/images/150905/bmnrfdcd.png
http://fs1.directupload.net/images/150905/kqn5j6nw.png
http://fs2.directupload.net/images/150905/xoryrnpp.png
http://fs1.directupload.net/images/150905/p8w4osoz.png

D25 - Smash, shieldbreaking but zero damage to Knight's health, but stun inflicted:
http://fs1.directupload.net/images/150905/867cyvzs.png
http://fs2.directupload.net/images/150905/yuc68nyi.png
http://fs1.directupload.net/images/150905/nsg32fr6.png
http://fs2.directupload.net/images/150905/9bwduab5.png
http://fs1.directupload.net/images/150905/d93ueaq8.png

D25 - Smash, shieldbreaking but zero damage to Knight's health, no stun inflicted:
http://fs2.directupload.net/images/150905/yxh3glwu.png

D25 - Smash, anormally weak smash that can not break the shield:
http://fs1.directupload.net/images/150905/5pow34co.png
http://fs2.directupload.net/images/150905/9fbq8nlh.png

D26 - Bash:
http://fs1.directupload.net/images/150905/5o58ek7n.png
http://fs1.directupload.net/images/150905/y4xbv72x.png
http://fs1.directupload.net/images/150905/g4y32zbu.png

D26 - Smash, shieldbreaking but zero damage to Knight's health, but stun inflicted:
http://fs1.directupload.net/images/150905/fr5msx8d.png
http://fs2.directupload.net/images/150905/t7kxgdlq.png

D26 - Smash, shieldbreaking but zero damage to Knight's health, no stun inflicted:
http://fs2.directupload.net/images/150905/s7hkhswn.png
http://fs1.directupload.net/images/150905/cwx2285o.png

D26 - Smash, anormally weak smash that can not break the shield:
http://fs2.directupload.net/images/150905/oqemwp46.png
http://fs1.directupload.net/images/150905/b79jnf3g.png
http://fs1.directupload.net/images/150905/3jzdyo52.png

Black Trojan:
D27 - Bash:
http://fs2.directupload.net/images/150905/hczilln7.png
http://fs1.directupload.net/images/150905/reopcr7c.png
http://fs2.directupload.net/images/150905/f5245l2m.png

D27 - Smash, shieldbreaking but zero damage to Knight's health, but stun inflicted:
http://fs2.directupload.net/images/150905/m7gqm24m.png
http://fs2.directupload.net/images/150905/h2ppczk5.png
http://fs1.directupload.net/images/150905/otk2f7dg.png

D27 - Smash, shieldbreaking but zero damage to Knight's health, no stun inflicted:
http://fs2.directupload.net/images/150905/ae5z8kas.png
http://fs1.directupload.net/images/150905/ba68ul2l.png

D27 - Smash, anormally weak smash that can not break the shield:
http://fs2.directupload.net/images/150905/iawt3h2i.png
http://fs2.directupload.net/images/150905/onj4wjbk.png

White Trojan:
D27 - Bash:
http://fs1.directupload.net/images/150905/8pr7rsgc.png
http://fs1.directupload.net/images/150905/grn37nga.png

D27 - Smash, shieldbreaking with additionally about 2% damage to Knight's health, no stun inflicted:
http://fs1.directupload.net/images/150905/wtlelhqz.png
http://fs2.directupload.net/images/150905/6u892wgg.png
http://fs1.directupload.net/images/150905/jczvx227.png

D27 - Smash, anormally weak smash that can not break the shield:
(no data q.q)

A word to the amount of the smashes of each category:
-The normal case has always been the smash that completely breaks the shield but not harming the Knight's health at all plus dealing stun.
-Around 2 of 10 smashes (without messuring these numbers, only a guess!!!) the smash is only destroying the shield. No damage to Knight's health and no stun caused.
-And finally the anormally. Sometimes it happened very often in row and near other trojans I had to wait for 15 minutes (yes, I was fighting one single trojan in d26 for 15 mins) to get a second weak attack to confirm the number. I have the assumption that the weak attack is somehow connected to the bash, but when I tried to reproduce it, I failed. My thesis is that the smash a trojan is usually dealing after he bashed to the knight is somehow weaker and that you somehow can trigger this effect even without the need of dash. well maybe it is the next-hit after a buffed smash? However, this still needs to get figured out.

My first though was that the 'weak' smash would be the one without stun and this the shield needs to bear less, but actually there is a stunless hit, that breaks the shield, too!

A word about what I expected and what I encountered:
When I started all this, I ofc did not expect to encounter two different amounts of smash damage - obviously. I also was surprised that that no-normal-no-shadow-damage-DSS can survive EACH SINGLE ATTACK any of the trojan is doing (except for the white smash one). If I had known this before, then I would have brough my good ol' chaos instead of the dragon scale stuff. After all I am really surprised that the difference between a shield with normal damage like BTS and DSS is so minor. And once again I am asking myselfe: if the actual normal defense bar matters so little, then how awesome must Omega Shell be? Please, someone.... bring Omega into the tests <3

lun, 09/07/2015 - 23:17

#50

Hexzyle

@Jenovaforumchar
I commend you for thinking this through enough to also wear Dragon Scale so we can also measure the overflow damage.

Sorry I haven't been compiling your recent data, I've been waiting until I get someone's or my Ironmight data (when I get it) because that's what I've been looking forwards to.

@Skepticraven
I'm aware that's a better wording, but from past experiences people (whom don't deal with mathematics often) tend to misunderstand that format. Otherwise I'd even go as far as using (262,263].

Except you said 262.00000001, specifically saying that it's greater than 262 but an unspecified amount greater, so saying "262" is incorrect, no?

Recherche

Analysis of Shield Defenses (and determining where the Grand Tortoise stands)