The thing messing up matches now is that they factor in wins/losses, you guys must be winners and they must get their butts handed to them in the top 100.
@mhalttu has been good at confirming these.
At least under the old system we felt we had a chance, now it’s just no fun. We are always terribly outmatched. Our alliance has a nice mix of players. Because of this, many lower strength, new players just sit around feeling silly. The other team has 28 players well above 25 th lvl. They all have def teams well above 3000. Our alliance has over half players under 20, new guys with def team strength of maybe 2400-2700. Our oppenant all have fully leveled 4-5* heroes. We have mostly 3 and a few 4*. How is it fair?
1 Like
This the second matchup since the new update and we also have been greatly over powered with no hope what so ever of winning, the teams we come up against are mostly 3400 to 3980 in power with only a couple below 2900 and our allience is of the opposite with only a couple over 3000 and only just.
We where far better matched under the old format without question.
Our allience is giving it a couple more goes at it but if this keeps up then honestly what is the point. It’s only made the stronger teams attack the weak ones is all this change has done so far which to me is if no surprise.
1 Like
Does this mean that it no longer takes each members top 30 heros into account but instead only one form each element.
I Think its more the win/lose your alliance have achieved. We might get some missmatch.
Might 
1 Like
That sounds like it refers to troops only, the 30 heroes metric isn’t mentioned in that quote
2 Likes
I’m finding this app to start to chest in the raids. I have 6 world energies fir the raids and I notice when I have won 3 in a row it starts to make me lose on purpose even to very low ranked apponets. I have to wait a couple hours before I am “allowed” to win again
Kerridoc, thank you for taking time to reply- greatly appreciate that from you.
A couple of open points to your reply- taking account of win loss history and “we find you an easier team next time…” is an open admission that the war matching is deliberately biased- deliberately designed to engineer so-called wins and so-called losses. How is this in any way fair?
Most of the members want a genuinely fair match- like an unbiased die- and even if we roll a 6 on the die 100 times in a row- that history has absolutely nothing to influence that the probability of rolling another 6 is still 1 in 6…
… unless you would deliberately engineer the die to roll a 1 regardless? Are you sure using history is a fair process to “twist” the odds? Remember- most people pay is some way into this game so as a service provider the company has an obligation to provide a fair, open and transparent service.
To the Mann–Whitney test- how you outline os one way- another would be to collect stats- average top 10, average top 20, average top 30, matching variance, etc and then- to filter and refine the search more- add a Mann–Whitney test to select those teams with near identical Us. As a divide and conquer approach- assuming the goal is to find a near identical power match- the average, variance and distribution will align.
Oh, and by the way, of course we are losing heavily… how can an average team power of 2150 over turn a defense average of 3150? Plus- the points average is 20% in the opposition favour- plus enemy aid reduces our defense… understandably- a Mann–Whitney u test of42-68 shows just how imbalanced we are.
And for the record- I don’t want to be given an easy win based on foreseeable losses- i just want a fair match
1 Like
Yea I know but it doesn’t say it does still either and it’s that NOW word that could be interpreted to be either way.
This point needs to be made clear by the SG devs so that we know exactly if this is in addition to the current 30 heros or instead of.
Either way this troop power plays a big factor as most mid range players who are still trying to reach TC20 levels fast would worry about troop power to much but mostly concentrating on leveling strong heros.
Also taking into account even within this forum I myself have read things like " don’t place to much importance on upgrading troops untill you get tc20" (example only as it may not be word for word) comments.
The other question to this is;
Troops ate used with heros which makes up thier special skills which in turn forms part of their TP score. CORRECT?
Would this mean that also taking the best (which means strongest) troops points and adding it to the TP total of 30 heros for AW’s mean that these troops power is being counted twice for each element considering thier more than likely already being used by a hero.
By placing troop power into it greatly changes matchmaking a great deal.
Unfortunately I dont fully understand what you mean.
The system takes into account your war performance together with other factors like strength of top 30 heroes and strength of the strongest troop of each color.
When the system has had time to adjust you will hpefully paired with opponents that have a similar war performance which should lead to more even pairings.
Unfortunately this needs some time so lets wait for about a month to see how it works out.
Please understand there will still be matches where you will be outmatched but there should be no prolonged winning or loosing streaks.
Meanwhile best of luck with your next war.
Thanks for replying. Actually this is my third out matched. Luckily, i can do statistical calculations to see if there is any point participating in a war effort.
My point is- your matching algorithm is fundamentally flawed.
Compounding matters worse- the points allocation- eg our alliance average power is 2150; ten members vs opponents average 3150 also 10 members… total points are circa 1500 each- means average 150 points per member. However, the 150 average means for us to beat 3150 team power- we get only 150 points- and for opponents they get 150 to beat just 2150 team power.
But we know it’s not really “per member “ divided- it’s scaled within our alliance- those below 2150 average are worth less than 150 snd those above average are worth more…
Which means- two guys at war- both have circa 3400 team power- but one gains you just 150 points and the other team gains you 195 points…
And anyway… what means “easier team” - do we just match loser with loser and winner with winners? So the team average 4500 who lost 4 wars in a row will be our next opponent?
Or, does “easier “ mean that actually you can calculate a matching balance… how do you know that a matching of average 3150 vs 2150 is fair or not?
At a guess I say “you don’t know “ because if you did, then you would understand mean average, variance, deviation, Mann–Whitney u test, t test(though not applicable due to non normalize team powers).
We’ve had three close matches recently where victory has really come down to who uses more flags.
Fortunately our alliance has some insufferable war monger who bugs everyone constantly to use their attacks…
2 Likes
Do you mind me asking the alliance involved- you vs them? I will assume their raid defense lineup is broadly similar to their war defense lineup- and I would like to understand what means close in terms of average team power, variance, rank, distribution, etc.
Just need the names- all other stats i can calculate
