Alliance Wars Matchmaking (Discussion & Developer Response)


#208

A war against a similar alliance we lost by 3 points.

So we kicked all 30d no show teammates.

Or second war was:

3969 Twilights Bastion
To
2317 yPA?
? Character appears to be a curve up, straight top, straight down


#209

Thanks. 7/8 of the opposing tee

am were more powerful than our top two players. Photos of match up attached.


#210

Hmm… Inactivity is quite bad with current matching system, that is clear. But one thing:

Main reason could be the way team power is calculated. Perhaps some kind of simulation could be used to make numbers better. Let’s take one example:
Defense team: tp 4100, containing 5 5* levelled up to max.
Two attack teams trying to take it down: 2x tp 2050 containing 3* heroes. I honestly doubt that they can do a thing against tp 4100 defence team… Perhaps they may get a point or two… So team power does not add up well…

Perhaps changing algorithm how power of each hero is calculated? It was enough that teams having similar team power are more or less equivalent in performance, but now… There are additional requirements, could team power calculation be improved so that two tp 2050 teams is somewhat equal to one tp 4100 team?


#211

@mhalttu - In the beginning of AWs, our team was matched pretty closely to the other teams. Now the other teams are much more powerful. We have been slaughtered. Team power should not be used as a matchup as the power score doesn’t mean a lot in that respect. It’s how you use your heroes together. A lot of games that are like this rank the heroes based on a lot of different aspects including how well they work with the others on your team. Titan score is probably your best predictor along with he average level score of the team. Trophy count doesn’t mean diddly because some purposely keep theirs low or they might have been attacked a bunch in one night (one member was attacked 40 times in one night).


#212

Ok in all honesty the top 5 heroes is all that should be raken into account or have greatest wieght put on. Seems that us that have worked on having 6 teams are being severely penalized. We have had 2 wars since last update and in 1 enemy had a guy with all maxxed 5 star heroes and we dont have a single member with any 5 star heroes that has even started last acsension tier and they were better from top to bottom. Most recent was even worse. Enemies lowest player had a team power of 2200 with rest of alliance having a score over 2400. We have 6 members who have a strongest team power score below 1900 with one at like 1540. How in the world are we even suspose to compete?


#213

Another status update.

I checked the results of the second war, and again on average it was more even than the previous wars. Please note that I’m only talking about averages here - I understand that your alliance may have gotten a very unfair opponent.

As Gryphonknight pointed out, we did make a bunch of other changes to Alliance Wars in the previous update. We’ll run the next war using Titan Score and check the results. At the same time, we are hard at work on a solution that puts more weight on the best five heroes, and possibly also takes into account the nonlinear nature of team power.


Analysis obsolete war strategy ARCHIVE for reference only
Consider this, DEVs, you're missing a vital info regarding AW matchmaking
#214

Wouldn’t it be nice to get some sort of acknowledgment from SG and the developers that they have made mistakes in the new formula for the matching system. An apology wouldn’t go a miss but alas no here at SG money and not the customer is at the heart of everything they do.


#215

Why not let the leader pick a war?
So you can have a 1 defensive and 1 offensive war.
Each war with 6 energy of course.
That way nobody can complain.
Use a leaderboard to determine rewards.


#216

The non-linearity of the power of the heroes is undoubtedly a big problem in that system, I think they should also give priority and correct the way in which the heroes are currently pondered


#217

Didn’t work…this time our opponent is 30k less than our alliance in titan score… Ugh!!! Another slaughter, only this time we’ll win.


#218

Why not have a new alliance war score that’s based on past battles?


#219

Please undo this!

My alliance (mostly beginners or people who have played for max 3-4 month) just got paired with pretty good players. In fact, their best 12 players are all better than our 2. best player. Our 7 worst players are all worse than their worst player.
They have 71200 trophies and we have 56500. But we get paired with them because they don’t do titans (or haven’t done a few titans for a week, so their score got lowered or…?)
This really seems unfair and after being beaten pretty bad for the last 2 rounds (and there is no way we can survive the next attact) I fear people are going to quit because honestly, this is just demotivativing :frowning:


#220

We are making the match based on the titan score, not the alliance score which includes trophies.


#221

Okay, yes matched by “titan” score worked, we are similar in “titan” score. However, we are 30k stronger in “alliance” score which uses trophies and titan combined…sorry for confusing that.

That said, that’s why using titan score doesn’t work either and why an update was necessary to begin with - there was a need to fix. The update didn’t fix and it some cases, as outlined in these forums, it made it worse. I understand from your post that on average, it got better.

So, again - as has been suggested by many - why not develop and AW score. Until then, using the full alliance score will probably set up the best match-ups.


#222

It is not that simple. People change alliances all the time, and that kind of a score reacts very slowly. A new alliance may be created by top players, and they’d have to first slaughter at least ten alliances before their score would match their true strength.


#223

Are you saying, you scrapped top-30 heroes system?


#224

Not permanently. We are working on a new version that gives more weight to the top five heroes, which are most likely to be used as the defense team.


#225

I was responding to myself when you wrote this realizing my mistake too.

That said, you can clearly look at our match-up and see that it will be a very unfair match-up for our opponent. We will win and it won’t be fun. It certainly won’t be fun for the other side. They will go into it knowing they don’t have a chance.

I’m not convinced of the 5 strongest hero rating system with weighting them differently. I get it’s not simple otherwise we wouldn’t be having this discussion, but it’s not super complicated either. An alliance war score that takes into account multiple factors, which could include past battles, seems to be the best approach.


#226

Perhaps you could also consider giving less weight to 5 star pure defence hero’s as f.i. Owl who if there are enough attacking 4 star hero’s will never be used for war. Off course same goes for 4 star defence hero’s in that matter.


#227

Same with titan score, that did not work for my alliance, and members come and go regularly. Thanks for the reply.