[Question] How much does damage reduce attack?

[Agent327]

I'm not sure if we're on the same page. Set strategy aside. Are you saying there's no rule of thumb to estimate how much less damage an injured unit will be able to inflict vs the same unit with full health, under the same conditions?

We are not on the same page, cause you are asking for the impossible and do not seem to understand you are.

Start a fight. Pick a unit you are facing and look at the expected damage the moment you can hit it. It is a variable, not a constant. Same goes for retaliation.

 

[Ben Jones the Newb]

I'm not sure anyone has done much analysis of how much difference it makes wounding a unit to how much damage that they deliver. With so many variables (including the range of the attack damage done) it's not easy to isolate factors to test them individual. The more favorable time for that would be very early on but that's the time when you have the least units to do the testing.

I guess if the damage drop-off was big enough to matter, it would be common knowledge. One way to test would be to find undefended cities and let the spearmen hit first, skip a turn, and let him again. Record his initial damage, his health after retaliation, and his second damage. Assuming it survives retaliation, and the same unit is used each time (e.g. a soldier), 30 repetitions would give enough data to have a good idea. Sounds like it's not worth the trouble, especially since those little green bars are so hard to see

 

[UBERhelp1]

I guess if the damage drop-off was big enough to matter, it would be common knowledge. One way to test would be to find undefended cities and let the spearmen hit first, skip a turn, and let him again. Record his initial damage, his health after retaliation, and his second damage. Assuming it survives retaliation, and the same unit is used each time (e.g. a soldier), 30 repetitions would give enough data to have a good idea. Sounds like it's not worth the trouble, especially since those little green bars are so hard to see

I think you don't understand the question you're asking, or I'm misunderstanding your question.

Attacking a unit doesn't lessen it's attack strength, UNLESS you are using a unit with the "heat" skill. Considering how fighting works, there are far too many variables (abilities, terrain, boosts, base stats, AI's choices (which are predictable), and more). For example, having a choice between killing an artillery unit vs a heavy might make a difference (the heavy could do more damage on a lesser turn, but the artillery can attack units further back and almost anywhere on the map).

At the point where anyone really would start to care about testing this, they are already autobattling everything. Ultimately, attack strength is not directly related to health. My hover tank can take damage when on a non-plains tile, but wipe out the rest of the enemy after it moves onto plains, for example.

In practical terms, here's the best way to choose which units to take out. Choose the one that has the most imminent threat or will be threatening down the road. These include artillery that will destroy your rare unattached units (I hate plasma artillery when using hovers...), or heavies with a lot of health you will have to start cutting away (think Assault Tanks). However, you may want to leave some units with extra health and not kill them. Those units can flip your rogues for you on the cMap, and that's what you want.

Hope this helps a little and I'm not just rambling

 

[Ben Jones the Newb]

We are not on the same page, cause you are asking for the impossible and do not seem to understand you are.

Start a fight. Pick a unit you are facing and look at the expected damage the moment you can hit it. It is a variable, not a constant. Same goes for retaliation.

I expressly said it's a variable range. If you did the exercise I described above, a simple linear regression model in R would give you a very good idea of what that value would be. The output would be something like, damage dealt=initial attack power * -log(damage taken), with a range (e.g. +/ 20 damage), and the confidence the damage dealt will fall within that range (more samples = greater confidence and narrower range). So you could take that formula, plug in initial attack and damage taken, and get a precise value. But more valuable would be the trendline. Obviously no one is gonna calculate that crap every attack.

Also, a linear relationship isn't necessarily constant. Real graphs don't fall neatly on a line. Consider the size vs age of babies. There's a very strong positive correlation, but you can't expect babies to be precisely average size; there's an expected range, You can be very confident a newborn won't be >20 lbs. Less confident that it won't be >15 lbs, etc.

I'm sure the real formula is piecewise and complex. They ain't posting that for us, and there's no need for me to break out R Studio to precisely estimate it (though it would be a breeze with enough data). I was hoping folks would know stuff like whether heavy cavalry has way less attack power when they're down to one health bar. Sounds like that's only true for artillery. Didn't mean to get into statistical weeds when I asked the question. It's not impossible to answer, but it's probably not worth collecting the data.

 

[Ben Jones the Newb]

I think you don't understand the question you're asking, or I'm misunderstanding your question.

Attacking a unit doesn't lessen it's attack strength, UNLESS you are using a unit with the "heat" skill. Considering how fighting works, there are far too many variables (abilities, terrain, boosts, base stats, AI's choices (which are predictable), and more). For example, having a choice between killing an artillery unit vs a heavy might make a difference (the heavy could do more damage on a lesser turn, but the artillery can attack units further back and almost anywhere on the map).

At the point where anyone really would start to care about testing this, they are already autobattling everything. Ultimately, attack strength is not directly related to health. My hover tank can take damage when on a non-plains tile, but wipe out the rest of the enemy after it moves onto plains, for example.

In practical terms, here's the best way to choose which units to take out. Choose the one that has the most imminent threat or will be threatening down the road. These include artillery that will destroy your rare unattached units (I hate plasma artillery when using hovers...), or heavies with a lot of health you will have to start cutting away (think Assault Tanks). However, you may want to leave some units with extra health and not kill them. Those units can flip your rogues for you on the cMap, and that's what you want.

Hope this helps a little and I'm not just rambling

This has gotten way more mathematical than I intended. I was hoping folks had a rule of thumb, like if your heavy knight has one bar of health, its attack is reduced by 20%. Sounds like it may only be true for artillery. I was making the point that if damage reduces attack, the relationship can be estimated with simple statistical models. I described how to collect the data, but I don't think it's worth doing.

 

[UBERhelp1]

This has gotten way more mathematical than I intended. I was hoping folks had a rule of thumb, like if your heavy knight has one bar of health, its attack is reduced by 20%. Sounds like it may only be true for artillery. I was making the point that if damage reduces attack, the relationship can be estimated with simple statistical models. I described how to collect the data, but I don't think it's worth doing.

Ok so that is entirely wrong. The health of the attacker has absolutely nothing to do with the damage it deals (unless it has an ability that explicitly changes that).

 

[DevaCat]

A former forum member xivarmy was into analyses of this type; too bad he got fed up and left. I seem to remember a discussion about how damage was calculated and it involved in addition to whatever formula a + factor added in, the nature of which is known only to Inno.

Anecdotally, I’ve had badly injured enemy units hit and do less than “normal” damage to mine, and I’ve had them do “normal” damage to mine. The factor at work? I don’t know. I kill the injured before they can hit me again.

 

[Ben Jones the Newb]

Ok so that is entirely wrong. The health of the attacker has absolutely nothing to do with the damage it deals (unless it has an ability that explicitly changes that).

I'd be breaking the forum rules to say where I got the information, but I didn't come up with idea. It sure seems like units with lower health do less damage, but as @DevaCat points out, anecdotes aren't strong evidence. I'll have to take your word for it.

@DevaCat , I can understand his frustration. There's some 30 rules that limit speech. The last time I recall such heavily restricted speech was during meals in Basic Combat Training, where we weren't allowed to talk at all.

My background is in biostatistics. When I read several claims that injured units deal less damage but the relationship wasn't clear, I suggested a pretty simple method to collect data. I was hoping folks would provide their data so I could test those claims and possibly give a function (or at least its curve) that might improve strategy, especially for younger folks who struggle to win battles. It's really elementary stuff that could be calculated by hand, but if it would violate rules I'll bow out.

I've got a bad habit of crunching data like that. I made 3d graphs in UFC 3 that proved which moves were best. I've proven which attributes matter in several different games. In MLB The Show 17, you could trade player cards and equipment and stuff online. It was tedious, so I created a system that monitored the buy and sell prices for me. That worked so well I modified it for buying and reselling stuff online....all I have to do is click the green button. Package comes to my door, I just slap a label on it and my work is done. My point is, this stuff might sound weird and useless, but the payoff is great. In all of those cases I kept the results to myself. Guess I'll do that again.

 

[Agent327]

My background is in biostatistics. When I read several claims that injured units deal less damage but the relationship wasn't clear, I suggested a pretty simple method to collect data. I was hoping folks would provide their data so I could test those claims and possibly give a function (or at least its curve) that might improve strategy, especially for younger folks who struggle to win battles. It's really elementary stuff that could be calculated by hand, but if it would violate rules I'll bow out.

Why not adress that at the place you get your info?

You are now at the baker telling the butcher told you something strange.

The info you refer to does not come from Inno, so Inno is not responsible for it.

 

[r21r]

reading this thread made me believe that injured troops deal less damage.
is it just me or ?

:S

 

[Johnny B. Goode]

Attacking a unit doesn't lessen it's attack strength,

Technically true, but misleading in the context of this discussion. A unit's health has no effect on its attack rating, but it definitely does affect the amount of damage it inflicts. It may be anecdotal, but I can't believe I am the only one who has noticed that units (both mine and the enemy's) that are at full strength simply do more damage in most situations than ones that have sustained damage themselves. If the unit's damage is minor it may be hard to see a difference due to other factors in the formula, but if a unit has major damage it is very clear.

 

[Ebeondi Asi]

-

 

[Ben Jones the Newb]

Why not adress that at the place you get your info?

You are now at the baker telling the butcher told you something strange.

The info you refer to does not come from Inno, so Inno is not responsible for it.

Because I assumed the owner of the game would be more knowledgeable. The only evidence I have that their claim is untrue is your claim that it isn't. Maybe you're right, but I'm scientifically skeptical. I don't know what your source is. Unless you've read the scripts recently, and you understand the language, my claim would be hearsay within hearsay.

I've tested other games, and I've yet to find one that behaves that doesn't behave differently than developers describe. That's probably because they're very complicated with several interacting files. Also, developers tend be assigned to different systems, so they don't understand (or even have access to) every component in those interaction. Think about it: how could they, especially in a game as big as this one. That's not criticism; it's the same reason physicians specialize in neurosurgery or gastroenterology, there's just too much data for even the smartest humans to absorb.

But statistics don't lie. Several people think they're seeing the same thing. Assuming they're reasonably intelligent and don't have conflicting interests, dismissing them is bold. Testing is easy: you just need a little data and know-how. The unmentionable source isn't a forum. They can't collect the 30 data points or so I'd need to prove whether a statistically significant correlation between damage taken and damage given exists.

 

[Johnny B. Goode]

Because I assumed the owner of the game would be more knowledgeable.

Unfortunately, this is not the place to reach the "owner" of the game. All you'll get here is other players and volunteer moderators. And the volunteer moderators' jobs do not include being in on every technical nuance of the game mechanics. Only the game development/programming/coding staff in Germany knows exactly how everything is supposed to work.

 

[Ben Jones the Newb]

Unfortunately, this is not the place to reach the "owner" of the game. All you'll get here is other players and volunteer moderators. And the volunteer moderators' jobs do not include being in on every technical nuance of the game mechanics. Only the game development/programming/coding staff in Germany knows exactly how everything is supposed to work.

Sorry, I was expanding on my answer when you posted this. I explicitly stated in an earlier post that I'm not asking for technical nuances, like piecewise functions for example.

I agree there's no reason to expect a forum moderator to have a perfect understanding of game mechanics. I don't know why you're describing the moderators' job to me. If you think he's claiming more complete knowledge than he has, that's between you two.

I have a hard time dismissing the folks who think they're seeing what I'm seeing. After all, they're the ones I expected to get opinions from. The other unmentionable source has been accurate without exception. It's entirely possible that the moderator is correct. I don't know what source he's basing his claim on. In my view, the evidence that an injured unit may inflict less damage is stronger than the evidence that it certainly does not. I don't have any bias or conflict of interest.

So where does that leave us? I submit it's worth keeping an eye on. It could improve strategy, especially for young players. If enough people say they see injured units dishing out less damage - or not - I think we would get a pretty good idea of how things work.

(edited to correct grammar)

 

[Johnny B. Goode]

I agree there's no reason to expect a forum moderator to have a perfect understanding of game mechanics. I don't know why you're describing the moderators' job to me. If you think he's claiming more complete knowledge than he has, that's between you two.

I was just pointing out that, despite being "staff", moderators do not have any special knowledge due to their position. Since you mentioned seeking knowledge from the "owner", I thought it would be beneficial to point out to you that this is strictly a players' forum, with the exception of Announcements. I am definitely not in any way trying to say that any moderator is claiming more complete knowledge than they have.

The other unmentionable source has been accurate without exception.

If that other source is the one I think it is, then it is also a player source. And as such cannot possibly be "accurate without exception", because I have never seen a player (including myself) who is always correct.

So where does that leave us? I submit it's worth keeping an eye on. It could improve strategy, especially for young players. If enough people say they see injured units dishing out less damage - or not - I think we would get a pretty good idea of how things work.

Well, I don't know about you, but in every game I have ever played that had fighting in it, damaged units/troops always dish out less damage than healthy ones. As far as "how much" (the question in the thread title), all you'll ever get from players is speculation and small sample size answers that would not stand up under scrutiny. Given the many factors involved in combat, it would be herculean task for a player to reliably figure out how much less damage a unit inflicts in even a handful of cases.

 

[Ben Jones the Newb]

I was just pointing out that, despite being "staff", moderators do not have any special knowledge due to their position. Since you mentioned seeking knowledge from the "owner", I thought it would be beneficial to point out to you that this is strictly a players' forum, with the exception of Announcements. I am definitely not in any way trying to say that any moderator is claiming more complete knowledge than they have.

If that other source is the one I think it is, then it is also a player source. And as such cannot possibly be "accurate without exception", because I have never seen a player (including myself) who is always correct.

Well, I don't know about you, but in every game I have ever played that had fighting in it, damaged units/troops always dish out less damage than healthy ones. As far as "how much" (the question in the thread title), all you'll ever get from players is speculation and small sample size answers that would not stand up under scrutiny. Given the many factors involved in combat, it would be herculean task for a player to reliably figure out how much less damage a unit inflicts in even a handful of cases.

That's fair. I was asked why I didn't tell my source their claim (that injured units do less damage) was wrong; the point I was trying to make is I don't have any evidence to support that claim. The burden of proof would fall on me, not them. The only realistic place to find that evidence is right here.

I didn't say it is accurate without exception. I said it has been accurate without exception. It is certainly reliable enough that it would be unreasonable to reject their claim just because someone says, "nu uh." On a side note, I've always been required to cite the source of little claim I make, so being prohibited by rule is just wild.

I described a method to collect the data in a statistically valid manner using undefended cities. If someone wanted to do that, I'd invite working out the nuts and bolts of the study, but I think it would be overkill. Why do you think the sample size would be too small? If 20-30 people described just one battle, that would be compelling. We're already at nearly 40 posts. This isn't supposed to be a peer reviewed scholarly paper, we're just talking about general behavior. It's interesting that you say injured troops always dish out less damage. I don't know.

 

[r21r]

lost in translation.

only reasons i personaly use injured troops while fighting are :

1) keep a rogue with lower health than the single troop so aggro goes on him and the single troop survives
2) im 100% sure that with the current injuries my fighting army will win without losses

or 3) racing an enemy guild so i have to risk my troops instead of losing time replacing them.

now, IF (as yourself said) somehow damage works as "heat" on the receiver, is interesting, but you did all this way to tell us to test it ... ?
i know 90% what to expect when fighting and i don't know if this will change anything i do, maybe if we prove it, we might end up taking less damage but i doubt anyone is using 1/10 HP troops into fight so i don't really know what are we after here.

it could be a bug but it's irrelevant for players, we always look how to single kill not how to leave the enemy with the less HP

 

[Agent327]

Because I assumed the owner of the game would be more knowledgeable.

They probably are, but as far as I know they have never made any statements on the subject, so you are knocking at the wrong door.

 

[UBERhelp1]

The other unmentionable source has been accurate without exception.

It really isn't. It's frequently been wrong with events, and players have gotten mad over that. The quests are almost always messed up somewhere, and new versions keep showing up to stuff that's been around for 5+ years. That site should be read thoughtfully, but you shouldn't ultimately rely on something there entirely.