War Matchups & War System Issue

Maat1985

Hi.
I have posted before about war matchups using both war rating and alliance rating when selecting a matchup. Most people keep telling me I am wrong. But how it works up top is different as there is only a very small pool and you have to be the best to climb that high.
But not all of us want to climb that high. Most of us just want to be ranked properly based on our skill.

This is something I have been watching for sometime and have only recently started compiling the data. i have posted below data from the 10 most recent wars from both my 2nd account (officer in that ally) and my main account where i am leader.
Not only does it show all matchups are relatively close in alliance rating. It also shows that my weaker ally is able to climb to and sit higher in tier despite having much poorer performances.
The 8.5 mil ally in tier 11-14 is facing similar opponents and losing wars when we have 97% exploration and 3 boss kills
The 4.5 mil ally is facing similar opponents and can win wars with only 70% exploration and 1 boss kill sometimes due to weaker competition

So despite being in tier 8 wars for the weaker alliance are won and lost with only 75% exploration on average and mostly only 2 boss kills are required.
The wars for the stronger 8.5 mil ally are won and lost with 97% exploration and always require all 3 boss kills.

So despite an 8.5 mil Ally scoring 140k in tier 11 we lose
While a 4.5 mil all in tier 8 can score 100k and win

This makes no sense the alliance which is clearly stronger and better is restricted to a lower tier despite much better war performance due to the way matchups are working.

This has resulted in my 8.5 mil ally sitting in #80 gold 3 at this point while 4.5mil ally sits at #1250 gold2. We both do 3bg wars 3 times weekly but despite the lower war scores the extra wins and higher multiplier negate this and enable the lower rated ally to sit higher in both tiers and in season. And also enjoy being able to win wars whilst not having to do as much in the wars.

This is something i have seen for along time and only just compiled the data for now.

@DNA3000 i am tagging you because i know you wanted to see this information from one of our previous discussions.

Comments

Maat1985

(my ally rating is only approximite as it was not recorded exactly but is within a couple thousand actual rating)

(my ally rating is only approximite as it was not recorded exactly but It barely changed from start to end of season)

NoOnexRO

Very interesting. I am also leader in a 6 million alliance and with a small account officer in a 1 million alliance.
I can confirm the analysis.

Maat1985

@NoOnexRO Post any info and proof you have screenshots or such that will help bring this to light so others may actually se and believe this is the case.
Yes it stops weak allies getting steamrolled by strong allies but it also causes other issues. So we need to find another metric that can help give even fairer matchups then is current... some kind of balance...

Maat1985

Maybe it could be fixed from the season score standings if instead of a flat 50k win bonus the win bonus was tiered. 15k per boss clear and a 5k bonus for clearing all bosses. That way an ally that wins when they are good enough to clear only 1 boss does not get the same bonus as an ally who performs better and can win by clearing 3 bosses.
15k bonus 1 boss clear win
30k bonus 2 boss clear win
45k + 5k = 50k bonus 3 boss clear win

NoOnexRO

@Maat1985 I will try to return with a similar document you posted above.
It may take a day or two to compile all the data and I can only use the last 10 wars from the game's history.

Maat1985

@NoOnexRO when i get home this evening (around 12hrs from now) i will try and upload my template as a google doc. That way people like yourself and anyone else can make a copy and collect data. The formulas are in place. It will take two mins to fill in raw data.

DNA3000

Maat1985 wrote: »

@DNA3000 i am tagging you because i know you wanted to see this information from one of our previous discussions.

I've been watching our own wars this season. I cannot speak to the difficulty comparison because I don't have two alliances to compare, but I can confirm we as a ~11 million alliance are seeing a similar set of match up parameters. We get matched to alliances within about 10% of war rating and alliance rating essentially all of the time.

This does have potential consequences for how difficult different alliances' war match ups are when match ups are occurring in a dense environment (meaning: lots of alliances of similar war and alliance rating are seeking matches at the same time). When the game can segregate alliances by alliance rating, it seems to be doing so (at some tiers this might be impossible and you won't see this pattern, especially at the highest tiers).

At least in your case your analysis appears to be sound. How frequently this happens is impossible to say without more data, but it clearly shows to me that this scenario is very possible. In some tiers you can get an advantage by dumping alliance rating, which should never happen.

Maat1985

DNA3000 wrote: »

Maat1985 wrote: »

@DNA3000 i am tagging you because i know you wanted to see this information from one of our previous discussions.

I've been watching our own wars this season. I cannot speak to the difficulty comparison because I don't have two alliances to compare, but I can confirm we as a ~11 million alliance are seeing a similar set of match up parameters. We get matched to alliances within about 10% of war rating and alliance rating essentially all of the time.

This does have potential consequences for how difficult different alliances' war match ups are when match ups are occurring in a dense environment (meaning: lots of alliances of similar war and alliance rating are seeking matches at the same time). When the game can segregate alliances by alliance rating, it seems to be doing so (at some tiers this might be impossible and you won't see this pattern, especially at the highest tiers).

At least in your case your analysis appears to be sound. How frequently this happens is impossible to say without more data, but it clearly shows to me that this scenario is very possible. In some tiers you can get an advantage by dumping alliance rating, which should never happen.

I believe at my low tier there is a large advantage by dumping ally rating. Which i dont want to do cusni dont want to sell champs.

And yes my hope with this thread is to be able to collect more data from more people.

Maat1985

@DNA3000 one question is if my smaller ally had another couple boss killers without increasing the ally rating much how many more wins would we be getting?? How high tier would we get too? How much would this increase our season score?
Just how high would a 5 mil or lower ally be able to get their war rating before this begins to become a non issue....
But this is really bad cus if everyone in my stronger ally sold our 3* champs we would drop to a lower rating and then get matchups like my weaker one...

I feel this is becomming a bigger issue as more ex pros retire from top tier and join weaker low tier allies... (i have some in both my allies).
Also is a bigger issue with war seasons than it was before.
I am not to happy woth this as my second account playing badly in wars is looking to get better rewards than my main who is busting thier a$$ every single war.

DNA3000

Maat1985 wrote: »

@DNA3000 one question is if my smaller ally had another couple boss killers without increasing the ally rating much how many more wins would we be getting?? How high tier would we get too? How much would this increase our season score?

That's likely to be highly variable. If I had access to all of the AW data Kabam had, I could answer a more general question which is: for a given alliance rating, what is the range of war rating for which there was some minimum number of alliances. That range would be the range within which there would be a theoretical advantage to rating dumping.

If that isn't obvious, go back to your hand sketches of the problem from season one. At the top and bottom of your "silos" of alliances they should get very pointy as the number thins out. When the number of alliances way up at the top drops to a tiny amount, you won't be able to match with "similar alliances" any more because there aren't any (looking at the same time as you). At that point the game will be forced to look for matches with alliances with different rating, and your ability to "hide" inside a low rating will vanish.

The difference between the center of your silo and that point near the top is a general answer to your question: how high can this strategy go? But it would take tons of data to answer that question.

unknown

This content has been removed.

Maat1985

DNA3000 wrote: »

Maat1985 wrote: »

@DNA3000 one question is if my smaller ally had another couple boss killers without increasing the ally rating much how many more wins would we be getting?? How high tier would we get too? How much would this increase our season score?

That's likely to be highly variable. If I had access to all of the AW data Kabam had, I could answer a more general question which is: for a given alliance rating, what is the range of war rating for which there was some minimum number of alliances. That range would be the range within which there would be a theoretical advantage to rating dumping.

If that isn't obvious, go back to your hand sketches of the problem from season one. At the top and bottom of your "silos" of alliances they should get very pointy as the number thins out. When the number of alliances way up at the top drops to a tiny amount, you won't be able to match with "similar alliances" any more because there aren't any (looking at the same time as you). At that point the game will be forced to look for matches with alliances with different rating, and your ability to "hide" inside a low rating will vanish.

The difference between the center of your silo and that point near the top is a general answer to your question: how high can this strategy go? But it would take tons of data to answer that question.

Yeah i understand that and it was more a theoretical question as i know it would be one that would be hard to answer. Cus 4.5 mil ally will begin to get matched up 7 mil then 8mil then 9 mil increasingly as it got higher. I feel this would get increasingly faster cus just as each bubble of alliance narrow at the top so does the entire pool.

Maat1985

reposting a comment i just had with someone in another thread cus it is extremely relevant to this topic.

Maat1985

To stop rating manipulation. Instead of total ally rating.
How about a rating metric similar to prestige.
But using a players top 15 champs. As this would give a much better metric of strength in war.
Rating is bad cus some have low rating but strong top champs. Or high rating but lower top champs.
Prestige is bad cus it only calculates on top 5 so if u have 5 strong champs but no depth it is useless for war.

Top 15 champs give a much more accurate indicator of strength in war as 8 alone need to be used in war. And then 3 for aq and 5 for solo so that makes 16 champs peeps need to keep as strong champs

NoOnexRO

I also thought at a similar solution @Maat1985 - even though I wanted to consider the first 8 champions for the metric.

But what if we take this one step further?
What if the system of match-ups were based on what players join the war?
At this moment you start to match-up against another team. The system gets an enemy based on the metric and from that point on it's up to your team members to join the war.

What if we open the BGs? We allow members to join and only after that period ends the team will get an enemy?
The enemy will be based on who joined the war battlegrounds and not on the alliance members that might not even be on the war.

Maat1985

@NoOnexRO
Interesting idea....
Start the war....
Place defenders....
Find opponent....
Could be interesting.....
Then you get a matchup based on the rating of your defence.....

Possible exploit.... placing a weak defence to get matched up against a weak opponent....
Could be part of a solution but maybe not the whole solution.

NoOnexRO

The match-up will not be based on the defense ratings.
The match-up will be based on the metric calculated based on the PI of members that joined the war.
And, as you suggested, not all the heroes would count only first 8 or first 10.

Maat1985

@NoOnexRO
you may have an idea there.....

also here is a link to excel online for my file.....

https://1drv.ms/x/s!AlkQtWTl4kUFknbyhAEhL2h1poNh

you cant edit. will need to save your own copy..... just enter in ally ratings, tier, points, bosskills and exploration.
my formulas will do the rest.....
(grey highlighted rows and average column)

Maat1985

just came across this and it shows that the manipulation of this is already being seen to happen.

NoOnexRO

Thank you @Maat1985
I have entered the data but... and there is a biiiig but... the alliances ratings are not the ones they were when the war took place. I took the rating now so my data (for the small alliance at least) will not back your findings. We didn't changed the rating much, the enemies did - had increases and some rather big. So, sorry for that.

I will use the document from now on and save the data the moment we have the war, not 5-10 days after.

https://goo.gl/mVK1Rv

NoOnexRO

And a small off-topic:
It would be interesting if wars would be as arenas are: limited by heroes.

A war with 2* only
A war with 3* only
A war with 4* only
A war with 5* only

unknown

This content has been removed.

Maat1985

@NoOnexRO all good. very helpful data.

it makes sense to see you are lower tier that 2 boss kills will decide the win or loss......
seeing you in the low tier there facing similar matched opponents whilst my smaller ally fights similar matched opponents in a much higher tier really highlights the ally rating being a calculating factor.
what intrest me is how it seems to also show difficulty progression based on ally size more so then just tier.....
8.5 mill ally, tier 12 needs 3 boss kills and >95% exploration to win most wars
6.5 mil ally, tier 15 needs only 2 boss kills and maybe around 80%+ exploration
whilst a
4.5 mil ally tier 8 needs only 2 boss kills and maybe around 80% exploration.....

looks like above maybe 8mil rating things begin to get much tougher.......

unknown

This content has been removed.

Maat1985

@Bajjsbwuqjw
we are on the cusp of gold 2 and gold 3 in my 8.5mil ally. between #150 gold 3 - #1300 gold 2 and as you see we need 95%+ to win.....
we are around #1200 gold 2 in my 4.5 mil ally and we need only 70-80% not even all bosses down.

as i said before both my allies are 3 wars a week and my 4.5 mil ally who does less to get its wins is looking to get a higher season reward.
we are really trying to push the 8.5 into gold 2 now.....
so alot of allies in gold 3 do get atleast almost 100%.

unknown

This content has been removed.

The1_NuclearOnion

@Maat1985 as requested sir:

Rating: 6,767,864
WR: 1,397
Tier 7 multiplier x3.2

Currently rank 372 Silver1 bracket

Maat1985

I wonder...
I wonder if there is a substantial difference in the skill level of the allinces you face running 3 wars per week as opposed to only 2 wars a week....
This would require my data to be collected from same same rated allies doing 2 amd 3 wars a week.

The reason i am thinking like this is i assume any alliance who wants to get the best possible season rewards will be doing 3 wars. So are those allies pushing for better rewards on average better allies??? I dunno.....
War runs for 6 days.....
To do 3 wars u need to matchmake days 1,3,5
If you matchmake for only 2 wars days 2,4 you will face only other allies happy to do 2 wars.

The1_NuclearOnion

We're running two BG regularly. A solid 2 bg is much easier to manage than 3.
3 bgs require ALL members to be 100% active.
2 bgs only requires 20 out of 30 to be active regularly allowing a laid back atmosphere and some flexibility for life.