Spell reduction trait not working correctly with split damage

My sunbird does 115 damage. It should kill both targets but doesn’t. This wasn’t a one time occurrence, I’ve noticed many battles this weekend where Tina is blocking way more than she should.

3 Likes

They are going to tell you it’s working as intended. Since split damage could send 110 damage to the first troop and 5 to Tina.

I couldn’t find any GoW guide that has the mechanics for scatter/split damage, only this post from a player:

• Scatter/split damage - divide all your damage, and deal them to all 4 enemies. The damage amount on each troop random, but it can take out troops if their life is less than 1/4 of your total spell power.

What I take this to mean:

1. If calculates if any troop has less life than 1/X (X=remaining troops), and deals at least enough to kill them. (Can overkill happen? Is that what occurred in OP’s video?)
2. Divides rest of damage “randomly”, taking above into account.

In this case, it’s very possible Tina would survive. Your Sunbird’s damage divided by 2 is 57.5, which is not enough to kill Tina due to her Spell Block trait. RNG was not on your side when it split the damage.

However… this lead me to question how random is random? It always seems to divide the damage relatively “evenly”. What I mean is that I’ve never seen it deal only 1 or 2 damage to one troop, while dealing significantly more to another troop (when no troops are killed by the cast). Just to illustrate what I mean, I tested my Sunbird in explore on Warlord IV to ensure it couldn’t kill anything on its initial casts:

Battle 1
Cast 1 = 27, 27, 22, 21
Cast 2 = 20, 16, 30, 24
Cast 3 = 15, 16 (killed), 9, 21 (Sunbird had taken health damage, so unknown total damage)

Battle 2
Cast 1 = 22, 24, 32, 19
Cast 2 = 19, 19, 25, 27

There’s certainly some randomness in there, but not completely “random”. I assume this has been talked about before, but I couldn’t find any topics about it.

EDIT: So I tried another battle to see how damage is split when a troop is killed from split damage. I didn’t keep track of how much total damage Sunbird should have dealt in my Battle 1 example.

Battle 3 (against 4 Wall of Bones)
Cast 1 = 21, 23, 32, 21
Cast 2 = 20, 27, 20, 23
Cast 3 = 27, 26, 22, 15
Cast 4 (90 dmg total) = 33, 20, 16 (killed), 21
Cast 5 = 30, 28, 32
Cast 6 = 31, 28, 31
Cast 7 (90 dmg total) = 17 (killed), 45, 28

In Cast 4 and 7, 1 troop was killed but it was only dealt the exact amount of damage needed to kill it. The rest of the damage was divided among the rest of the troops. Why did this not happen in OP’s example?

EDIT2: Further testing with 1 troop that has Spell Block.

Battle 4 (against 3 Wall of Bones and 1 Ethereal Sentry in 2nd slot)
Cast 1 = 26, 13, 20, 24
Cast 2 = 20, 12, 24, 21
Cast 3 = 31, 9, 19, 22
Cast 4 = 13, 13, 22, 29
(Health left = 12, 13, 29, 18)
Cast 5 = 12 (killed), 10, 29 (killed), 18 (killed)

Cast 5 should have had a total of 90 damage. 59 was dealt to the walls, and killed all 3. 20 was dealt to the Ethereal Sentry, resulting in 10 actual damage. Where’s the remaining 11 damage?

EDIT3: Also previously reported and discussed in this thread. It appears to be due to the fact that scatter damage never deals enough damage to overkill, so it’ll never deal enough to kill an enemy with a spell reduction trait.

In my video Tina actually took 34 damage. That means the game chose to throw 68 damage at her yes? 68 - 50% = 34. Where did the figure 68 come from? Tina had 36 health and armor so I’m confused what is going on.

As to the game splitting sunbird’s damage and 57.5 not being enough to kill, I don’t think that’s how it works. If say I had 1 enemy with 5 hp and a 2nd enemy with 110 hp (but no spell reduction) sunbird would kill them both every time. But it’s working weirdly and differently with Tina for some reason.

I can answer this, and it’s working as intended, albeit not clearly. Scatter damage used to only deal enough damage to kill the troops that it hit, but with the introduction of the spell armour traits, this would always leave those troops alive because the base damage it was reducing was only equal to its remaining life, instead of the full damage. This was originally so that if one troop died, then the rest of the damage would still be dealt to the rest of the team, and there was no need for the excess damage.
Now, Scatter damage will do that same calculation, but the excess damage will be dealt randomly among all the targeted troops. It won’t prioritise the damage to a troop with a spell armour trait.

As we only display the damage actually dealt, this is something we could improve on and communicate better as you’re playing.

1 Like

So if I understand correctly in my video half of the “excess damage” was sent to the top enemy troop and therefore wasted? Wouldn’t it be better for that damage to only get split among enemies still alive?

It does to a point. It calculated TINA would be killed because it had a total of 36 life. Using your video as an example, the total damage would be 115. First, that damage is split between troops that are alive randomly. Chances are, it will detect that Beserker will die first as it only has 5 health. It will then deal the remaining damage to TINA until it also would die by dealing 36 damage to it. At this point its dealt 41 total damage.
The remaining 74 damage is then randomly split between Beserker and TINA. As TINA took 34 damage, the total dealt to her was 68 (as you surmised). So the excess damage split between them was likely 42 and 32 respectively. The Spell Block trait (and others like it) is not taken into account when calculating the total damage.

2 Likes

Should it be? I would claim yes. The current method leads to unintuitive outcomes like TINA surviving casts that clearly had enough damage available to kill it.

2 Likes

Thanks for the reply and the detailed explanation. In a perfect world the 42 damage wouldn’t be “wasted” on Beserker but instead go to Tina. Not a huge issue though, I’m just happy I know now what’s going on. It was bugging me trying to figure out why Tina took 34 damage.

1 Like

It is really awesome to have a dev explain game mechanics in such detail, thank you!

1 Like

We can review the intention behind it. It’s a unique case for sure.

@Magnusimus/@Fleg I am happy to help

5 Likes

So, is that a “we can review it, we don’t ever intend to” or a “we can review it, we’ll try to fit it in for 4.4”? The case isn’t really unique, the expectation for split damage is to never overkill a troop while there’s still one alive. I’m beginning to wonder how Faerie Fire factors into this.

Speaking algorithms, what you want to do when calculating split damage is to handle all spell reduction as temporarily increased life/armor. Tina has 50% reduction, so for calculation purposes her life/armor should be doubled (1/(1 - 0.5) = 2) prior to applying split damage, then halved (1 - 0.5) once the split damage has been applied. This approach works for all resistance values, even negative resistances like Faerie Fire.

I am going to hazard a guess that any changes won’t make it into 4.4. The development targets for that were probably set once 4.3 and 4.3.5 were out, and it won’t be easy to squeeze in something that takes a non-negligible amount of coding and isn’t a bugfix. That said, I am not on the dev team and they have surprised before with what they can accomplish. Let’s both cross our fingers—but not hold our breaths.

It’s just that I’ve seen a lot of customer facing statements, and this one looks carefully crafted to avoid committing to anything while making it appear otherwise to the casual reader. Which might be purely accidental, I’d just like to be sure prior to raising my expectations. I really appreciate the insight into split damage calculation, info this detailed is hard to come by.

Getting back to Faerie Fire, two troops present, one with 50 life, the other with 75 life and Faerie Fire, do both of them always exactly get killed by 100 split damage? Or will the first one likely survive, because split damage tries to assign 75 hits to the second troop, which gets multiplied by 1.5 afterwards?

It is more so that I can’t promise how or if we’ll make any changes. I can say that it’s likely there won’t be any change to it in 4.4 though.
That being said, it’s also helpful knowing what the expectation is for everyone. It helps us know how to address it.
And Faerie Fire behaves the same way in that it won’t be taken into account.

2 Likes

Would you be able to add it to the Known Issues list to raise awareness? Split damage is currently considered the weakest kind of damage. It doesn’t scale well with Magic (like area damage or splash damage) and it doesn’t kill any threats quickly (like single target damage). Its selling point up to now was that no damage gets lost. Spell damage reduction and spell damage amplification both significantly reduce the damage output of split damage by a random amount, making overall results unpredictable. This can be quite a nasty surprise for game modes like Siege Wars, players may want to rethink their team compositions.

It also helps us players to know when something is going to be addressed, to be able to plan ahead. Like the huge number of weapon upgrades considered harmful, due to turning predictable weapons into random board manipulators likely to help the opponent. It’s been almost a year now, would it be to early to hope a consensus on how to address that issue has already been reached?

Sorry for veering somewhat off topic, can’t just let the rare source of information visiting this forum escape with all the knowledge.

6 Likes

We usually keep known issues to be known bugs. As at the moment it’s not considered a bug, I don’t think it really fits there. Though we’ll look at adding a Help Centre article for the different types of damage, and go into more detail about it there.

Unfortunately that I cannot help you with.

1 Like