GBG Nerf?

[* Cass *]

Seems like everyone is complaining about increased attrition and damage. Loadouts that generally might have been 4 attrition are now 16+, easy, regularly. Used to get through a full day with maybe 10 attrition at burnout and am now much closer to being maxed out, even with fewer hits! It isn’t just a case of bad luck and bad odds. This is a regular issue. With 0 attrition and fresh troops, they come back with major damage, if not losses. Guild members and friends across guilds are talking about how they cannot make it to the same attrition levels that they used to easily make it to. Something seems way off this season. Any others?

 

[RazorbackPirate]

Sounds like you're blind battling. No way to tell what's going on.

 

[Emberguard]

Go into a manual battle, press Auto. See what the computer is doing during the battle. It could be something as simple as the terrain randomising into something that didn't suit your troops this time round.

 

[* Cass *]

It is many people commenting on it. Not just me. And every round, every day. Maybe really bad luck but not the whole season. Yes,I am autobattling but something changed. Attrition of 20 on 4 SC sectors means that I should be averaging 500 fights for the round. this is with maybe 200.

 

[wolfhoundtoo]

She obviously means more than 1 sector assuming that is why you think she's drunk.

 

[Emberguard]

It is many people commenting on it. Not just me. And every round, every day. Maybe really bad luck but not the whole season. Yes,I am autobattling but something changed. Attrition of 20 on 4 SC sectors means that I should be averaging 500 fights for the round. this is with maybe 200.

I get you’re not the only person saying that, you mentioned it in your original post. However if you’re not willing to go into the battle and see how it plays out then there’s no info to go off of on why you’re getting that result. Even Manually fighting the terrain can make the difference between whether all your troops die or you sustain no damage for the entire fight.

 

[My Push Account]

I think she is referring to attrition growth, not troop loss. Four siege camps offers 96% attrition protection, so she is expecting four attrition per 100 fights, resulting in it taking 500 fights before she reaches 20 attrition. The sample size of 200 seems a bit informal, but if correct, would suggest that the attrition growth of four siege camps is happening at a rate between 6.57% and 14.94% which is beyond the 4% posted values.

 

[icarusethan]

I think she is referring to attrition growth, not troop loss. Four siege camps offers 96% attrition protection, so she is expecting four attrition per 100 fights, resulting in it taking 500 fights before she reaches 20 attrition. The sample size of 200 seems a bit informal, but if correct, would suggest that the siege camps are operating between 6.57% and 14.94% which is beyond the 4% posted values.

unless we have a huge sample size(90 days 100 people over 40k attempts or more ), talking about odds change is just ridiculous.

 

[* Cass *]

Agreed. I am all for stats. What I am getting at, how ever informal this is, is that many players are noticing a much different play in terms of attrition. If you aren’t, then congratulations. This whole season has been uniquely different. I posted this question to see if others here would agree or not. Evidently, those of you responding are not.

 

[Emberguard]

I haven’t been doing enough this season to notice a difference

 

[CommanderCool1234]

A popular foe youtuber made a video a while back where he said that after his several thousand battles (not sure of the exact # of battles) he said that 1,2, and 3 siege camps seemed to be pretty accurate in how often they gave you attrition but 4 SC sectors seemed to be closer to 85 or 90% chance of not getting attrition.

 

[My Push Account]

unless we have a huge sample size(90 days 100 people over 40k attempts or more ), talking about odds change is just ridiculous.

No, that is not how statistics works. You assess the data you have and the mechanics of the math account for the sample size by narrowing or broadening the width of the confidence interval which determines what we can confidently state about the data. In fact, a single sector of GBG can be the first indication that something is not working properly. For a four camp sector, if the collective attrition is more than the following values, there is very likely something wrong or not as published:

Diamond >12, Platinum > 10, Gold > 8, Silver >6, Copper >4

All of this would support Cassandra’s observations of it not working properly, if the claimed attrition and sample size are accurate.

Your proposed sampling is not a great design. You would not experiment on something over a 90-day period that can be or will be changed during the data collection. It is clear now that Inno Games is changing at least some things without communication so we might not even know about it if they did, which would cap sampling to about 2 weeks between updates. Additionally, the mechanics and probabilities of the game do not change from one day to the next so a sample collected all in one day is no different than those collected over two or more, if there are no changes being made to the mechanics. The number of people does not matter, as this is not a representative population study and the posted probabilities are supposed to work the same for everyone. There is nothing special about 40,000 attempts. Even if the experiment results in precisely 1600 attrition, it has narrowed the confidence window down to .4% width, a modest improvement from the .5% width achieved at 25,000 attempts or even the 1% width achieved with 5,000 attempts.

 

[My Push Account]

Agreed. I am all for stats. What I am getting at, how ever informal this is, is that many players are noticing a much different play in terms of attrition. If you aren’t, then congratulations. This whole season has been uniquely different. I posted this question to see if others here would agree or not. Evidently, those of you responding are not.

Have you noticed any difference in attrition growth based on how many sectors are being used to provide the four camps? Basically, there are four ways to get to four camps supporting against a sector:
3 Camps + 1 Camp in 2 sectors
2 Camps + 2 Camps in 2 sectors
2 Camps + 1 Camp + 1 Camp in 3 sectors
1 Camp + 1 Camp + 1 Camp + 1 Camp in 4 sectors

It would be interesting if the published rates become more divergent with having to leverage more sectors, or vice versa.

 

[RazorbackPirate]

No, that is not how statistics works. You assess the data you have and the mechanics of the math account for the sample size by narrowing or broadening the width of the confidence interval which determines what we can confidently state about the data. In fact, a single sector of GBG can be the first indication that something is not working properly. For a four camp sector, if the collective attrition is more than the following values, there is very likely something wrong or not as published:

Diamond >12, Platinum > 10, Gold > 8, Silver >6, Copper >4

All of this would support Cassandra’s observations of it not working properly, if the claimed attrition and sample size are accurate.

Your proposed sampling is not a great design. You would not experiment on something over a 90-day period that can be or will be changed during the data collection. It is clear now that Inno Games is changing at least some things without communication so we might not even know about it if they did, which would cap sampling to about 2 weeks between updates. Additionally, the mechanics and probabilities of the game do not change from one day to the next so a sample collected all in one day is no different than those collected over two or more, if there are no changes being made to the mechanics. The number of people does not matter, as this is not a representative population study and the posted probabilities are supposed to work the same for everyone. There is nothing special about 40,000 attempts. Even if the experiment results in precisely 1600 attrition, it has narrowed the confidence window down to .4% width, a modest improvement from the .5% width achieved at 25,000 attempts or even the 1% width achieved with 5,000 attempts.

That's a whole lot of words to support nothing. Instead of lecturing us on how to measure, why don't you measure, Mr. Measure Man?

Always wanting to accuse and indict Inno of wrongdoing, based on rumor and innuendo. Nothing more than a sampling based on confirmation bias. But sure... blah, blah, blah, no point in measuring now, Inno will just change it. C'mon, man!

The bovine excrement is deep with this one.

 

[My Push Account]

That's a whole lot of words to support nothing. Instead of lecturing us on how to measure, why don't you measure, Mr. Measure Man?

Always wanting to accuse and indict Inno of wrongdoing, based on rumor and innuendo. Nothing more than a sampling based on confirmation bias. But sure... blah, blah, blah, no point in measuring now, Inno will just change it. C'mon, man!

The bovine excrement is deep with this one.

Checking the scoreboard, I have commented on five sets of data submitted by other users on here. In four of those cases, I stated the the data supported what has been posted by Inno Games. This would be the first data provided by anyone else here that I have said supports something not working correctly, with the caveat that the sample and attrition was as reported. That seems like a far stretch to say “Always wanting to accuse and indict Inno...” and would also invalidate your claim of “confirmation bias”. I am not interpreting the data in anyway, simply taking the data provided here putting it into formulas that have been long established and posting the results. You, and anyone else, are free to do the same, but so far the only response to any posted data is roughly “you need a larger sample size, tens of thousands” without even knowing what the initial data suggested.

You have yet to refute anything that I have posted, but like others, you seem to be more focused on who I am rather than what I have actually posted. Might be your own form of confirmation bias?