Casualty Rates

[Bronson]

Popular forum logic was that you look for rivals with defence way lower than your offence in order to decrease casualty rate, well this below post from Crime City Mark might blow you away......

"Comparative attack and defense have no impact on casualties in any of our games. You could have a billion attack and they have 1 defense and you'd still have the same casualty rate.

Just because your glass cannons do a ton of damage, doesn't mean they're invincible."

[War Priest]

Comparative attack and defense have no impact on casualties in any of our games. You could have a billion attack and they have 1 defense and you'd still have the same casualty rate.

Just because your glass cannons do a ton of damage, doesn't mean they're invincible.



Seriously!!!

So I could attack someone with 50% less D to my A and lose just as much as if I was attacking someone with just 5% difference???


Yeah, that's exactly right.

Now you can stop looking for weaker rivals to reduce casualties... :)

[Bronson]

New color man. You can hardly read that.

its my first time using color lol yeah that was hard on the eyes how about now???

[Luke7676]

It's nice to know this now. Glad it was cleared up.

[War Priest]

its my first time using color lol yeah that was hard on the eyes how about now???

Haha, it's better. I included the quote that way nothing is misunderstood.

[Aidan]

All i can say is this is not true for me and many others. I'll post screenshot tomorrow to prove this.

[War Priest]

All i can say is this is not true for me and many others. I'll post screenshot tomorrow to prove this.

Coincidence?

[stricker]

All i can say is this is not true for me and many others. I'll post screenshot tomorrow to prove this. there are too many variables at play and they are layered at that... for example, there is a unit casualty rate but there is also a %reduction rate associated with a lot of units also! ...and anyway, if some stroke of luck we did figure it out, the dev gods would do the "tower of babel" confusion shuffle just for the fun of it... LOL

so, two scenario snapshots, aidan, does not a system make. there are several folks here on the boards that are quiet good at recognizing generalized patterns but imo, that's the best we'll every be able to do!!!

[Aidan]

Coincidence?

Doubt it. I have the screenshot now. Just need to do some cropping

[Aidan]

there are too many variables at play and they are layered at that... for example, there is a unit casualty rate but there is also a %reduction rate associated with a lot of units also! ...and anyway, if some stroke of luck we did figure it out, the dev gods would do the "tower of babel" confusion shuffle just for the fun of it... LOL

so, two scenario snapshots, aidan, does not a system make. there are several folks here on the boards that are quiet good at recognizing generalized patterns but imo, that's the best we'll every be able to do!!!

I'm lil bit weary now to post it. What if they fix me? I understand there are variables but you cant call it a luck if you get the same outcome everytime

[stricker]

I'm lil bit weary now to post it. What if they fix me? I understand there are variables but you cant call it a luck if you get the same outcome everytime
...you need to be 'fixed' aidan!!! LOL

[JohnnyR]

In total agreeance with Aidan. My experience has been different.

[cheiz]

Interesting! Was trying to figure out why I was losing high value units against much lower stats and sometimes not against higher stats. Was thinking maybe the boost buildings made the difference, but apparently that should not be the case as those only affect A/D levels. So I guess maybe the "strong against" units could be more of an influence on the losses. Has anyone else been keeping track of this?

[stricker]

...don't get me wrong. my casualty rate seems extraordinarily high most of the time also!!! ...and my paranoia side doesn't discount the conspiracy theory possibility that individuals can be 'fixed'... but that would be on the edge of extreme prejudice on the part of funzio, so probably NOT! ...my comment to aidan is more like instead of looking at an act... look at their ways....... what have been the 'ways' in the mw theme!!!

[Bronson]

...don't get me wrong. my casualty rate seems extraordinarily high most of the time also!!! ...and my paranoia side doesn't discount the conspiracy theory possibility that individuals can be 'fixed'... but that would be on the edge of extreme prejudice on the part of funzio, so probably NOT! ...my comment to aidan is more like instead of looking at an act... look at their ways....... what have been the 'ways' in the mw theme!!!

Dafuk!!! What "ways"???

[stricker]

lol... small bits a pieces of info... generic we are looking into it reply... good strategies being used by the masses suddenly are shaken up by an update or event... etc. etc. etc.

those 'ways'... patterns... whatever you want to call it... ;p

[Bronson]

lol... small bits a pieces of info... generic we are looking into it reply... good strategies being used by the masses suddenly are shaken up by an update or event... etc. etc. etc.

those 'ways'... patterns... whatever you want to call it... ;p

I got you! CCM is posting on the other thread again so ill copy his replies over now

[Bronson]

So is there anyway to attack someone and have a better chance of not losing units??
.

Crime City Mark - Have an army full of units with low casualty rates.

You know, I've mentioned the whole "A/D have no impact on casualties" thing a number of times in the past on the forums

[Freekizh]

You know these guys are sneaky..half truths can be misleading.

[War Priest]

Bronson, no man. Old photo...

[cheiz]

Crime City Mark - Have an army full of units with low casualty rates.

You know, I've mentioned the whole "A/D have no impact on casualties" thing a number of times in the past on the forums


You know these guys are sneaky..half truths can be misleading.

For me that strengthens my idea to look (out) for "strong against" units

[frenda]

For me that strengthens my idea to look (out) for "strong against" units

I'm not buying that line. "Strong against..." would work only if you know the exact balance of the other guys' units. I'm taking disproportionately high losses in my low to very low casualty rate units. And the BS about there being no difference in loss rates whether attacking a high or low defense rival is ridiculous. What fairy-tale combat model is that based on? If MW is going to suspend all rules of combat and logic, then you can just as much fun playing roll the dice and if you get even you win and if you get odd you lose. It's time the game was tightened up some, it's really losing all appeal to me at this point.

[stricker]

I'm not buying that line. "Strong against..." would work only if you know the exact balance of the other guys' units. I'm taking disproportionately high losses in my low to very low casualty rate units. And the BS about there being no difference in loss rates whether attacking a high or low defense rival is ridiculous. What fairy-tale combat model is that based on? If MW is going to suspend all rules of combat and logic, then you can just as much fun playing roll the dice and if you get even you win and if you get odd you lose. It's time the game was tightened up some, it's really losing all appeal to me at this point.

.......2nd'd.

[frenda]

.......2nd'd.

And the opening rate on the event sucks! Even with the cash only option I've only gotten 1 out of 4 right now. This early in the event and the rate already sucks, geez, can't wait until the last few hours... NOT!

[cheiz]

It was a hypothesis, still needed testing

I've been trying to avoid all strong against air units (takes more time, because you need to check the whole inventory) and attacked regardless of rival stats (taking in account I could actually win with regards to defense + boost)

Did this the last 24 hours and I noticed much less casualties of my air units. Still lost some high value items, but I did not take in account the other strong against units.

I must also admitt, the sample wasn't that big, maybe 100 attacks, so not very conclusive, though enough for me to keep trying this for a while

[Ramshutu]

I've been researching strong against interactions with unit losses. There is no appreciable connection between the two.

In terms of unit losses, while not technically 'units', Crime City explosives are expendable, and are only lost when A/D is close between two losses.

[frenda]

I've been researching strong against interactions with unit losses. There is no appreciable connection between the two.

In terms of unit losses, while not technically 'units', Crime City explosives are expendable, and are only lost when A/D is close between two losses.

Concur, and since we know in CC that explosives are expendable when we buy them, it's no surprise when you do lose some.

[cheiz]

I'm not buying that line. "Strong against..." would work only if you know the exact balance of the other guys' units. I'm taking disproportionately high losses in my low to very low casualty rate units. And the BS about there being no difference in loss rates whether attacking a high or low defense rival is ridiculous. What fairy-tale combat model is that based on? If MW is going to suspend all rules of combat and logic, then you can just as much fun playing roll the dice and if you get even you win and if you get odd you lose. It's time the game was tightened up some, it's really losing all appeal to me at this point.

Hi Frenda, I do BTW totally agree on your comments on it being absurd that A/D stats have no influence on loss rates

[Maverick50727]

FYI,
My strategy all along has been low consumption with strong against. Everyone who knows me well knows that. Maybe that could be a factor in my super low loss rates as opposed to an outright bug. But I still won't rule a bug out.

A thing to note here is that many of the better units (higher A/D stats) have better casualty rates. When people increase their A/D stats higher than people under them, they do so by buying better units to replace those weakest ones in thier "battle group". So even though not by design, the rule of thumb rule of attacking weaker players works by that very nature. The better units often have the lowest casualty rates. It also goes to explain occasional losses to people much weaker than you.

I recommend a few people find rivals who aren't losing very many units when you attack then do some indepth analysis of their units.

I find all of this very interesting.

[Poopenshire]

I just completed 24 attacks in the last few hours. I have lost enough Super Hornets and Stealth Frigates for ~3000 in valor, and not even a 10th of that from fights. Both of these are LOW casualty rate. Please stop saying nothing is wrong with the rates. there is something obviously wrong with the rates and Valor units are dying in a manner not conducive to low rates. I have valor units dying faster than units with a High casualty rate.

[Q Raider]

Some facts and figures for you guys to mull over. Been logging lost Valor items since I got the EAD to determine loss rates. I also upgraded my stats using all Valor units a couple of days ago after researching what I thought was reasonable.

The results indicated something very different with the effective loss rate almost doubling.

Before figures
Attacks 955/39
Raids 615/0
Loss ratios
SH 1 in 4846
SF 1 in 9896
EAD 1 in 1554
HM 1 in 1465

After figures
Attacks 158/3
Raids 150/0
Loss Ratios
SH 1 in 2207
SF 1 in 5287
EAD 1 in 870
HM 1 in 1088

As it has been stated recently that relative A/D has no bearing on losses then I looked for another reason.

Prior to updating my Valor units comprised 39% of my total defence stat, after it was 43%.

SH went from 53% to 57% of total Air Power
SF went from 16% TO 21% of total sea power
EAD went from 38% to 42% of total Ground power
HM went from 32% to 36% of total Infantry power.

Current thoughts are that running Valor units above 40% of total strength is not adviseable. Am aiming to get mine back under 40% and see if the loss rate improves.

[Q Raider]

So for my money if you see someone with Valor representing over half their attack strength then they probably have the low loss glitch.....

[KCh]

So for my money if you see someone with Valor representing over half their attack strength then they probably have the low loss glitch.....

Not necessarily. I have about 1500 valor units making up my attack strength but take a lot of losses if not I'm not hitting certain targets. I have about 15k fights won. The people with the low loss glitch are the ones that have way too many valor units for their amount of fights and raids. Some of these people only have about 8k-12k fights won but they have over 2000 valor units and attack strong players without caring because they aren't taking any losses.

[Zydrike]

To me, if I attacked anyone 20% of my strength, I hardly loose anything, however, when I attack someone 30% of my strength, I start to lose heavily. Coincidence, my casualty reduction is sitting @ 30%, not sure if that's what Funzio means by "reducing the casualty".

[Q Raider]

KCh
Was a statement based on my observations from level 77 through to current of 87. You must earn a lot more Valor at higher levels. Currently my Valor cost per attack or raid works out as 8.5 and attacks at my level do not return those amounts regularly. I have chased the degradation missions up to 210 allies (I have 116) just trying to break even. Actually have currently run out of targets so will have to ally up to start seeing some more.

Also look at a couple of other things.

Too many loot items, one guy had 150 heavy bombers, they were in his attack strength stats and he had only completed a few dozen PVE missions.

Too many GIGN's they dropped like flies for me.....

Very high attack to raid ratio.
Low PVE mission counts.
Little or no Gold items purchased.

[Aidan]

To me, if I attacked anyone 20% of my strength, I hardly loose anything, however, when I attack someone 30% of my strength, I start to lose heavily. Coincidence, my casualty reduction is sitting @ 30%, not sure if that's what Funzio means by "reducing the casualty".

Yup, that's how it is for most of us but somehow funzio see it differently. Bug? Glitch? Misunderstanding? Mehh i dont know.

[Zydrike]

Yup, that's how it is for most of us but somehow funzio see it differently. Bug? Glitch? Misunderstanding? Mehh i dont know.

Hi, Aidan, what's your casualty reduction rate? Not sure if casualty reduction is used to protect you with weaker players, e.g. you will hardly lose unit again the player with 10% of your status if you casualty reduction is sitting @10%

[Aidan]

Hi zyndrike
I was agreeing to your first statement. Low or zero losses when attacking a much much weaker rival. High losses when attacking stronger rivals

On your theory that casualty reduction ties to player str, i dont know. I dont think so. I have 10% from infirmary and 2% from BS Heli. So 12%.

[Zydrike]

Hi zyndrike
I was agreeing to your first statement. Low or zero losses when attacking a much much weaker rival. High losses when attacking stronger rivals

On your theory that casualty reduction ties to player str, i dont know. I dont think so. I have 10% from infirmary and 2% from BS Heli. So 12%.

I was suspecting casualty calculation build up by 2 steps, first, based on you attack and opponent defense determines how many units you lose, and then based on the spread and casualty of your unit, decide which unit you lose. That's why meat shield works. Any thoughts?

[Aidan]

I find that meatshield works too but CCM's post contradict that. I'm changing my tactics now. I'll be replacing all my meatshield with low and very low casualty units. See how i fare. if i still have high losses, i'll be suing CCM. Lololol


Comparative attack and defense have no impact on casualties in any of our games. You could have a billion attack and they have 1 defense and you'd still have the same casualty rate.

Just because your glass cannons do a ton of damage, doesn't mean they're invincible.


Have an army full of units with low casualty rates.

You know, I've mentioned the whole "A/D have no impact on casualties" thing a number of times in the past on the forums.

[Ramshutu]

I find that meatshield works too but CCM's post contradict that. I'm changing my tactics now. I'll be replacing all my meatshield with low and very low casualty units. See how i fare. if i still have high losses, i'll be suing CCM. Lololol

There is a cap on the losses you suffer. Were it not for this, the chances are that you would loose many more units.

Because there is a cap, it means that the server goes through your unit list one by one, until this cap is reached at which points it stops the calculation.

This means the unit loss rate and number of units is not the only contributing factor, the position of the unit in this list is important as well. For example, if you have 2000 units of a given type with a 1% loss rate, and 100 units with a 95% loss rate, and your loss cap is 4 units. if the 2000 units are processed first, the chances are that you will 4 units of the 2000 due to loss rates, and unlikely that you will lose any of the 100 units as a result.

Now, the important point to note, is that I haven't quite worked out what the order actually is. If it is low atk to high atk, this means that meat shielding will work. If it is some other order, you have to be a little smarter with the meat shielding.

[Aidan]

Interesting theory you got there ram. May i know how you came to that conclusion?

[Ramshutu]

Interesting theory you got there ram. May i know how you came to that conclusion?

The cap can be inferred by calculating what losses you should be suffering. I worked out that given my unit breakdown, I should be loosing on average about 12 units per raid. I was actually loosing 3 max.

I suspect the cap is dynamic, and changes for each fight, it can be any value between 0 and N.

[LAX]

I think there is a new twist to casualty rates and which units you lose. I have experienced low casualty and weaker unit losses when i attack a low ranked guy like a private with similar or a little less Att/Deff than me. When i do the same to a sergeant which is the same rank as i, losses are heavy min. 2 units; like a hawk drone, APC, avenger, gign.

no statistics to back it up, just plain old observation.

[Freekizh]

Because there is a cap, it means that the server goes through your unit list one by one, until this cap is reached at which points it stops the calculation.


Just because there is a cap doesnt mean that there needs to be an order. I can assign a number 1 to N for each unit into battle, generate a random number, and the unit with that number is then consumed. Repeat loop until cap reached.

[Ramshutu]

Just because there is a cap doesnt mean that there needs to be an order. I can assign a number 1 to N for each unit into battle, generate a random number, and the unit with that number is then consumed. Repeat loop until cap reached.

Yes, but that would mean that the relative numbrer of a given type of unit was the only deciding factor, which is not how it works. For example with the above, my losses should be (in terms of regularity) frigates, stealth frigates, super hornets.

[Freekizh]

Yes, but that would mean that the relative numbrer of a given type of unit was the only deciding factor, which is not how it works. For example with the above, my losses should be (in terms of regularity) frigates, stealth frigates, super hornets.

That was just a simple illustration that ordering is not necessary. If you wanted losses to follow the expected values, there are numerous implementations..e.g., weight the units given numbers by it's casualty rates, etc etc. Anyway it was a simple point, and not meant to definitively say how it actually works.

My view is that ordering is not strictly necessary, I.e., according to consumption rates like you are describing.

[Ramshutu]

That was just a simple illustration that ordering is not necessary. If you wanted losses to follow the expected values, there are numerous implementations..e.g., weight the units given numbers according to casualty rates, etc etc. Anyway it was a simple point, and not meant to definitively say how it actually works.

My view is that ordering is not strictly necessary, I.e., according to consumption rates like you are describing.

Before i start, Consumption rate = casualty rate.

Now Of course it's not, strictly speaking, necessary. The servers may, of course, take an average of the DOW jones industrial average to generate a pseudo random number that can be used to weight a google search for all units that are being calculated and use this to determine which units are lost.... But it simply is not practical.

I am looking at this as a black box, and as a programmer. The unit loss algorithm can be calculated in many ways, but given what I know about it's outputs, and what is practical, the options have to be one of the following:

1.) the complete unit list is searched, with a "loss check" on each unit being made. If the check fails, the unit is lost. The lost unit is added to a list, and up-to N units are selected from this list and flagged as 'killed'

2.) the unit list is searched, with the same loss check as above. If the check fails, the unit is lost. The lost unit is added to the list, and when this list is N units long, the items are flagged as killed, and the search ends.

Now, from a pure programming perspective, #1 is the preferable option and it gives something that closely resembles the causality rates reported in the game data. However, for this to be what is currently happening, the unit losses experienced will be in proportion to casualty rates, which they appear not to be from my analysis.

From a server load perspective, #2 is better, as it vastly reduces the amount of processor time required to deal with a fight. This may not seem like a big deal, but even a small fraction of a second optimisation to the algorithm will have a massive cumulative effect with the number of concurrent fights.

This latter point is the reason that I think it is massively unlikely that some random selection or ordering of units is taking place. Any significant re-ordering, or randomisation algorithm over and above existing lists and searches will add significant processing time.

The fight stats, at least from a few weeks ago when I was last playing about, shows that for a particular unit you loose, the stats seem to work out properly, but taken as a complete analysis of your full unit breakdown, there is most definitely a skew in the units you loose. A randomised mechanism of selecting the unit to lose would not account for such a skew. Infact the only way to account for such a skew is a systemic bias in the algorithm, and in such an algorithm the only realistic source of bias will most certainly be list position bias.

[Freekizh]

I think you're making it too complicated than it needs to be. What is this garbage about the Dow? Do you even know how a random number generator works? Are you saying the Dow is generating random numbers? Let's not start that one!

You don't need to be a programmer to write simple loops or to understand conceptually wot can be implemented. Anyone can code this up in excel. In an hour max. You can make it more efficient in many ways by reusing randn, Vectorization methods, etc, so CPU cycle/sever garbage aside. Even if hundreds of thousands of fights take place each second, that is not a burden for such loop.

My point remains, you don't need ordering, as you admitted. The only diff between my speculation and your speculation is that u represent it to be how it works. And speculation without data is meaningless.

[Ramshutu]

I think you're making it too complicated than it needs to be. What is this garbage about the Dow? Do you even know how a random number generator works? Are you saying the Dow is generating random numbers? Let's not start that one!

You don't need to be a programmer to write simple loops or to understand conceptually wot can be implemented. Anyone can code this up in excel. In an hour max. You can make it more efficient in many ways by reusing randn, Vectorization methods, etc, so CPU cycle/sever garbage aside. Even if hundreds of thousands of fights take place each second, that is not a burden for such loop.

My point remains, you don't need ordering, as you admitted. The only diff between my speculation and your speculation is that u represent it to be how it works. And speculation without data is meaningless.

Any non programmer can knock up a searching algorithm in excel. Unfortunately this is not excel.To make this work with a thousand concurrent fights on servers, interacting seamlessly and probably asynchronously with random access databases to obtain correct results all of the time requires someone who actually knows what they are talking about, and is the main reason why people who do it for a living can be paid quite a lot of money, rather than marginally more than shovelling chips in burger king.

Ive presented what I know the raw statistics to be, Compare the losses of 200 ATs vs 200 snipers for example, and presented a detailed reasoning why this could be the case.

While you are quite at liberty to disagree, the simple fact remains that there is definitely a bias in the system. This bias can only be explained by a non random skew in the selection of units. I have given the most likely explanation for this, given my experience in the field. While you can disagree with this, your counter suggestions match neither the facts nor the typical practical considerations of a multi user server system that are generally many orders of magnitude more complex than 'knocking something up' in excel.

All I am trying to do, is provide testable explanations, based on stats, data and my experience to help optimise people's strategies in order to minimise losses. You can argue that the data is wrong, or that there is a more appropriate programming solution that yields the same result, but simply disagreeing by throwing out suggestions that dont work isn't particularly constructive.

[Freekizh]

Again your rambling on fancy lingo garbage.

Why need to bring up useless facts about fancy server etc etc when we are talking about the concept? Do you think other people know nothing about this or hosting up complex product on servers or mainframes? Do you think everyone is ignorant?

Why obscure a simple concept with useless drivel about servers and Dow jones random numbers..haha I am still dying from that one.

[Ramshutu]

Again your rambling on fancy lingo garbage.

Why need to bring up useless facts about fancy server etc etc when we are talking about the concept? Do you think other people know nothing about this or hosting up complex product on servers or mainframes? Do you think everyone is ignorant?

Why obscure a simple concept with useless drivel about servers and Dow jones random numbers..haha I am still dying from that one.

I'm bringing up the issue of servers rather than 'concepts' because the games run on servers, and to ignore this, when it can have a major impact on what you can or can't do is a bit stupid.

If you want is appropriate and possible, fine, but this is why I mentioned the Dow jones. While you may not have grasped the sarcasm in the post, I was trying to illustrate 'just because something is possible, does not mean it is appropriate.'

Now I don't think everyone is ignorant, but it is obvious you do not understand or are at least very naive about the issues and problems that come about when you are trying to develop this type of system.

[Freekizh]

Duuude are you the type of guy he makes out that his job is so difficult that only he knows how to do it? Do u really really believe that running a rudimentary randomization routine that an 8th grader could do on an Atari to be so insanely difficult that it can only be done in a certain way? Have you ever done this yourself? Please dont insult my intelligence with your naiivity.

If the answer is yes it is so difficult, then guys like you are the reason why IT budgets cost 100MM instead of 100k with massive cost of overruns.

[Ramshutu]

Duuude are you the type of guy he makes out that his job is so difficult that only he knows how to do it? Do u really really believe that running a simple randomization routine is so insanely difficult that it can only be done in a certain way? Have you ever done this yourself? Please dont insult my intelligence with your naiivity.

As I've said, Randomisations are really easy to implement. Running it on a server where there are many other issues to consider other than it 'working' is a lot harder. In my profession, I have spent many, many years dealing with seemingly 'simple' problems that cant be solved simply because of speed, loading or resource issues, including a great deal that are very similar to this. I don't profess that other people can't do my job, they can. I am, however, quite certain that you can't.

Anyway. Like most posts on the internet, This is deteriorating and getting seriously off topic and serves no purpose other than looking like a douche, I will not continue to engage with you on this. What I'm saying in the above posts relating to the algorithm speaks for itself, if you want actually want to deal with the points I have raised rather than to simply say that I have no clue what I'm talking about, I will be happy to talk further.

[Freekizh]

Haha u r hilarious. Thanks for the exchange. You never open to any ideas except ur own.

That's why I stopped trying. I am glad there are so many wankers like u in the world to amuse me.

[Warfiend]

Why obscure a simple concept with useless drivel about servers and Dow jones random numbers..haha I am still dying from that one.

The servers are entirely relevant.

[Freekizh]

The servers are entirely relevant.

No bud I do not thinks its relevant becos I think it's calculated on your device. There are two reasons why I think this is the case..the casualty rates being contained and updated on our devices, and the low casualty rate glitches people have. If the server did all the calculations, I cannot see the need to have those rates on our machines or the observance of the low casualty glitch.

Even if this was not the case, the randomization routine is not at all intensive. Pseudo random number generators are very fast.

[JMC]

In my experience, except when the casualty rate was way too high, A/D score comparison has always mattered. Currently, under 10% strength is usually zero casualty. Between 10% and 15% is fairly low casualties, but still take some about 1 in 5 fights. 15% to 25% is probably 2 in 3 fights will give casualties. Between 25% and 35%, i will get casualties consistently, almost every single fight. A rare fight, maybe out of 100 will net zero casualties. Above that there is absolutely no chance of getting a zero loss fight.

The types of units i lose seem to be the same at each range, just the amount i lose changes. Meatshields work to some extend, but they do not work every time like some people seem to think. In fact they only work about 50% of the time. The other half of the time, i am losing valor units and high value units.

[Freekizh]

In my experience, except when the casualty rate was way too high, A/D score comparison has always mattered. Currently, under 10% strength is usually zero casualty. Between 10% and 15% is fairly low casualties, but still take some about 1 in 5 fights. 15% to 25% is probably 2 in 3 fights will give casualties. Between 25% and 35%, i will get casualties consistently, almost every single fight. A rare fight, maybe out of 100 will net zero casualties. Above that there is absolutely no chance of getting a zero loss fight.

The types of units i lose seem to be the same at each range, just the amount i lose changes. Meatshields work to some extend, but they do not work every time like some people seem to think. In fact they only work about 50% of the time. The other half of the time, i am losing valor units and high value units.

That's right JMC meat shields don't work all the time that's why I don't think there is a loss priority or ordering like people think.

The biggest conundrum is that CCM said A/D relativity doesn't matter, and the only way I can reconcile this with what I have observed and what people are saying is that A/D is often, but not always correlated to number of units the rival brings to battle.

Conceptually, if you are attacking a low unit rival, he/she should have low number of A/D stats usually. But not always - so the theory is that if you attack a low A/D guy with low units you should get lower losses because less of your units are engaged. But not all low A/D guys have low units, some have just low density but high units, which should engage more of your units and therefore cause you more casualties.

[Warfiend]

No bud I do not thinks its relevant becos I think it's calculated on your device. There are two reasons why I think this is the case..the casualty rates being contained and updated on our devices, and the low casualty rate glitches people have. If the server did all the calculations, I cannot see the need to have those rates on our machines or the observance of the low casualty glitch.

I'm not going to pretend to know exactly how it's done, because I don't, however, I think there are indications that the server plays a role in it.

First is the perceived casualty rate fluctuation during events. Granted, this could be a case of the server simply sending a different variable to the client to perform the calculations. Or it could simply be coincidence shared by a lot of people and it doesn't actually change.

Second is that when I do an attack, if anything is wonkie with my network connection, it hangs and will even time out if the connection doesn't improve. Virtually all my attacks are done after checking my rivals stats and inventory, so all unit related variables should be on my client at that time, so it is definetly doing something with the server when it's doing those calculations.

Finally, I've had my client crash countless times while attacking, before I can see the results. When I go back into the game, my inventory is updated according to the attack having transpired, despite my client having crashed.

I admit none of this is conclusive evidence that the server is involved in the calculations but in my opinion, they point to the server playing a role. Let me stress that I am not saying you are wrong. I've been wrong about many things, and this could be one of them. I'm just putting my opinion on the topic out there.

[Q Raider]

I have the same scenario as Warfiend in regard to both method prior to attack and attack crashouts.

Prior to the update a couple of weeks ago I actually could not perform an attack at all as the game crashed every time. First raids also crashed most of the time.

When I ran back up my inventory, missions, experience points and Valor were updated but the cash was missing.

After the update (and until this weekend) the crash rate has been much lower.

Given that there is effectively a dual event going on at present then it seems logical that the server does play at least some part in the calculation process and is currently struggling to keep up hence the increased crashes being experienced. Though at least this time I am getting the cash....

[JMC]

That's right JMC meat shields don't work all the time that's why I don't think there is a loss priority or ordering like people think.

The biggest conundrum is that CCM said A/D relativity doesn't matter, and the only way I can reconcile this with what I have observed and what people are saying is that A/D is often, but not always correlated to number of units the rival brings to battle.

Conceptually, if you are attacking a low unit rival, he/she should have low number of A/D stats usually. But not always - so the theory is that if you attack a low A/D guy with low units you should get lower losses because less of your units are engaged. But not all low A/D guys have low units, some have just low density but high units, which should engage more of your units and therefore cause you more casualties.

You're probably right about that. Lots of my targets that have very low stats have only a few hundred units. However some of them manage to have garbage stats with 1600 or even 2000+ units. And those guys always net much higher casualties, regardless of their stats.

All 2000 of my units brought to battle are still engaging, however i guess the amount you have engaging vs the amount they have, is a big factor in the casualty rate. The ratio may be what matters. 3:1 is usually a very good rate for me.

Haven't really paid too much attention to this before. Wonder if there is a high stat player that i can beat with under 750 units. Then i could test the casualty on him. If this is true than that gives me another theory on what could be causing the casualty glitch.
For example, Maverick, has about 9850 units total. Perhaps instead of displaying it as 2000 vs 2000, it is giving him 9850 vs 2000 (although the stats of 7850 arent included). This would then give him a 5:1 unit ratio and a very low casualty rate.

[Freekizh]

Wonder if there is a high stat player that i can beat with under 750 units. Then i could test the casualty on him.

Yes that is another way to test the theory.


I've had my client crash countless times while attacking, before I can see the results. When I go back into the game, my inventory is updated according to the attack having transpired, despite my client having crashed.


Warfiend you could be right. I personally have not experienced any of those problems on a consistent basis when using wifi as I do not rely on my 3G connection.

Even during this holiday weekend perfectly fine - quite a few people have found quirks when having a bad connection. I find that the program itself doesn't send very much data to and from the server when roaming on my 3G. Some other programs I have use 1-2gb a day, which is many times the data use of this program.

[Q Raider]

Not sure if this is an indication that the server is doing work in the casualty cals or not.

I am often completing an attack/raid and reading the result which incudes a lost unit.

Scroll to the bottom and often there is no unit being displayed. THis weekend it is taking around five seconds for it to appear.

An additional quirk recently is if I lose a Valor unit then the game is crashing 95%-100% of the time.

[JMC]

Not sure if this is an indication that the server is doing work in the casualty cals or not.

I am often completing an attack/raid and reading the result which incudes a lost unit.

Scroll to the bottom and often there is no unit being displayed. THis weekend it is taking around five seconds for it to appear.

An additional quirk recently is if I lose a Valor unit then the game is crashing 95%-100% of the time.

After the update to version 2.0, if you scroll to the bottom of your list too quickly, the units dont have time to load. Scroll back up a bit and then down, and they will all be displayed.

[JMC]

Just attacked someone with 19500 defence with only 980 units. Got zero casualty on him 3 times and only ever lost 1 unit at a time other than that. So far the theory looks about right.

[Pace]

Popular forum logic was that you look for rivals with defence way lower than your offence in order to decrease casualty rate, well this below post from Crime City Mark might blow you away......

"Comparative attack and defense have no impact on casualties in any of our games. You could have a billion attack and they have 1 defense and you'd still have the same casualty rate.

Just because your glass cannons do a ton of damage, doesn't mean they're invincible."

My LLP attacked someOne with less than a tenth his A (2400 vs 200) and less than a tenth his units (around 240 vs 20): no casualties in seven straight attacks. Seems to me casualty rate is dependent on your army size vs his army size.

[Freekizh]

Just attacked someone with 19500 defence with only 980 units. Got zero casualty on him 3 times and only ever lost 1 unit at a time other than that. So far the theory looks about right.

I've been testing it on KA for the last week or so. I was just sick of all these damn casualty threads :p

[JMC]

Yep, i don't see CCM lying to us. The only thing that would make sense to determine casualties is army size, and so we can pretty much confirm that is how it works now.

[Freekizh]

Yep, i don't see CCM lying to us. The only thing that would make sense to determine casualties is army size, and so we can pretty much confirm that is how it works now.

Army size and casualty rate of the units as well, I would think - the larger the army size, the more pronounced the casualty loss. So it still kinda follows the basic maths, just with a few minor tweaks, like relative army size, etc. to take into account differences between defender and attacker.

Conceptually it all makes sense. CCM is sneaky but he wont lie.

[bobcnn]

Unit casualty rate is really useless.
I always lost more units of low and very low casualty rate than high casualty rate.

[Corporal Denney]

My observations lead me to believe the algorithm is different for raids verses attacks. If I attack someone I seem to lose from my meat shield, low value units mostly. But if I raid I lose high end units, to the point that I seldom earn more than the cost of replacements.

As far as attacking those with low stats, I attacked some redneck the other day who had the word nigger in his name (that just pissed me off), he had about 1000 a&d. I hit him 10 times, no money won but a lot of valor points that put me over the limit for two goals and I leveled up. I didn't lose a single unit. Then last night I hit someone who hit me a couple of weeks ago when I was weaker (a little revenge) and raided his high payout buildings. I pocketed about 2 mill and it cost me 2.5 mill to replace my units. Every raid cost three units and I was lucky if one was only a commando.