Alliance Wars Matchmaking (Discussion & Developer Response)


#1182

■■■■, I’m ready for just a simple ladder system based on win-loss ratio, pair with similar ratios. Would take some time to start even out, but… it would work (and I would like to think devals could dig past match results from somewhere… unless their game DB is really messed up design.)


#1183

The best matching algorithm won’t solve the problem, that alliance members get frustrated by war and lose interest. In fact the opposite is happening. A better matching increases the possibility, that alliance members lose their interest, when many wars are lost (a downward spiral is happening). With a matching algorithm, which randomly results in a greater advantage for an alliance, the motivation to stay active in war is maintained more easily.
So i would suggest to separate war and alliance. Offer war as a separate challenge. People can sign in and paticipate to it and are grouped randomly with other players. Advantage is, that those poeple are motivated to fight. When they are not, they won’t sign in for the challenge.
The individual opt out button is a good but may be half way solution for the “downward-spiral-motivation-problem” of an alliance. So make war a separate challenge.


#1184

First draw ?

Le bastion tranquille vs 125rus : final score = 2707 vs 2707 :slight_smile:


#1185

They don’t need an algorithm, they don’t need some convoluted, complicated matchup equation system. Just pair Alliances up within 25 to 50 ranking above or below them. Period. You will get more better matchups than worse. And most importantly, DO NOT ANNOUCE THAT YOU ARE DOING THIS. People cannot manipulate a system if they don’t know what that system is. If the matchups are good, no one will care that they don’t know how they are made.


#1186

That’s pretty close to what I suggested a while back, except I think alliance ranking could totally replace the existing alliance score. No need to change it in secret. Teams would settle into their proper place in the rankings over a short time, just as they do in raids.


#1187

Could not agree more. I’ve seen the alliances around us in terms of total score. Our alliances are very similar.


#1188

So not surprising:


Seems relatively close?
We attacked 155 times.
They attacked 90.
They had at least 40 kills. Under the old points system that would be 600 more points- a complete blowout.
Yes we had numerous attacks go for less than 5 points, but with this matchup at some point it devolves into us attacking 3000+ tp teams with our levelled 2* and Unlevelled 3s. We had absolutely no chance in this war (again)


#1189

Just an update. We lost our last war by a ‘mere’ 50%. 2k verse 3k. PS they only did 2 attacks in the last 7 hours. Sooooooo given the mere 50% final result does that result still rank as a fair match? Even with the reduced points scheme? Just asking. Will a represent of SG pls respond to my post, directly?
Thanks


#1190

We are still being overmatched. This time by only 14000. Why can’t we be the favorite once?


#1191

as a member of a “training alliance” I feel that we are now never going to win a single war.
why ?
because the matches are being done by alliance scores and titan scores and doesnt factor that many players have been & gone from our alliance helping to… over time accumulate those scores.

As our players get stronger they move up to allied alliances to fit better with their gained levels .
This means that the titan scores & alliance general score does not reflect the groups true strength more a past glories strength.

Sadly new players come in & any hoping to get a fair AW fight with us never will as we are massively over matched.
The only way we could be matched fairly now would be by taking the mean score of hero’s within the players set ups .
I am lead to believe a 4* hero fully levelled up scores near the same as an levelled 5* which is madness
Find a way to score 3*/4*/5* heros based on how much the are levelled so that eg a 5* would score 5 & a levelled 5* scored 5x70 or what ever the number of full levels there is for a 5*

Half of our alliance would struggle to field 2 teams yet alone 6 but our titan/team scores miss match them.

Something is skewed badly wrong when I can score a max of 50 points against a player of 3400 and the same player can score 70+ against me of similar strength …

Hardly a shock why so many are starting to opt out


#1192

7 wars, 7 mismatches, 7 no chance to win. Courtesy prevents me from using every four letter word in the English language to describe my contempt for your incompetence. You have to fix this.


#1193

@drenai the Alliance War scores USED to be calculated by alliance/Titan scores. That has since been changed.

The calculations used now are roughly: 1. The sum of the top 30 heroes for each eligible alliance member. 2. A heavier weight applied to the top 5 heroes (the assumption is that these will likely be the defense team).

Since alliance and Titan scores are no longer used, the hard work of past members should not impact your current Alliance War score. It is my understanding that these scores are recalculated before matchups are made. This means the Alliance War score assigned to each alliance is based on current members only.

This sounds like your alliance may have less members than the matched alliance (although I think the variance should only be ±2…or is it 1…?). Also, the total field points are divided equally between all players based on their defense team scores. This means, if you have a team of 3,400 but everyone else has 2k teams, your team would be worth significantly more. By the same reasoning, if your opponents have teams of more the same level, the points would be spread more evenly.

Hope this helps with some of your frustrations :slightly_smiling_face: You definitely have some hurtles to overcome with the large differences in team sizes.


#1194

since AW has been tinkered with our alliance has always been over matched by 10-20k on alliance score alone , we now have some members that opt out as they are unwinnable. we have 3 players with +3k defence scores and usually matched against alliances with 10+ over 3k & the remainder of enemy in the 2500+ range , the bulk of our alliance fail to reach 2800.

All our scoring is lower even when players hit the same attacker , without fail they score lower , I dont understand the permutations just see the result . The closest reslut we have ever had since the implemented changes was a loss by 1500 , usually they are 2-3k losses.

Some of our group play AW simply as there is a minor materials rewards but … tbh i cant see the point.
as it is now we have no chance even of a minor win & players are either leaving the group or abstaining as the enjoyment has been sucked from aw.
even the most basic ladder league set up would slowly sort this out with the chance eventually of being matched against a similar alliance that gets zero win rates since the changes.


#1195

This. SG has a event ladder for raids. Why can’t they just port the code over for wars?


#1196

Is this what the upcoming elo-based matching system will do? Never dealt with one so a little vague on all that is involved :slightly_smiling_face:


#1197

Yes, generally speaking. A proper ELO system is more complex, but it’s the same general idea.


#1198

1.14 is here and still no individual opt out button. I have several vets wanting to leave the alliance bc they dont want to war. when is this coming


I found the flaw in AW matchmaking system and it is probably not what you would expect
#1199

That feature was actually postponed to 1.15. My apologies for the delay. The current plan is that you will be able to opt-out from the wars simply by removing your defense team.


Alliance Wars Opt Out
Whatever happened to individual AW opt out?
#1200

I’m going to have to keep hitting the like button for this one mhalttu! :star_struck:


#1201

Will there be any other war/matchmaking improvements in 1.15? I’ve probably plugged my two suggestions already a few times, but I think these are both fairly easy to implement and will make for much more competitive battles / fairer matches:

  1. Fix team power (or use a fixed team power alternative) which doesn’t overstate the abilities of heroes with higher stars that aren’t fully ascended. (5* 3/70’s and 2/60’s simply aren’t as strong as their team power would indicate when compared with lower star heroes, etc.) Team power is misleading across star levels
  2. Adjust the points assigned per team so the 1000 points available are divided up by the number of attacking teams instead of the number of defending teams. AW Scoring problem (and fix!)