Shield heating and shield UVs: worth it?
jeu, 03/19/2020 - 12:39
Previous shield research was lacking so I gathered some new data. Damage data was gathered by measuring the shield HP bar in pixels on a screenshot after the hit. I chose to test against 2 attacks from both sides of the damage spectrum, lumber swing being one of the heaviest hits and scuttlebot bullets being one of the weakest hits. I assume most other attacks will fall between this range.
First set of data: Damage to shield taken at depth 19 from a Lumber (in mission 6-2 Beyond the Axes of Evil) on Elite.
(A Lumber swing deals elemental, normal and stun damage to shields).
A Grey Owlite Shield without any UVs at heat level 1 took 95.833% damage to the shield HP from a lumber swing
A Grey Owlite Shield with a Stun MAX UV at heat level 1 took 83.888% damage to the shield HP from a lumber swing
A Grey Owlite Shield with a Stun MAX UV at heat level 10 took 78.888% damage to the shield HP from a lumber swing
A Grey Owlite Shield with Stun MAX + Elemental MAX + Normal medium UVs at heat level 10 took 73.055% damage to the shield HP from a lumber swing
Second set of data: Damage to shield taken at depth 19 from a Scuttlebot (in mission 6-2 Beyond the Axes of Evil) on Elite.
(A Scuttlebot bullet deals pure elemental damage).
A Grey Owlite Shield without any UVs at heat level 1 took 23.055% damage to the shield HP from a scuttlebot bullet
A Grey Owlite Shield with a Stun MAX UV at heat level 1 took 23.055% damage to the shield HP from a scuttlebot bullet (no change since the attack doesn't inflict Stun)
A Grey Owlite Shield with a Stun MAX UV at heat level 10 took 20.833% damage to the shield HP from a scuttlebot bullet
A Grey Owlite Shield with Stun MAX + Elemental MAX + Normal medium UVs at heat level 10 took 16.111% damage to the shield HP from a scuttlebot bullet
Discussion:
1. Stun max added a good bit of extra protection against the lumber swing. The difference in damage taken was 12.5% (83.888/95.833).
2. Heating the Grey Owlite shield from 1 to 10 adds a nice bit of damage reduction. Against the lumber it took 6% less damage (78.888/83.888), against the scuttlebot it took 10% less damage (20.833/23.055).
3. The elemental max + normal med UVs reduced damage from the lumber swing by 7.5% (73.055/78.888). The elemental max reduced damage from the scuttlebot bullet by 22.5% (16.111/20.833)
4. The combined damage reduction from fully heating the shield and adding Stun MAX + Elemental Max + Normal medium UVs is 24% (73.055/95.833) against a heavy lumber swing and 30.2% (16.111/23.055) against a scuttlebot. Reminder that the reduction against the scuttlebot is exclusively caused by the elemental max UV and heating.
Since damage and defense interact by subtraction, these findings make sense. Defense increases will have a larger effect on the net damage coming from weaker attacks. The most spectacular result comes from the Elemental max UV against the weak scuttlebot, 22.5% damage reduction. In comparison: elemental max and normal med combined give a much smaller damage reduction versus a strong lumber swing. Heating provides a more balanced bonus to the shields defense, since it adds both normal and elemental defense. The result is still a 10% damage reduction against the pure elemental damage scuttlebot. Against the lumber it provides a 6% damage reduction. The Stun UV acts as defense on shields, but since only strong attacks carry stun, there are no small attacks to compare to. The reduction that Stun Max achieves against a lumber swing is a respectable 12.5%.
Conclusion:
Shields should be considered a regenerating health pool that will tank lots of hits over the course of a mission. Shield defense can be improved by heating and UVs. Combined these can provide over 30% damage reduction depending on monster attack. Status UVs can provide a large bonus but not all attacks inflict status. Both UV types (matching defense UVs and status UVs) are very useful to increase your shield's defense. Heating adds a balanced bonus to both shield defense types and seems comparable to 2 medium defense UVs.
Some players might be so good at dodging that they never tank a hit with their shield more than once in a row. In that case the shield would never break and so additional defense does not help them. However, it is plausible that in Shadow Lairs, 1 hit will break an unheated shield without UVs more frequently because monster attack is so high. In that case, UVs and heat might mean the difference between tanking a hit or having the shield break and status inflicted on you.
tl;dr: heat your shields and get UVs if you can afford it, it's definitely not a negligible bonus. Obviously though, heating and UVing weapons and armor takes priority over shields.
sam, 03/21/2020 - 12:45
#2
Response 1
Yes I have seen Glacies shield research and used the same methods to gather data. His findings revealed a lot of the shield mechanics which has been very useful. However I disagree with some of the conclusions he drew from that data. I meant the research on shields was lacking in that there doesn't seem to have been anything new after his research and it has been 7 years. I wanted to provide some nuancing to his conclusions.
https://forums.spiralknights.com/en/node/90509
https://forums.spiralknights.com/en/node/89111
The wording in these threads regarding the effectiveness of heat/UV is confusing. E.g. "Each Tick on a Shield's bar is equal to about 2% resistance to that damage-type", "As for UVs, each UV rank is adds about 1% to the absorbed damage", "The fact that heating only provide a 2% difference is absolutely horrible". "Non Status UVs are pointless on Elite. Each Tick gives you +1% resistance".
This wording makes it sound as if the % are relative to the total damage absorbed, as in you take 100% dmg at heat level 1 and 98% at lvl 10, or that a Max Shadow defense UV brings damage taken from 100% to 96%. This is not what Glacies means when compare this to his data (which he doesn't link to in that thread, making it more difficult to clear up the confusion for the average reader). When he says the difference is 2%, he means the difference is 2 percentage points in damage taken with the shield. So if an attack hits for 25% of the shield hp unheated, it would be 23% heated. If the attack hits for 85% unheated, it would be 83% heated. It should be clear that in the former situation the effect is much larger compared to the latter. Since he generalizes in his conclusion by saying "Non Status UVs are pointless on Elite. Each Tick gives you +1% resistance" and "The fact that heating only provide a 2% difference is absolutely horrible" only adds to the confusion that he might not be talking about percentage points and is talking about a relative percentage increase to the defense.
I agree with him that [usually and on average] each UV rank will reduce damage taken by about 1 percentage point and that heating reduces damage by about 2 percentage points for each damage type on the shield. I disagree that this automatically means both are redundant, as I show in my previous post that this entirely depends on the incoming attack. I think he may have confused himself with his own wording, and jumped too conclusions too quickly.
sam, 03/21/2020 - 13:19
#3
Response 2
Regarding Status vs Defense UVs, I just don't think it is very relevant to say that Status UVs are more powerful than Defense UVs.
- Yes, against an attack that inflicts Status (e.g. lumber swing), the Stun Max will have a larger effect than the Elemental Max. However, Elemental Max is more generally useful since there are many more attacks that contain elemental compared to attacks that contain stun.
- For most attack+status combinations there is a shield with an innate status resistance that is more powerful than what a status UV provides. A fire max UV on an elemental shield isn't very useful because if you want to defend against Elemental + Fire you should just use Grey Owlite shield. If you want to defend against Elemental + Freeze you can use a heater shield.
Some notable exceptions are the following:
1. Stun high/max is very useful on a lot of shields, because while each level generally has 1 main status type (fire/shock/freeze/poison), stun can occur as a secondary status from any heavy enemy. In addition, heavy enemies are usually going to be the ones to actually break your shield, so extra defense against them is great.
2. Any status/defense combination that doesn't exist on a stock shield. Examples are Piercing+Freeze shield, to some extent also piercing+shock (because while there is a piercing+shock tortoshield, the MSD is a big tradeoff). Other examples could be freeze, shock or poison on a plate shield because even though there is a poison+piercing shield (dragon scale), you might also really want the high normal defense that plate shield offers. A final example could be curse UVs on any shield that you might want to bring to a curse level.
Conclusion:
The answer to your question is 3-fold:
1. Is there a shield with inherent status resist and similar defenses that outclasses my status UV? (e.g. crest of almire and dread skelly cover all 4 statuses for shadow+normal defense, the only useful status UVs for those are stun and curse, and then there's still teddy bear buckler which has a better stun resist combined with freeze). Only a select few status UVs are irreplaceable by inherent status resists from other shields.
2. Which attacks do we consider the most important to tank? We probably want to cover large attacks since those are the most likely to result in shield breaking (therefore Stun UVs are sometimes really useful), but bullets are probably the most frequently shielded attacks (therefore matching piercing/elemental/shadow UVs are generally very useful)
3. We shouldn't be trying to prove that 1 type of UV is the end all be all, and instead go for a combined effect from all buffs. In some situations a status UV (usually stun) is going to make the largest difference. But a shield that has both (stun+defense UV) is going to be better in all situations. And a shield that has no UVs and is unheated will be the weakest for sure (and my data against the lumber shows it is a pretty large difference)
ven, 04/17/2020 - 22:24
#4
Amazing Post
This shield defense data was FAR more easy for me to understand than all the previous shield data I've seen. The way you presented it just makes sense. From the data I read previously, I was under the impression that shield UV's and heating were worthless. Like total they would only reduce damage by like 4% (from 100% to 96%). Even with all UV maxes. I must have misinterpreted previous data. Thanks a million!
I've theory crafted with the idea of specialized Omega shells to determine the effect of UV's on a shield. I want to take a Volcanic Plate Shield with no UV's vs an Omega Shell with Fire Max, Stun Max, Elemental Max and see which takes more damage from Vanadukes mace. Before I thought there was no point in testing because shield UV's were insignificant, but now I think it might still be worth testing.
dim, 04/19/2020 - 17:54
#5
Simple question: will shock
Simple question: will shock resistance decrease the damage shock itself does to your shield, or just the damage from shock-related attacks?
dim, 04/19/2020 - 18:31
#6
It's the damage shock does to your shield. 5* shields with shock resistance take no shock damage at all.
The shock damage only occurs when the status would've inflicted, which is why a non-resistant shield can take different amounts of damage from the same attack.
A good way to test status damage would be to block zombie breath or wisps, both sources that deal no damage outside their statuses.
dim, 04/19/2020 - 21:31
#7
So then no, the the damage
So then no, the the damage from shock spasms themselves would not be altered.
dim, 04/19/2020 - 22:16
#8
Oh, I misunderstood!
You're correct. Shock spasms do a fixed amount of elemental damage, and as such only the elemental defense of your shield would matter in blocking it.
lun, 04/20/2020 - 08:02
#9
↑↑↓↓←→←→ba
There were a handful of other studies, including these:
https://forums.spiralknights.com/en/node/110324
https://forums.spiralknights.com/en/node/94227
I wrote up the shield HPs that I found here:
https://wiki.spiralknights.com/User:BlarghBlargh#Damage_Analysis
The reason people comment about the omega shell line is because that shield has 25% more hp than all other shields.
It's been awhile, but the 12.5% "damage" for status looks correct. On omega, it turns into 10% (because more hp). (Of course, the % changes based on the strength of the status) -- Status UVs on shields (if they don't come with the shield) are super useful.
I never fully finished the "heating" stuff, but there is an equation that someone came up with for how much damage is applied to armors. I believe shields function on the same stuff (only difference being the numbers are similar to 2 pieces of matching equipment). The interesting thing that I did discover was that lvl 10 4* items are for the most part better than unheated 5* (until ~lvl 9, depending upon which depth you are on... I even suggested how to reverse this "downgrade").
That logic follows even more strongly for shields, because they generally don't change bonuses at all (exception being heavy plate -> volcanic plate gaining fire resist). I haven't measured the resistance differences between 4* and 5*, but if they're anything like weapons (which I have measured)... lvl 10 4* shield is about equal to a lvl 8.5 5* heated shield (falls between 8 and 9). Lvl 10 5* is only about 6% more resistance than a lvl 10 4* (only in stratum 6). Using your numbers...
"Against the lumber it took 6% less damage (78.888/83.888), against the scuttlebot it took 10% less damage (20.833/23.055)."
A lvl 10 4* shield (compared to a lvl 1 5* shield) would take 5.5% less damage against a lumber and 9.4% less damage against a scuttlebot.
A lvl 10 4* shield (compared to a lvl 10 5* shield) would take 0.5% more damage against against a lumber and 0.6% more damage against a scuttlebot.
--
This assumption (shields scale similar to weapons) should probably be tested, because the "upgrade" bars of weapons seem to scale more than shields based on the images. The images have been known to be deceiving, though.
mar, 04/21/2020 - 06:25
#10
Oh, shock is legit elemental.
Oh, shock is legit elemental. I thought it was environmental like fire.
mar, 04/21/2020 - 16:10
#11
@skepticraven
I just did some tests on d19 and d24, it seems like 4 star lvl 10 and 5 star lvl 10 are very close in terms of tankiness on those depths. I don't know what happens deeper into stratum 6. The main question in my opinion though is what happens to 4 star scaling at the new shadow lair depth.
As for the upgrade to 5 star, there are a couple more shields that gain unique stats at 5 star. Crest of almire gains a unique shadow + fire + shock combination, Heater shield is a unique elemental + freeze combination (I don't like the tortoshields' MSD), ironmight plate shield gains piercing defense and barbarous thorn shield increases the sword dmg bonus (although it is not really a tanky shield and I never use it). So while 4 star shields might be very good, there are some 5 star versions you just can't ignore (especially the crest of almire).
I personally upgrade all my shields to 5 star because I want to put the strongest shields to the test in the new Shadow Lairs, and there's really only a few shields required to cover most defense+status combinations so getting those radiants isn't too much work. The radiant drought isn't nearly as bad as it used to be.
mer, 04/22/2020 - 03:32
#12
↑↑↓↓←→←→ba
Yeah, there are a bunch of unique stats for 5*... I just listed one as an example.
As for 4* on D30, I don't know what happens. If we look at how weapons function, don't expect them to drift too far.
Finally, all my shields are 5* because I have heated them. I make the 4* recommendation for those who don't have radiants lying around (because it is generally a major downgrade until lvl 8+). Only upgrade when it can be fully heated or it gains new bonuses.
mer, 04/22/2020 - 10:53
#13
The SL depth is rumored to be depth 40 though, so I was expecting it to make a bigger difference. I don't think anyone has recorded data for 4 star items post-SL update.
As for Tier 3, I did some more exploring. I didn't save my initial data from D19 and D24 (but those are easy to access through missions anyway), however here's the summary of what I found:
Data measured with Dragon Scale and Wyvern Scale shields (5 star and 4 star, both level 10)
D19: At the start of tier 3, the 5 star shield seems to be slightly stronger than the 4 star version (+- 2% less damage taken with 5 star)
D24: Interestingly enough, the 4 star shield seemed to have caught up with the 5 star. The 5 star was just barely stronger on D24 (+- 1% less dmg taken with 5 star).
D27: Against a Wolver (no status) on D27, my 5 star shield took 32.777% damage, while the 4 star took 34.666%. This is a difference of about 5.5% less damage taken with the 5 star (32.777/34.666)
I believe the 4 star spends the entire stratum 5 to catch up with the 5 star (the endpoint is probably d23, if someone has the required shields, it might be possible to confirm this in guild training hall against the bots). After they converge on d23, the 5 star slowly diverges away from the 4 star curve through stratum 6. I would have measured on D28 but there wasn't a good level for the shields that I have.
mer, 04/22/2020 - 16:16
#14
Shadow Lair
I just went to do more testing, this time in the Shadow Lairs:
Void Wolver, D40(?) (Unkown Passage) , Elite
Shield - Heat level - UV - Damage taken% (px/px)
Dragon Scale Shield - 10 - Stun Max + Elemental Low - 76.111% (274/360)
Wyvern Scale Shield - 10 - Stun Med + Freeze Med - 76% (228/300)
Aegis - 10 - Freeze Max + Normal Low - 56.388% (203/360)
Royal Jelly Shield - 10 - Freeze Max + Piercing High - 61.111% (220/360)
Surprisingly, the 4 star and 5 star shield took the exact same damage from the wolver bite. Or do Void Wolvers do pure normal damage? That might explain, since those shields don't have normal defense.
jeu, 04/23/2020 - 06:13
#15
↑↑↓↓←→←→ba
I know pre-update shadow lairs were D30. The way to verify would be using deathmark/armorless (howlitzer heads) attacks and checking player weapon damage. I suspect that they only modified the monsters.
The easiest way to check monster damage type is the color of the health bars that fly off the player when damaged. Red = normal, Green = elemental, and so on. Another way to confirm would be switching between "armorless" sets for the damage type (dragon scale, radiant silvermail, divine) to confirm that damage is identical. Be sure to have the same HP because %'s are much easier to check than measuring pixels/exact numbers.
jeu, 04/23/2020 - 08:26
#16
Yup, I know, but I didn't actually get hit by a void wolver except on my shields. I only realized when looking at the screenshots afterwards when the damage was identical. So I guess I might have to go and test again, but getting hit on purpose by the same attack without something else hitting you first is difficult in unknown passage... I'll try luring a slime next time.
Have you seen Glacies's epic threads about shield protection research? I would not describe them as "lacking". Of course, more research is always nice.
My sense, from earlier studies and from yours, is that status UVs are generally more powerful than damage UVs. Do you agree? (I mean, against attacks that contain status. Against status-less attacks, status UVs are of course worthless. Just as damage UVs are worthless against damage-less attacks.)
Thanks for your findings.