Flawed Match Making

About 6 seasons ago, Kabam changed their match making algorithm from strictly war rating based to something else. The old system worked well until alliances started tanking in the off season in order to get easier wars at the beginning of the next season.

I've taken the time to look at the alliance ratings for the top 50 in each tier. I know this isn't perfect as a small handful of alliances may have sold champs. But generally speaking it's a good gauge to the relative strength of an alliance. What I have found is shocking.

You're match making system is clearly broken. The top 3 in master looks OK but then it starts to fall apart from there. There are (2) 23m alliances in Master while you have a 64m and 46m in Plat 1. Certainly the 23m alliances in master are not better at war than alliances double their strength in Plat 1.

I think your primary match making criteria is prestige with war rating second. This is why you see so many different strength alliances within the same tier. There is no way that 18/19 million alliance in Plat 1 ever gets matched up against 46 and 64 million alliance ranked lower than them. In a fair system, you would not see such a wide range of alliance ratings within the same tier. Look at Gold 2, there's a 43m at #7 and a 12m alliance at #12. In what scenario is an alliance 1/4 the size only slightly worse in war ranking than a much stronger alliance?

In all tiers outside of top 3 master, you have much smaller alliances in Gold 2 on up. I'm not saying these alliances don't deserve to be there as they've won wars to get to their war rating. However, they are not getting matched with alliances at the same war rating and prestige. What you have is the best 4k prestige alliances at war gets a spot in Plat something. And so on it goes through the various prestige ranges.

It's great that you have increased the rewards for war season rankings but you really need to look at how you match alliances. You have also sort of fixed the tanking problem during the off season. At this point the integrity of rankings for seasons is completely gone. Why should larger alliances even care about war when much smaller alliances finish in top tiers and they don't ever get a chance to face them?

Please Kabam give me a reason to believe in wars again...

Comments

Mcord11758

Prestige has very little to do with skill. You can have prestige easily over 10.5k just by spending without ever defeating the champion in 6.2 and you can explore all of act 6 and still have prestige sub 10k

GOTG

If so please give us some or even 1 high skill alliances with small rosters. We would absolutely love to get beaten by them. But we never had such honor because at gold 2 we always got matched to same size alliance and every war is so equal and difficult for both sides.

Worknprogress

While I'm sure there is a possible advantage to be gained, it's by no means guaranteed. People were calling it cheating where in my own experience it was a strategy that had just as much chance of blowing up in your face as it was favorable. You're at the mercy of matchmaking no matter what. Is it possible that you were more likely to get favorable matches with a lower rating, absolutely and probably likely. It definitely wasn't a given though bc I played in alliances that got absolutely horrible matches after tanking, got ruined in some early wars, and then never could regain rating fast enough to make up for being in a lower tier (after the early losses)

Without basically doing elimination style tournaments with seeded matches, you'll always be at the mercy of matchmaking. Sometimes it will work in your favor sometimes it won't. Regardless of what strategy alliances go with, the matchmaking system has been broken from the jump.

I'd much rather see a seeded tournament than this nonsense we have currently. Points systems only work in sports like European soccer bc its level. Everyone in the division plays everyone else in the division. No one has "easier" schedules than the others. Championships based on points where the top teams arent even playing each other for large parts of the season are ridiculous.

Midknight007

Dude, it was your last post before I commented. Yeah 5 days passed, but you kept bringing up about his original data set and being dismissive of what others had stated.

We cannot prove that alliances are being matched by War Rating And Alliance Rating... War Rating may be a factor, but it is also being weighted by Prestige. Quik also stated this. He loaded up a retired alliance and their 4m AR was going against 800k with similar War Ratings and Prestige.

While at first Quik thought it was possibly about Alliance Rating, we did not have access to Prestige at that time. With visibility in prestige being enabled, we are seeing that the trend is a mix of WR and Prestige. While it isn’t necessarily fair for a lower prestige to face higher prestige, Kabam not allowing them to face is causing lower Prestige alliances in higher brackets. The War Rating alone indicates that they have more skilled to their competitors of similar prestige. But consistently pitting their lower ranked opponents can become problematic.

In one example of a P4 at 8200, they had to go up or down Tier to have a match of equal Prestige. With the limit of alliance in these brackets, they have a strong chance of facing the same alliance over and over. That can be boring, and it can also be unfair to a gold 1 that has similar Prestige but keeps losing to them. Also, the G1 would have to play on a map with stricter nodes, which is also somewhat unfair... their prestige and rosters may not be up to par for T4 and T5. For example, Psychic Thorns ramps up very fast at that point.

If War Rating is supposed to be a sign of skill, then it should allow for across Prestige fights within the same bracket. Our opinion is that the matches wouldn’t be fair, and that the War Rating is artificially inflated because of algorithm Is preventing matchups.

If Kabam wants to keep these boundaries to promote fairness, then we may need separate Seasons based on Prestige... like 1000-5000, 6000-8000 and 9000+. Each can have it’s own master, platinum, gold, etc... and the rewards can reflect the current division they are in. Otherwise, you have lower skilled teams placing higher with softer schedules if the War Rating is a true reflection of their ability.

Otherwise, we have Division I, Division 2, and little league in the same Tournment where theoretically Little League could be in Master because they are overly skilled compared to their competition and the gap in points will prevent their competitors from catching them.

DNA3000

Cheating doesn't have to have a guarantee of winning, or even consistently working, to be cheating. It simply has to offer an unfair advantage.

I've been advocating for something like this for a very long time now. It isn't as easy as it sounds, but it can be made to work. Strangely, Kabam isn't calling me to help implement this.

For reference: I kinda took all the ideas I had from issue 14.0 forward and collected them into a single thread back in July: https://forums.playcontestofchampions.com/en/discussion/comment/917169#Comment_917169

Midknight007

I agree with you on the divisions...

On the tanking strategy... It is not guaranteed, but the chance are higher to face a lower skilled team by dropping WR. If the Algorithm first looks at AR, then looks at Prestige... A 9800 Prestige G1 might face a normally Plat 2/3 alliance with 10300 prestige. Prestige is not an indication of skill... it is more an indication of investment (whether that be time or money). A higher prestige might have a better defense, which can cause the lower skilled alliance problems.

With brackets being so disjointed... it is going to become even more of a jumbled mess. How many 6000-8000 alliances are there? How many 9000+? If they are going to be spaced in the same tournament and brackets without cross competition... how many tiers down would a P1-P4 alliance have to go down to find a match? It will create stagnation at the top with little to no variety of matches. You will keep fighting the same people as the brackets become more congested.

S-aerating might be a great ideal. It may not solve for shells, but it can give us time to sort things out further.

Worknprogress

I didn't say winning was guaranteed, I said an advantage wasn't guaranteed which it isn't. Without any widespread data I can't say for sure obviously that the somewhat lowered rating wasn't giving a much if any better chance at matches with a substantially higher chance of wins. All I can say is that from personal experience I never saw a large swing one way or the other.

There are just too many variables that have to be taken into account when using nothing but a number for matchmaking. A seeded elimination tournament would be at the least more fair in my opinion.

When you weight something like prestige in war matching it throws things off since prestige is how you score in AQ. That assumes an alliance wanting to compete in AQ also wants the same level of competition in AW. That doesn't make sense when you take your average high prestige champion's usefulness in AW. Typically, they're not the greatest.

If an alliance is ranking for max prestige, they're more often than not forgoing some war effectiveness unless they're spending enough or competing at a level high enough to rank incredibly frequently. Neither is a direct reflection of war ability.

Prestige being weighted for war matching has just never made even a slight amount of sense to me personally

Speeds80

I’ve tried I raise this issue several times, the issue is it’s created a bunch of high level relaxed alliances like mine who end up playing for scraps like tier 11 3* shards and gold 2/3 rewards. but who only ever get to face high level alliances out wars are often 5-10 deaths total, the issue I have is that my leaders with secondary accounts in 10m alliances are competing and beating us for higher rewards where they wouldn’t have a chance facing our alliance, because they only ever have to face 8.5-11.5m alliances, the problem I have is that the rewards I should be competing for are unfairly being rewarded to alliances who I never get a chance to go head to head against because my alliances prestige is too high, it would make sense for me to join a 10-15m alliance and walk through the opponents defenders and my 5/65 korg will get 5 kills a war, and I would place much higher in war seasons.
Vet alliances are suffering and people are leaving the game over it

QuikPik

It's not only impacting vet alliances. Basically any alliance that has high-ish prestige is effected.

Looking at Plat3, there are 7 alliances in the top 50 where the prestige gap is greater than 20% with the top prestige alliance in that tier. That's 14%.

In Plat4, there's 8 alliances with greater than 20% gap; that's 16%.

In Gold1 that number jumps up to 24 out of the top 50 for a whooping 48%.

In Gold2 it is even worse with 27 alliances that have greater than 20% prestige gap. That's 54% !!

Worknprogress

Those number also don't take high prestige alliances that don't care about war and run less than 3 bgs into account. Take ours for example that finishes top 5 in AQ but finished near the top of G1 bc we even had one war we only ran 1bg

QuikPik

@Worknprogress while true, there are (6) 10k prestige alliances in Gold 1 plus a lot more at 9k plus in the top 50.

Worknprogress

Right, was just bringing up that alliances running less than 3 bg definitely had an effect on skewing the high side of some of the lower ranks. Obviously not affecting Master and likely not P1 either

DNA3000

The specific statement you made I was commenting on was the fact you strongly implied that this wasn't cheating: "People were calling it cheating where in my own experience it was a strategy that had just as much chance of blowing up in your face as it was favorable." It is cheating whether in your experience it had as much chance of blowing up in your face or not. That has no bearing on whether cheating is cheating.

Worknprogress

It's no more cheating than professional athletes being rested in unimportant games to be better rested for an upcoming important one. Cheating is breaking a rule and there's no rule that says you have to keep your war rating at the highest you possibly can at all times. Starting the season with a lower rating is a strategy not cheating. Whether it's an effective one or not I'm definitely unsure of though.

Mcord11758

All my point was, an alliance can have 10+k prestige and still not be great at the game. Doesn’t mean that all 10k prestige alliances don’t have skill. Prestige is not a true measurement of skill but you can say it is a decent indicator

GroundedWisdom

Actually, I've had the same conversation with the OP, and I agree with DNA's stance. For some reason, they have the idea that the Leaderboard must be a reflection of Total Alliance Rating. There's not much convincing them otherwise in regards to it being wrong.

QuikPik

Let's be clear, I only used alliance rating in my initial analysis because that's the only public information we were privy to. I always thought match making was governed by prestige. I have since updated that list to show alliance prestige.

DNA3000

Resting players costs the actual team wins and losses. And almost no one calls deliberately losing cheating. That's not the problem. The problem is the combination of deliberately losing and alliance swapping. That combination is not analogous to resting athletes. The closer analogy would be if an NFL owner owned two teams, and deliberately put all his best players in one and his worst players in the other, tried to win the Super Bowl with the first and deliberately lost every game with the second to get the best draft picks. Then next season he took his best players from team one and swapped them into team two along with those best draft picks and moved the lesser players into team one so that now team one would try to lose every game and get the best draft picks while team two then tried to win the Super Bowl.

This isn't possible because there's actually a rule that says no owner can own two NFL teams. In fact the prohibition started with baseball a long time, for similar reasons as above.

Cheating is not "breaking a rule" in the sense of explicitly breaking a stated rule. No sport defines cheating in that way. No set of rules can accommodate all possible options. Most recently the Houston Astros were fined and penalized for cheating, even though technically speaking there's no rule against using electronic camera equipment to steal signs (or there wasn't at the time). This was considered understood to be against the rules of Major League Baseball, consistent with prior rules that implied players on the field stealing signs is part of the game, but anyone using technology to assist stealing signs has an unfair advantage. No one needed an explicit rule to know this was cheating, and Major League Baseball didn't need an explicit rule to punish them for cheating.

WikiLeaks

War rating system already determines an alliances capability of winning a war. Not using it to matchup alliances makes no sense at all.

This is an open and shut case.

DNA3000

It is being used to match alliances. What's being discussed is that the game for a while now has been using an additional criteria. If you're 2000, say, there might be many alliances with about 2000 rating. The game *should* pick either randomly from those, or use a system where every alliance has a similar chance of matching any other in that range. Instead, the game looks for other 2000 war rating alliances that *also* have similar alliance rating (or prestige: it was definitely alliance rating at one time).

This has the effect of "stratifying" the alliances. All the war 2000 alliances with 10 million rating face each other, all the war 2000 alliances with 20 million rating face each other, etc. Since the war 2000 rating alliances with 10 million rating only face other 2000 rating alliances with 10 million rating, and they don't face alliances with 20 million rating, they are facing a set of alliances that could be significantly weaker on average.

Now, all alliances of similar rating should have similar strength. But if two different alliances start with the same rating and wildly different alliance rating, they won't fight through the same competition to climb in the ratings. If they both start at 1000 rating and both end at 2000 rating, one of them could have faced weaker competition to reach that rating. That means two alliances with 2000 rating aren't necessarily the same strength, and this problem can compound over time, with two alliances with the same rating having divergent strength.

For war rating to be a valid measure of strength and thus a valid match criteria, all alliances of similar strength must have a similar chance of facing each other. They don't have to actually face each other constantly, they just have to collectively face each other often enough to "validate" their ratings with common opponents.

A couple of years ago this was pointed out by another player who likened the problem to war rating functioning like win/loss record while alliance rating functioned like divisions. It was as if all 10 million alliances were in a different division than the 20 million alliances, since they never faced each other. And so a 2000 war rating as a 10 million alliance didn't necessarily equate to the same strength as a 2000 war rating for a 20 million alliance. This is highly problematic. It was supposed to have gone away when Kabam switched to the mandatory match up system we have now, but if it was reinstated that's completely broken.

The problem is Kabam refuses to discuss how the match up system works because they think doing so will allow people to exploit its properties. But that presumes they did it right in the first place, and if they are using alternate ratings as secondary match criteria, then they didn't. Secondary match criteria is mathematically provable to break the kind of war rating system they use. This should be obvious to any game systems designer, but apparently not.

Worknprogress

Except that's not what it's like. It's more like a team getting sold and moving to a different division that's weaker with the same exact players it had before which definitely does happen.

DNA3000

That analogy breaks down for many reasons, but the most critical one is that owners cannot arbitrarily sell or move franchises for competitive advantage because the other owners wouldn't allow them to. It never happens in the real world that franchise owners keep jumping around every season shopping for the best competition because everyone knows that's wrong, so no one allows any owner crazy enough to try to actually do it.

Your analogy literally proves the opposite of what you're trying to claim. Teams don't arbitrarily move to new competitive divisions because they aren't allowed to, because everyone knows it is wrong. Moving a team to a completely different competitive environment only happens rarely, only with the direct or indirect approval of the other teams in that competitive environment, and only when there is sufficient justification for allowing it.

I think everyone would be entirely happy with applying your analogy to MCOC: alliances can only manipulate match systems to alter their competitive environment if all other alliances or a governing representative of them, directly approves such things, and attempting to do this without their approval generally results in a permanent ban from competition (which is what would generally happen if a franchise owner attempted to change competitive venue without their league's permission).

Worknprogress

And yet those same teams in multiple sports tank regularly to get either guaranteed better draft picks or a better chance in a draft lottery. Dont see any of them getting banned for it

DNA3000

And neither do people here get banned for losing on purpose. You seem to not understand the concept of what's actually going on. Maybe that's why you never saw any advantage to doing this: your alliance wasn't doing it right.

Worknprogress

You just proposed banning them.

I understand what's going on. People are assuming something is effective bc a lot of people were doing it with absolutely no evidence to support it other than their own anecdotes which aren't worth squat.

I honestly could not care less what's happening in tier 10 as it's fairly unimportant. It's not a competitive pool regardless of what the people in it would like to think. I've both tanked and not tanked in T1-4 and saw absolutely no concrete benefit to it whatsoever.

Just bc people say something is actually a problem bc it makes them feel better saying they were wronged instead of they just weren't good enough, doesn't mean it's actually a problem.

You can be as verbose as you like but end of the day, you don't have any better idea on whether it made a lick of difference than I do.

The more word vomit you lot keep spitting out about "issues" though, the more support the nonsense gets bc it gives people the warm and fuzzies feeling like it wasn't their fault. Then the "fix" ends up being worse and we're in another thread complaining about something new that's actually worse. Fixing things that dont actually need to be fixed just ends up actually breaking something

Speeds80

My war today we had 1 defensive kill, 27m alliance competing for 3*/4* shards in tier 11, it’s very disheartening

Speeds80

War was awesome when it was just war rating, too many people complained about having to face alliances who were 10m above them but in my opinion they were there for a reason and if you are winning a lot you deserve to have to face the big boys, I guess Kabam want to create this bubble where up and comings get to compete for massive rewards early without having to climb the ladder

rwhack

Wow.

If you have a high AW rating you get "rewarded" with tough match ups all season. If you start low and go up it's an easier AW season. Prestige should have zero to do with a match up but it's factored in.

Best thinking I read was seeded playoff or round robin type stuff.

We matched the top 3 2 times each one season and barely made master. Right now it's whoever matches kenob the least. gets top 3 out of a few alliances. Kabam decided to put a "rematch" algorithm in. There's a youtuber talking on a video about it with a kabam rep. That's poorly conceived. We have 12 matches per season. Matching someone twice makes no sense. Matching the top 3 twice each made even less sense.

Midknight007

I only say it is a problem, because Kabam actively seeks to stop it. When AW was off straight War Rating, it wasn’t a big deal to us. We knew when we were facing a team that tanked or was a shell. Often we would go for the clear without being too hard on deaths for those matchups.

I don’t mind facing shells or tanking provided that we also get to face alliances that we outmatch within the same bracket. But when we keep facing tight matches due to prestige based matchmaking, facing the occasional shell alliance means an extra loss with little guarantee to make up based on how close the margins are. Our matches have been only 10-70 points difference between win or loss. One misstep on either team means a loss, so you tend to spend on boosts, revives and pots just to be safe.

I never asked for Prestige based system, but Kabam seems to have introduced it because teams with 7000 prestige would enter Gold 1 or Gold 2 and face a 9000 Alliance with swelling rosters and lose. This was because the went into a bracket they are not skilled or prepared to be in. Instead of explaining this to the lower alliances, Kabam made sure they could only fight between equally matched teams. This means that teams that are very skilled with high prestige and AR are killing themselves to perform are now ending up in lower Gold and even Silver tiers.

This will only cause less spending and faith in AW Season. People will stop spending for lower rewards. It Kabam’s economy... but I would suspect item purchases will go down if we stay on this current course.

My comments on trying to solve for shell/tanking is because Kabam made all of these changes to try to solve for these issues. Saying to simply go back to War Rating will not go far if we do not at least propose and concede on something they are trying to solve for. Unfortunately, in times like these, we must compromise in order to find a common ground with Kabam’s intent for seasons.

One thing is certain, if things continue as they have been, AW Season is going to have less and less participation and cause more burnout.

DNA3000

I say it is a problem because it is mathematically provable that such tactics destroy the validity of the rating system, which then makes match making systems which rely on the integrity of the rating system also invalid. As the game operator has a vested interest in avoiding that situation, and as all competition participants can be expected to understand that activities provably capable of doing that are exploitive, it is safe to say this is a problem that better design of the system as a whole can address.

I'm sure you disagree, but it was never my goal to convince you of anything. The goal isn't to prove you wrong to you, it is to prove you wrong to the people who could change the situation.