What happened to war matchmaking???

Printable View

Show 100 post(s) from this thread on one page

October 19th, 2016
nerfedname

Quote:

Originally Posted by ryryshouse

4) It doesn't matter what offense weighs, if you cannot see it.

Again, he claims you CAN see offensive weight, it's included in the gold storage amounts.

Quote:

Originally Posted by ryryshouse

explain that 20v I posted then please.

could be that hero levels don't count as much as lab research and camp space (did the other clan have maxed lab troops?). Also could just be a random mismatch in your favor which is still likely to happen from time to time.
October 19th, 2016
ryryshouse

Quote:

Originally Posted by cnaf

I would argue that offensive weight is not still very low, post update. Certainly don't have access behind the green curtain- but I've done extensive data compiling and crunching since this update came out, with the help of a variety of different types of clans running wars. Without a doubt, offense is playing a bigger role in the MM algorithm.

I also happen to think , the MM algorithm just sucks when looking at both offense and defense. Probably not apples to apples.

If my clan goes another 3-4 wars without seeing 20% engineered base on the map that would prove that out ( in my eyes ) - proving their algorithm just plain sucks.

They increase offense, we see engineers.

the decrease offense, we see trash.

It is fundamentally broken. They need to decile each base into a small/medium/large on a per TH basis and get a decent line fit.

So min TH11 weight is 88, mid is 100, max is 110
min TH10 is 70, mid is 78, max is 88
min th9 is 56, mid is 65, max is 70 and so on

stop the engineer, and combine offensive and defense into a single value so we CAN ALL SEE it. Then use whatever line fit they have.
October 19th, 2016
ryryshouse

Quote:

Originally Posted by nerfedname

Again, he claims you CAN see offensive weight, it's included in the gold storage amounts.

Yea, I feel you. Burden of proof goes to him, though.
October 19th, 2016
MrJackrabbit

To me the easiest way is to match vs resources received through raiding, war win bonus, mines and pumps, and gems. Those with the most resources collected should be a higher th level than those who don't. If you just waste resources like elixer while maxing walls at th8 that is your loss.
Any thoughts?
October 19th, 2016
ryryshouse

Quote:

Originally Posted by nerfedname

could be that hero levels don't count as much as lab research and camp space (did the other clan have maxed lab troops?). Also could just be a random mismatch in your favor which is still likely to happen from time to time.

No they did not have max troops.

Very consistent w/ match ups prior to adjustment. Have another one I will follow up with. Not as bad, but stacks basically the same way.

We are underweighted 2-6, way higher hero all the way down. These guy at least have teenagers....
October 19th, 2016
cnaf

Quote:

Originally Posted by ryryshouse

explain that 20v I posted then please.

Your post showed defensive base weights, and hero levels. No other offense shown aside from the hero levels. I've been using nightmare77's extensive spreadsheet, way too much time spent on it but tallied 37 wars from different clans since the update. In 31 of them, there was a clear cut advantage in defensive weight totals (especially top 5 or 10), and the same clan that held the advantage in this- was at a disadvantage on the complete offensive side. In 4 of the pairings- defense weights and total offense was nearly equal. In 2 of the pairings, defense weights were close, yet one side had a stronger total offensive advantage.

There are certainly other variables that come into MM pairings that can affect matches, variable that escape me. But I'm certain that I am in the neighborhood on my findings, and until otherwise debunked, I have a clear theory I am going with. These 37 matches came from clans with lot of different makeups. Rushers, maxxers, .5's and defenseless clans. And a defenseless clan with low hero levels.

Without knowing more specific details about your pairing, right off the cuff I find it interesting that basses 4-7 have a sizable defensive weight advantage...while the other clan holds the advantage in AQ and GW levels.

ETA- as stated, could be a plain bad pair. And also, as you mentioned...the MM could be rubbish. Won't argue there.
October 19th, 2016
ryryshouse

Quote:

Originally Posted by MrJackrabbit

To me the easiest way is to match vs resources received through raiding, war win bonus, mines and pumps, and gems. Those with the most resources collected should be a higher th level than those who don't. If you just waste resources like elixer while maxing walls at th8 that is your loss.
Any thoughts?

nah, dont think so man. dont see it being better then what they have. real solution is to publish how it works, so everyone can evaluate and provide feedback.
October 19th, 2016
MrJackrabbit

Quote:

Originally Posted by ryryshouse

nah, dont think so man. dont see it being better then what they have. real solution is to publish how it works, so everyone can evaluate and provide feedback.

true but they aren't going to do that. With total resources it would create a fair matchup for everyone. you posted no thoughts on how it would not be better.
October 19th, 2016
ryryshouse

Quote:

Originally Posted by cnaf

Your post showed defensive base weights, and hero levels. No other offense shown aside from the hero levels. I've been using nightmare77's extensive spreadsheet, way too much time spent on it but tallied 37 wars from different clans since the update. In 31 of them, there was a clear cut advantage in defensive weight totals (especially top 5 or 10), and the same clan that held the advantage in this- was at a disadvantage on the complete offensive side. In 4 of the pairings- defense weights and total offense was nearly equal. In 2 of the pairings, defense weights were close, yet one side had a stronger total offensive advantage.

There are certainly other variables that come into MM pairings that can affect matches, variable that escape me. But I'm certain that I am in the neighborhood on my findings, and until otherwise debunked, I have a clear theory I am going with. These 37 matches came from clans with lot of different makeups. Rushers, maxxers, .5's and defenseless clans. And a defenseless clan with low hero levels.

Without knowing more specific details about your pairing, right off the cuff I find it interesting that basses 4-7 have a sizable defensive weight advantage...while the other clan holds the advantage in AQ and GW levels.

ETA- as stated, could be a plain bad pair. And also, as you mentioned...the MM could be rubbish. Won't argue there.

Assume max offense for us, except the 9.5s and our bottom guy. ( so 15 max offense in the lab out of 20 ) All our war stuff maxed.

They had a few max, most were missing stuff. ( maybe 5 maxed offense , some had troops not unlocked yet! )

We have data for more wars then 37. You are better trending wars with the same make up then a bunch of one off samples.

My point is, they stated they tweaked a knob. It was to increase offense factoring in weight.

What we saw was equal defense matches, with other clans having offense advantages via engineering. TH11 in TH9 slots.

They turned it back, and the results are immediately very distinctly similar to matches previous.

the fix to me is obvious. Engineers need a weight adjustment that doesnt exist yet, with the current algorithm.
October 19th, 2016
ryryshouse

Quote:

Originally Posted by MrJackrabbit

true but they aren't going to do that. With total resources it would create a fair matchup for everyone. you posted no thoughts on how it would not be better.

I have already done that up thread...

Show 100 post(s) from this thread on one page