**Mastery Loadouts**
Due to issues related to the release of Mastery Loadouts, the "free swap" period will be extended.
The new end date will be May 1st.
Due to issues related to the release of Mastery Loadouts, the "free swap" period will be extended.
The new end date will be May 1st.
Options
Comments
War rating has an advantage other ratings don't: it self corrects over time because it is based on wins and losses. Shell alliances were a problem, but with seasons they become less of a problem because there's no mathematically valid way I'm aware of to use shells to generate a higher seasonal point total than you would otherwise be able to achieve.
It depends on your definition of fair, but if the definition of fairness is that two alliances should be matched up if they have a roughly 50% chance of winning when measured statistically, the mathematically correct way to generate rating is to infer AW strength through wins and losses and adjust rating only by that input. Inputs independent of performance (i.e. alliance rating) are virtually guaranteed to always generate worse results across all competitors, because they won't converge on the correct result through gameplay.
Anyway,I'll tell ya,I'm a mid tier player,in a 4mil,Decently skilled alliance,We were matched against people with rating higher than ours,sure,Difficult?Heck Yes!
Not possible/Unfair?Not in the very least.
We won atleast 3 out of the 5,I remember being tough.
A fly weight can win all his matches at 106lbs, but you put him vs heavy weights and he won’t do so well.
Maybe the prestige of the top 10 champs in a players roster is a better way to go.
Whereas if you treat war rating like chess, then war rating is comparable to ELO or FIDE rating.
The question is: do we segregate players by roster or do we allow everyone to compete together and reward the players with the stronger roster with the advantage of having better tools than the next guy? And the answer in this game is, we generally allow players the advantage of having a stronger roster. Arguably, the alliance with 2200 war rating with 6 million alliance rating might be working harder to get that rating than the alliance with 12 million alliance rating, but this game doesn't recognize the difference between strong skill and strong roster. This game rewards the combination of roster+skill in total only.
Since this game treats different players with different rosters and different alliances with different rosters usually equally, MCOC is more like Chess than boxing. In Chess, youth has advantages and years of experience has an advantage, but the 80 year old that starting playing Chess last month has to compete on exactly the same playing field as the 28 year old grand masters with 19 years of experience. If the 80 year old that just learned to play has a FIDE or ELO rating of 1300, then he plays against all other 1300s regardless of age, physical or mental condition, or years of experience.
That’s cause you search for only 2 BGs. Based on your alliance profile, gold 1 is too high.
@DNA3000
I’m not going to parse out the chess ratings, but what I will say is that in chess everyone starts with the same amount of pieces and the board has 64 squares. It comes down to knowledge and skill
Matching making based on Prestige is probably a better (not perfect) way to go than taking players with 5/50 rosters and matching them against rosters with 5/65 and r2 6*. The only downside for the stronger Alliance is that they may lose some war ratings points, The weaker alliance great reward for winning isn’t a stronger roster, it’s more tough opponents. It’s better for them to tank those fights, which hurts the alliances competing for top positions when they get matched with evenly matched opponents.
Maybe there should be minimum requirements to enter certain war tiers.
If the fly weight lands a lucky punch and wins the fight against a heavy weight, his reward is he gets to fight more heavy weights.
Alliance Rating corrects itself a lot quicker than War Rating does, though. In fact, Alliance Rating corrects itself virtually instantaneously. War Rating takes a very long time to correct itself. For example, recently there was controversy about a YouTuber and a group of top end players who originally decided to start a new Alliance. Had they gone with original plan and started from the bottom, how long would it take for their War Rating to correct itself and place them in the proper tier? Conversely, if a top tier alliance rebuilds with lower level players, how long before their War Rating corrects itself?
Since War Rating did not reset at the start of the current season, the positive and negative effects of shell alliances from before the season still affect everyone's War Rating. It's still possible to run shells, but it's more of a long game. With most alliances trying to run 3 BG's to get max points, you can just have 10 guys in a shell running 1 BG to inflate the shell's War Rating, then hop after the season's over. Which that's a whole different flaw with War Rating. The ability to win with one or two BG's is not representative of the ability to win with 3.
I would wholeheartedly disagree, there. Being realistic, War Rating is an alliance's past ability to win wars. Alliance Rating is a better reflection of an alliance's current ability to win a war. Would you say a 2 mil Alliance has a 50/50 shot against a 20 mil Alliance because their War Ratings are the same? I think most would say the 20 mil Alliance would have a closer match against another 20 mil Alliance even if their War Ratings are very different.
In the context of fair match making, alliance rating corrects itself never. You are presuming that alliance rating is an accurate measure of the strength of an alliance, and thus it is always "correct." But when you design a rating system for competition matching, you don't define fair matches in that way. The way this is done, and this is not an invention of mine, is you define "fairness" operationally by some theoretical result. For example, the most common definition of fairness used in all designs of rating systems is the notion that if two people have the same rating, statistically speaking one will win half the time. Or when there are rock-paper-scissors effects in play, then across all players with the same rating statistically speaking everyone will win half the time when matched against all other players.
If we have two alliances with ten million rating A and B and A *always* beats B, that's defined to be an intrinsically unfair match up according to most commonly used definitions of fair. If A always beats B, we intuitively know that A is stronger than B. But if we keep matching A against B, we are deliberately matching a stronger alliance against a weaker alliance.
The theory behind most rating systems is that we initially don't know who is strong and who is weak, and the only thing we know for certain is who beats who when they play against each other. That information is the only reliable information we can be certain of, because it actually happened. We then incorporate that information into a rating system that attempts to determine what the "true" rating of a competitor is based on their ability to beat other competitors. Given the right system of altering ratings based on wins and losses, we can state with mathematical certainty that under a given set of conditions the ratings will always converge on the "true" values. This is something that can be objectively proven.
Again, anyone saying alliance rating is a better metric for measuring alliance strength, or even a good one, has to address the problem of two alliances each having identical 4* and 5* (and 6* ) rosters, but one alliance having all of the 1*, 2*, and 3* champions maxed and the other having none at all. And this doesn't address the separate issue of not all rosters with the same rating are equally strong in alliance war, and doesn't address the completely separate issue of player skill. I can say this with certainty: given a specific population of alliances and given enough wars, eventually all of their ratings will converge on their mathematically correct ratings which determine their relative strength. This is mathematically provable. In the real world, ratings never perfectly converge because not enough wars are played and alliance configurations change, but there is still strong dynamic pressure that keeps the system always moving in the correct direction. Matching by alliance rating says "this is my wild guess as to what strength you are, and no matter how many times you win or lose, you will forever be matched against alliances with that same guess."
And as I mentioned previously, matching by rating creates a degenerate situation that is severe enough to make it unthinkable to match by only alliance rating. An alliance that only cared about winning wars and was willing to forfeit other parts of the game like allowing players to collect champions could instruct their members to dump all champions that were not useful to war. They would end up with a tiny alliance rating, and would then be constantly matched up against alliances far weaker than them. But not matter how many times they won, they would continue to get matched up against those same sheep over and over again, because their rating was unaffected by the number of wins they continued to rack up. This is a fatal design error.
I did directly address this point in the post you quote: in fact this represents the bulk of the post.
This would then penalize alliances with very skilled players capable of beating much stronger alliances and could move up in tier, but because they are more skillful than the average alliance of their prestige rating they would be barred from doing so.
Usually, what we hear are the complaints from people who find themselves matched up against higher alliances and think the match up is unfair. What we are probably not hearing are all the alliances who *benefit* from the match making system *allowing* them to move up to stronger competition, and thus getting better rewards. But you can't stop the one without stopping the other. I can guarantee if we put in controls to stop the former, the forums would immediately flood with complaints from the latter. And they would have the much stronger argument about unfairness and penalizing skill.
In your scenario where a skillful alliance would be penalized, because they would not be able to move up to higher tier until they have higher prestige is a bit of a specious argument.
1. This would not apply to the majority of alliances, so you are advocating for a system that catered to outliers
2. They would be able to compete once they reached the minimum requirements, which in this case would be based on prestige.
3. Challenger Rating is one sided, again this is unfair to competing alliances in the upper tiers. There is statistical probability of wins and loses when weaker champs are matched against stronger ones, it’s math you can’t argue that with anomalies.
Your other point about alliance ratings is valid, I’ve seen player with 500k total base ratings only to find out they have like 100 3* champs maxed out, and their top 5 champs are all 4*.
Using the prestige of the top 10 champs is a better gauge because these are the champs that are most likely to be used in War, but for all I know they could already be doing that. Some transparency from Kabam would be nice.
@DNA3000
I literally just said this to myself in my head and came to the same conclusion. It doesn't have to be alliance rating but there needs to be some measure of an alliance that can be assigned to make these fights more fair.
Well, you can use chess rating systems as an example, but those are designed for zero sum games with individual participants in which both participants have identical starting points. AW is none of those. In chess, you don't run into situations where one player starts the game down three pawns and the other player has an extra knight instead of a bishop. Chess ratings also work because there is very little, apart from mental illness, that can quickly impact a chess player's ability to win or lose a game. In AW, player traffic in and out of an Alliance can greatly swing an Alliance's ability to win wars very quickly. That's the main issue people are having with using War Rating exclusively. It isn't nimble enough to ensure fair matchmaking.
Your argument here is also flawed based on a fallacy, which is that two Alliances repeatedly matched based on War Rating will each statistically win 50% of the time but two Alliances matched repeatedly based on Alliance Rating will not. My Alliance recently lost 2 of 5 Wars to the same Alliance. Our War Ratings were virtually identical both times, but they had a significantly bigger Alliance Rating, tougher defenders on the whole, and generally seem to be more skilled as players. Neither war was even close, and if we played them 5, 10, or 100 times during that time period, the result would always be the same. No matter what rating you want to use, there is nothing to stop Alliances from getting repeated advantageous/disadvantageous matchups with the same Alliances. If two Alliances with a War Rating of 1000 keep getting repeatedly matched up and A always beats B, it's still an intrinsically unfair matchup.
I also don't see how you can argue that Alliance Rating doesn't correct itself. If 15 players leave a top alliance and get replaced with beginner level players, the Alliance Rating immediately changes. Alliance Rating grows at different rates for different alliances.
If matching by War Rating doesn't work in the real world, then it doesn't work. In a hypothetical perfect scenario it might work, but we're not in that hypothetical perfect scenario, so it doesn't make much sense to debate the point based on that scenario. Lots of things work on paper that don't in real life. You can argue that wins and losses are the only thing that matter when it comes to determining strength, but everyone knows that's not true. That's like trying to argue that the Philadelphia Eagles are currently the best football team in the world because they won the Super Bowl even though their owner went nuts, cut all their players, and replaced them with 8 year old kids.
I never advocated for matching solely based on Alliance Rating, so that's a bit of a straw man argument. It's possible to manipulate Alliance Rating just as it's possible to manipulate War Rating, and War Rating has and will continue to be manipulated. Even if they completely did away with War Rating and tiered alliances solely on Alliance Rating, which is not what I advocate doing, you would still want to have the highest Alliance Rating possible to get the best multiplier for Season rewards. Not to mention that selling all your champs except for your AW champs isn't feasible. Your 2* roster feeds your 3* roster and so on. You'll be opening a lot less crystals and will be crippled when it comes to grinding arenas, meaning you won't have a good shot at the next Blade and fall behind the power curve. You'll get easy Wars, but your in the bottom tiers, so your season rewards will suck. Plus, while your Alliance is basically sitting there capped at Alliance Rating, all those other little Alliances you beat up on will outgrow your Alliance Rating through natural progression and move out of your tier fairly quickly. Not really worth it unless you and 29 other peeps really get your rocks off by murdering noobs.
I'm not sure what else to say except pretty much everything you've said is wrong. Contrary to what you said, rating systems like ELO are not designed for "zero-sum games" where "both participants have the same starting points."
In fact, that rating system is designed on the presumption that both sides are not starting from the same starting point. The pieces on the board don't matter, because the rating system doesn't measure them: it measures the strength of the opponent which is presumed to be different and unknown until games are played. And contrary to what you've said, rating systems like the one being used for AW (outside discussions of the modification algorithm) not only work in the real world, to the best of my knowledge they are the only ones that are used in the real world. There's no real evidence that it doesn't work in general here outside of corner cases that all match making systems suffer from. And contrary to what you've said about alliance rating manipulation, not only is it possible to do, it actually has been done in the past for different reasons. People used to sell lower tier champs to use the resources to build stronger higher tier rosters specifically to sacrifice alliance rating to boost prestige. This very thing caused some issues when Kabam used alliance rating and not prestige to generate the compensation packages for 12.0.
And there is no way to manipulate war rating in a profitable way, because AW seasons incentivizes winning as much as possible. Pretty much any attempt to manipulate war rating by losing or by creating a new alliance generates far less rewards overall. You can only manipulate AW rating if you enjoy winning but don't care about earning rewards. In that case, you can manipulate yourself into easy and unprofitable wars. That's not really something I think Kabam should be spending a lot of time thinking about at the moment.
First Season8 War we've been matched with a 25M alliance. 8M bigger than us! 😒