Yes, wars are unfiar! Help to analyse wars, to make a solution here

General Discussion
1

Hi all. Due to a lot of complaints coming from players about unfair matchmaking in alliance wars, I decided to make a short analysis of it. SG is still trying to make wars more justice, but I don’t think that has helped for the system.
Ok, but what is wrong with matchmaking? Is any war possible to win or sometimes wars are impossible to win with an enemy? Check out there.
To win war, alliance must gain more points - and this is first thing to check.
I am basing on a wars of my alliance and this is what I found out:

1. Additional point for a win batle are a half of basic points, additionally rounded up. It means, that tere’s completely no matter how strong player attacks the enemy - if You win, you get additional half of basic points.

  1. Both alliances have the same sum of points in total. To explain it, there is a table with players strenght and their total points.
    obraz
    obraz781×347 8 KB

    Both alliances are worth the same amount of points (1517). And there is a very important matter to ask - is this justice?
    NO!
    Check this out:
    1st alliance:
  • 3 players ~4000+, 2 players ~3700, 1 player ~3600, 1 player ~3500, rest - worse
    2nd alliance:
  • 4 players ~4000+, 2 players ~3700, 1 player ~3600, 2 players ~3500, rest - worse

That seems first alliance have harder way to destroy whole enemy alliance to get the same amount of points, because enemy has more players with 4k+ strenght. It’s completely unfair.

From another war, there is a comparision between my and enemy alliance:

1st

  • 3 players ~4000+, 2 players ~3700, 2 players ~3600, 2 player ~3500, rest - worse
    2nd alliance:
  • 7 players ~4000+, 2 players ~3700, 1 player ~3600, 2 players ~3500, rest - worse

Just explain me - how first alliance have to destroy 4 more players 4k+ and get the same amounts of points, if enemy kills weaker players and get the same? Completely no logic… Points should depend on defense teams not on overall score for team.

3. Points are not proportional to defense team. Just look for example from screen:

obraz
obraz811×374 8.64 KB

Why player with 4000 defense team is less worth than player with 3700 defense team?

And another example from last war. Two enemies:

obraz
obraz2160×2160 1.54 MB

Why player with 3700 team is less worth than 3400?
If it’s harder to kill, should be worth more!

To have a clear proof for unjustice pointing system I made some charts, basing on a table posted upper.
They presents dependence between defense strenght and points possible to get for win a battle.
1st alliance:

obraz
obraz959×627 7.68 KB

There we can see that there’s no matter in current system how strong defense team is. Some stronger teams give less points than worse ones.

The same chart with second alliance:

obraz
obraz956×629 7.24 KB

And there is a collection of both charts in one:

obraz
obraz1033×625 17.4 KB

The conclusion is that points aren’t proportional to defense team, what makes next conclusion - not every war is possible to win.

Dear SG, the most honorous fight is when players can get better rewards for killing better players…

Let’s all discuss - if points would be proprtional to defense team - it will be a solution?

2 Likes
2

Points are based on total HP of the team.
While they are related to team power, it’s only due to better team usually (only usually!) having higher hp.

12 Likes
3

And your top 30 heroes too. Without knowing the full roster of your opponent the data couldn’t be properly charted.

7 Likes
4

Ok, but what If I have the same total 30 heroes strenght but oponent has stronger ones to make better defense team?

5

very nice job but i thought the matchmaking is based on the 1st 6 heroes of each players, that’s what she said…:slight_smile:

6

The bulk of the calculation is the top 30 heroes in your roster, a small portion is your def team.
Logic,
You have 5 maxed 5* and 25 Maxed 2*. Thats it, thats all.
or
You have 30 maxed 5*

Your offence opportunity with Logic 1 is very very small in the average war. You’ll probably get 100 Points max on a standard 30 V 30 match.

Your offence opportunity with Logic 2 is huge. Probably 6/6 max flags and 300 points.

The defensive points have to be biased toward offensive opportunity as well to make the matches as fair as possible. So just because you see a 4300 point team doesn’t mean they have as much offensive strength as the guy with a 3800 point team and 30 maxed out 4 and 5*'s.

7

Matchmaking

1:1 matchmaking is an insolvable problem. Devs target “good enough”.

But you could make some changes to war.

Changes

You could add some buffers to spread out some of the concentrations of imbalance that make matchmaking even more difficult than a 1:1 restriction.

6 attack teams AND 6 defense team
This would buffer the problems with revive, HotM defense, and the ( Fabian* strategy for war )

Wars last for 6 days
This would buffer the RNG problem. Everything except war energy, and points, resets every 24 hous.

Sacrifice play
Every 24 hours you can donate 1 war energy to any teammate. Removing 1 from your war energy bar and giving them an additional 1 war energy. Maximum +3 war energy per 24 hours. This would buffer the uneven teammate roster problem.

Wanted Hero Mission Chest
Each killed hero in war counts for the Wanted Hero Mission Chest. This would buffer the war loot problem.

World Energy
Every attack in war, rewards +3 world/ map energy. This would buffer all the above problems.

“Good enough”

But the Devs have moved on. They have a ton of other stuff to work on before they can get to trade which is a huge deal now that Pokémon GO broke the road block to effective MMO trading.

3 Likes
8

First time I hear this… How this operation can be done?

1 Like
9

Those are just improvements that Gryphonknight is suggesting for wars.

It would be really cool if that was possible, but I think the point of his post is that the devs consider wars to be “good enough” when it comes to matchmaking.

I agree overall with OP that the matchmaking can be skewed in several instances, but because it’s not horribly broken (a while back it was very bad) they likely consider further work on it to have little return. The player in me disagrees, but the developer in me agrees :unamused:

3 Likes
10

Lol :rofl: I already told my co-leader next time to give me a flag as soon as I find out how this can be done :joy:

1 Like
11

War is still a zero sum game though. For every inferior Alliance in a matchup, there is a superior Alliance. You may be at a disadvantage today and in a superior position another day.

IMHO, the main area for improvement is that Wars don’t consider bench depth. I run an Alliance of mostly newbies and while we have a few members with two or three 3000 TP teams, everyone else has a single low 2000 team. We regularly get matched up with opponents who appear to have multiple 3000 teams given the observed performance using flags 4, 5, and 6.

1 Like
12

THIS is an excellent idea. It would have my vote immediately…immediately after I removed one from another post

1 Like
13

The war score supposedly considers bench depth. It is claimed that each player’s top 30 heroes count, with the top 5 counting more.

Also, your top 5 troops count in the matchmaking.

I think there is a bug in there somewhere (or an ecploit) because we have faced many alliances with vastly superior defenses as well as offenses equal to or greater than our own. I can’t remember the last time we faced an alliance that did not have many more high defense teams than us.

1 Like
14

Here are my thoughts.

Matchmaking doesn’t seem to work well, if…
…the alliancees have only a few members or at least under 20.
…the alliance contains defenses from absolute noobs (1k tp) up to high power teams above 4k tp.

Matchmaking works great for almost full alliances with almost similar powered rosters.

2 Likes
15

I have facing the same problem, but we can handle it, because our member has more depth ofense over top 5 defense.

War Match Order from Top to Bottom Factor Priority

  • The best 30 heroes in hero roster of each Alliance member opted in for war
  • Out of the 30 heroes, the most weight is put on the best five heroes of each Alliance member
  • Player count of the Alliance (also subtracting those who have opted out) - they try to get an even number of players on both sides, but it can vary by up to three members.
  • The troop strength of the best in each element
  • War history (minor impact compared to Hero Power)

Point

It is not based on Team Power, it is based on HP (health).

1 Like
16

Those are the key words - “supposedly” and “claimed” LOL. Well, no algorithm will be perfect. And someone has to be the winner in Wars.

It’s just bit frustrating when an opponent with a 4300 TP team eliminates our (strongest) 3500 TP defense teams with each of their 6 flags. Then their own 3000 TP players go to town on our 2000’s.

2 Likes
17

Agree on this, based on observed results. With opponents roughly the same as ours based on the makeup of defense teams, scores are usually reasonably close. Blow outs either way happen only with the more unbalanced setups (one or two clearly superior opponents along with multiple at par or inferior ones).

2 Likes
18

Our matchup this time around was BAAAD. We have one 2900 TP member who’s benches aren’t deep at all, a bunch of 3400, 3500, 3600, and 3700, and 11 above 4000 TP members who’s war teams scale down significantly to the mid 3000s. We are not able to stand up to a team with 2 3600 TP defenses as their lowest two teams and 7 3800 tp teams, everyone else being 4000 or higher. What is our small fry supposed to do?? I think matchmaking needs to be fixed so that individuals are paired against individuals within the alliances. Not just the (in our case) strongest 690 heroes and 115 strongest troops averaged into a singular score. And Win/Loss should definitely be considered in a running ratio AND an average. I also think flag useage should fall into consideration. If a team won, “scraping by” with 3500 points, but only used half their flags, they should get matched with a team that regularly scores 3500 points but uses EVERY flag. If the alliance participation changes, the matchup would then be unfair.

19

Re: your flag suggestion

That would then lead to flag manipulation.

1 Like
20

Very true, and excellent point. I do not do things like that, I am a rule follower, therefore I do not often think about ways a system can be manipulated lol. Clever and sneaky people ruin it for everyone like me haha