**WINTER OF WOE - BONUS OBJECTIVE POINT**
As previously announced, the team will be distributing an additional point toward milestones to anyone who completed the Absorbing Man fight in the first step of the Winter of Woe.
This point will be distributed at a later time as it requires the team to pull and analyze data.
The timeline has not been set, but work has started.
There is currently an issue where some Alliances are are unable to find a match in Alliance Wars, or are receiving Byes without getting the benefits of the Win. We will be adjusting the Season Points of the Alliances that are affected within the coming weeks, and will be working to compensate them for their missed Per War rewards as well.

Additionally, we are working to address an issue where new Members of an Alliance are unable to place Defenders for the next War after joining. We are working to address this, but it will require a future update.

Analysis of AW Season 1: Tier vs Bracket

24567

Comments

  • DNA3000DNA3000 Posts: 18,558 Guardian
    Maat1985 wrote: »
    @DNA3000 alos if you are saying it may take 6 weeks to balance....
    For one i have seen this going on for much longer than that.
    Also with a season of only 8 weeks that is almost an entire season where the weak ally has a huge advantage

    Half a dozen wars would be about two weeks, and that's not two weeks at the start of the season, that is two weeks in absolute terms. In other words, factoring out jockeying, it should have already happened for alliances that have existed long enough. Churn would reset the clock for new or suddenly shuffled alliances.
    Also i showed you the scores to the most recent... to show it was easier... i showed you 1gt won with only 1 bg clear and 77k score meanwhile tcr lost and we got 3bg clears and 95% clear....

    I can see that but that can happen in any war: in our season we had wars won very easily and wars lost after extremely tough slogging at or near the same war rating: that's could be representative of the jumble of alliances that exist at every tier.

    If what you were seeing was consistent and pointed to a serious match up imbalance, and I don't say that's impossible, we would have to explain how those "easy" alliances maintain their war rating in the face of eventually having to fight tougher competition. In other words, every time your 3 million alliance wins an easy war, and assuming they keep doing that repeatedly against that level of competition, you're acting to put pressure to reduce the ratings of those alliances. They should fall below the level where your 3 million alliance would ever face them again. What causes those alliances to "bubble up" back to levels where both your 3 million and 8 million alliance consistently face similar war rating opponents at radically different difficulty levels?

    Also, let's say your basic premise is correct: that while there's no guarantee, it is far more likely that a 1000 war rating alliance with 3 million alliance rating is significantly weaker than a 1000 war rating alliance with 8 million alliance rating. That would mean that war rating is not converging on an accurate measure of strength fast enough. How would we fix that? The basic problem is that the only way that can happen is with striation as I mentioned above: because the 3 million alliance rating alliances are facing each other and the 8 million alliances are facing each other, the stronger alliances are beating each other up at just the right rate to freeze their ratings (because anyone that wins abotu half the time tends to hover at their war rating) relative to the 3 million alliances. The obvious solution to that is to break the striations: force everyone at the same war rating to fight everyone else regardless of alliance rating.

    In other words, if the 3 million alliances had to face the 8s and vice versa (assuming they had the same war rating of course) then no 3 million alliance could "hide" in a sea of 3 million match ups. They would have to, in a sense, "prove" they deserved their war rating by facing all alliances with the same rating, not just the ones with the same alliance rating.

    *IF* your position is correct and *IF* we can gather enough data to prove it is correct, the next problem is that the obvious solution is going to make a lot of players angry. Math never convinces anyone of anything, and it will be very difficult to convince people that an 8 million alliance facing a 3 million alliance is "fair." In fact, this kind of match up didn't happen in the past: we used to have match ups against wildly different alliance ratings prior to season one. It is possible Kabam added this feature to match ups to respond to complaints about match ups involving far apart alliance rating alliances. If so, and if you are correct, Kabam may have created the problem in trying to solve a different problem players were complaining about.
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    You may be right here.
    aubvxlpvyxeb.jpeg

    And i do agree with what you are saying @DNA3000 about 8mill beating up other 8mil while 3mil beat other 3mil. My point is how far can one climb before it chamges????
    Even if both were stangnant at around 1mil surely this is not fair as if the 8mil ever fought the 3 mil we know the likely outcome.
    As i sid i have seen this going for longer and i will continue to collect data.
    What i can tell you is i am very positive from experience that what i have shown wont change.

    And in regards to other allies havog not faced this. I disagree.
    I am almost certain that war matchmaking was chamged at some point around the time of changes to the new map.
    As i remember over 18 months ago in a weak 3 mil ally getting impossible matchups in tier 15 where we fought 11 mil allies. Nothing like this has happend in along time leading me to believe that matchmaking is changed and does infact include ally rating in its calculations.
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    edited April 2018
    As i said also i believe it changes vastly at the top end due to higher competition levels due to smaller pool sizes it is only natural.... and here would be less of this which is why i believe people dont see it as a problem. Most people are looking at the very top and seeing how it is affected.
    And most strong people who rebuild an ally rebuild with an extremely skilled crew and can win at a much higher rate than 50% when facing low average allies like mine
  • DNA3000DNA3000 Posts: 18,558 Guardian
    Maat1985 wrote: »
    As i remember over 18 months ago in a weak 3 mil ally getting impossible matchups in tier 15 where we fought 11 mil allies. Nothing like this has happend in along time leading me to believe that matchmaking is changed and does infact include ally rating in its calculations.

    About a month or two prior to the start of season one there was a thread where someone was complaining about wide alliance rating match ups: i.e. both alliances had similar war rating but differed in alliance rating by several million points.

    Kabam has silently changed how match making works probably more times than I can count or have directly noticed. When they introduced the new 14.0 war changes they changed match making more than once during that period alone (it was discussed in the threads related to those changes). That's part of what makes hard data difficult to come by. You need a lot to see patterns, but you also need to collect it in a relatively short period of time or you can't be sure something didn't change under the hood while you were collecting.
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    edited April 2018
    DNA3000 wrote: »
    Maat1985 wrote: »
    As i remember over 18 months ago in a weak 3 mil ally getting impossible matchups in tier 15 where we fought 11 mil allies. Nothing like this has happend in along time leading me to believe that matchmaking is changed and does infact include ally rating in its calculations.

    About a month or two prior to the start of season one there was a thread where someone was complaining about wide alliance rating match ups: i.e. both alliances had similar war rating but differed in alliance rating by several million points.

    Kabam has silently changed how match making works probably more times than I can count or have directly noticed. When they introduced the new 14.0 war changes they changed match making more than once during that period alone (it was discussed in the threads related to those changes). That's part of what makes hard data difficult to come by. You need a lot to see patterns, but you also need to collect it in a relatively short period of time or you can't be sure something didn't change under the hood while you were collecting.

    I feel also too that the larger discrepancy rating is certainly in part due to higher tiers and therfore smaller pool so less options. But it is more fair that a weak ally faces a strong ally than a weak ally being rewarded same or better than a stronger one.

    I am not saying i know the fair solution to this but it seems to me that this.
    If you include alliance rating in the calculations it causes the issues of inability to climb and weaker and smaller allies sitting side by side
    If you dont include the allinace rating then weaker allies fet shat on by stronger ones....

    They need to find the right balance where stronger allies will naturally be able to sit higher than weaker ones.
    Maybe implement a division style system?
    Ar a point scaling system where by if you win a severly inbalanced matchup where you are the weaker ally you get double or triple points... and if you are the stronger bully ally you get half points??

    I dunno every system i think of does have some flaws and room for exploitation.
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    DNA3000 wrote: »
    Maat1985 wrote: »
    As i remember over 18 months ago in a weak 3 mil ally getting impossible matchups in tier 15 where we fought 11 mil allies. Nothing like this has happend in along time leading me to believe that matchmaking is changed and does infact include ally rating in its calculations.

    About a month or two prior to the start of season one there was a thread where someone was complaining about wide alliance rating match ups: i.e. both alliances had similar war rating but differed in alliance rating by several million points..

    Also several million points difference can be many things.
    A 3 mill ally v a 7 mil ally is going to be much harder than
    A 15 mil ally v a 18mil ally as their top champs are likely similar.

    Also another reason why including alliance ratining is unfair imo is simple as this
    I have a 488k rating yet have only 1 4/55 champ. All my other top champs are 5/50.
    I know people with 300k accounts with multiple 4/55 champs.
    If one ally was full of peeps like me rating would be 14.6 million.
    If one ally was full of peeps like the second the rating would be 9mil.
    In this scenario it is likely the 9mil ally is actually stronger for war as each person has 2x 4/55 on attack and defence each where as the 14.6 mil ally has almost all 5/50 champs

  • DNA3000DNA3000 Posts: 18,558 Guardian
    Maat1985 wrote: »
    DNA3000 wrote: »
    Maat1985 wrote: »
    As i remember over 18 months ago in a weak 3 mil ally getting impossible matchups in tier 15 where we fought 11 mil allies. Nothing like this has happend in along time leading me to believe that matchmaking is changed and does infact include ally rating in its calculations.

    About a month or two prior to the start of season one there was a thread where someone was complaining about wide alliance rating match ups: i.e. both alliances had similar war rating but differed in alliance rating by several million points..

    Also several million points difference can be many things.
    A 3 mill ally v a 7 mil ally is going to be much harder than
    A 15 mil ally v a 18mil ally as their top champs are likely similar.

    Also another reason why including alliance ratining is unfair imo is simple as this
    I have a 488k rating yet have only 1 4/55 champ. All my other top champs are 5/50.
    I know people with 300k accounts with multiple 4/55 champs.
    If one ally was full of peeps like me rating would be 14.6 million.
    If one ally was full of peeps like the second the rating would be 9mil.
    In this scenario it is likely the 9mil ally is actually stronger for war as each person has 2x 4/55 on attack and defence each where as the 14.6 mil ally has almost all 5/50 champs

    I actually don't think they should include alliance rating, and I said so in a lot of different threads. But I'm in the minority there. The problem you're claiming to be observing (earlier, not in this post) is one potential problem, but there are others.

    The biggest problem I have is roughly similar to what you state above: there is an advantage to manipulating alliance rating lower, which you can do by selling lower champs. Most players don't want to do that, but you don't want to incentivize such behavior. That would be counter-productive in a game that otherwise is based on encouraging players to collect champions.

    You can't "manipulate" war rating except by losing, and that's not manipulation, that's losing. But you can manipulate alliance rating, so no match up system should rely on it to determine fair match ups.
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    edited April 2018
    @DNA3000 i agree totally with that statement.

    If they expanded prestige to calculate based of a persons top 10 champs and then used that to calculate war matchups that may well work to solve manipulation and would create dair matchups based on strength.
    But would still have the problem i was saying as to weak allies mostly fighting weak allies whilst strong ones mostly fight strong ones
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    Also @DNA3000 this is the matchups it has just found for us....
    1gt slightly higher war rating and higher tier....
    Both allies have once again found similar alliance rating matchups....
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    edited April 2018
    tz7d8vwdzlyc.png
    xcjro8mdp2wp.png
    bbm73e1v3rfo.png
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    edited April 2018
    Double post deleated
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    iz21p5ecq6f9.png
    kmadattppi3o.png
    qx02nwb16tg1.png
  • DNA3000DNA3000 Posts: 18,558 Guardian
    There is sufficient evidence that the game is using both to match: war rating and alliance rating. The next question is whether experience suggests that a 2 million alliance rating at 1000 war rating is always or far more often a weaker alliance in war in terms of ability to fight a war than an 8 million alliance of similar war rating. You'll have to fight out the wars to see if that continues to be true.
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    @Kabam Miike
    @Kabam Wolf
    This kind of matchmaking is not fair...
    Something needs to be done....
    My strong ally is struggling to climb whilst my weaker one is doin just fine...
    I know which of my two allies would win if we fought each other...
    I know who deserves to be higher than the other...
    Maybe now people will begin to understand what i have been trying to say this from a while and people dont believ it.
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    @DNA3000 this war will not be good to determine skill as just now the 8 mill ally we are facing has crumbled and half the members have left gofting us a win...
    Will have to wait till next round to start collecting more usefull
    Data... lol...
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    edited April 2018
    Certainly stronger allies are more likely to get a full clear just purely due to strength of attackers.

    Yes the defense may be weaker but soooo much strength comes from the nodes which is same same and therefore easier for a stronger attacker to bypass... also from my experience some weaker allies like 1gt have a few strong peeps that can and will clear bosses while linked...

    Whilst the other ally gets stuck as they cant beat our few stronger guys
  • DNA3000DNA3000 Posts: 18,558 Guardian
    Maat1985 wrote: »
    @DNA3000 this war will not be good to determine skill as just now the 8 mill ally we are facing has crumbled and half the members have left gofting us a win...
    Will have to wait till next round to start collecting more usefull
    Data... lol...

    To be statistically valid, you should track all data. Technically, this would be an example of a "not so strong" 8 million alliance. If the 8's are much stronger on average, future wars will eventually statistically overwhelm this one. If that doesn't happen, it suggests that part of the reason why either some 2's seem higher than they should be or some 8's seem lower than they should be is because there are enough outliers like this that change the nature of the competition.
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    @DNA3000 i will track it but i was just saying imo it is not an accurate representation. It will certainly be one of the outliers not in line with the majority.

    And i can tell yu as a starting point.
    My 8mil ally has assigned bgs and assigned paths and we push alot... we communicate all day and try and lush each ofher... we are not top tier level but decently organised and have some very skilled players and alot of average skilled players.
    My 3mil ally is not as acitve, we have assigned bgs, but not paths, we struggle to tet good exploration as people are just not active enough. We have only a very few strong skilled players with most being low skill weak acoounts.
    Theee is a hige difference between the two allies in terms of skill and performance.

    So this is an example of a 3mil ally being half assed....
    While a 8mil ally trying as hard as possible and let down sometimes due to a little lack of skill from some peeps.
  • DNA3000DNA3000 Posts: 18,558 Guardian
    Gmonkey wrote: »
    This is pretty disturbing that this can happen, do I sell a bunch champs to lower rating ensuring we get easier Matches. People will game the system for sure

    I wouldn't. If @Maat1985 's data holds up, it points to a significant match making issue that Kabam could make changes to address without warning. Every match making change they've done recently has been implemented silently, so you wouldn't even know the change had been made until you experienced it first hand.
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    @DNA3000 so far it is holding up...
    Tcr just lost to an 11 mil ally
    While 1gt just beat a 2 mil ally.
    Now placing tcr down in tier 12
    And puttin 1gt in tier 10.

    Please see this thread where i have posted the recent data.
    http://forums.playcontestofchampions.com/en/discussion/62874/sort-out-matchmaking-before-season-2#latest
  • DNA3000DNA3000 Posts: 18,558 Guardian
    Maat1985 wrote: »
    @DNA3000 so far it is holding up...

    I don't mean the match making algorithm: you've convinced me that match making is using alliance rating. I mean I'm curious to see if your observation that alliances with lower alliance rating are consistently stronger than alliances with higher alliance rating but the same war rating. That would imply match making was separating alliances into different "strength layers" within the same war rating range that contained radically different war fighting strength, like an inversion layer in the atmosphere.

    Hard to say in between seasons because I'm personally finding that war in general is easier, probably because alliances are trying less hard during the break, saving resources for when the season restarts.
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    Yeah wat would be really good and helpful @DNA3000 is if kabam would atleast acknowledge this issues and comment on it. But it would be amazong if they could give a bit of insight as to how matchups work and why they are set to work in such a manner.
    But they keep ignoring this @Kabam Miike @Kabam Wolf @Kabam Vydious @Kabam Loto @Kabam Zibiit . I would tag more but thats all that appeared.
    Cmon kabam you told us you would give us more transparency and comunication yet you seemingly ignore an issue as great as this
  • GroundedWisdomGroundedWisdom Posts: 36,193 ★★★★★
    My observation is that it takes both Average Rating and War Rating into account, but the primary factor is War Rating. Now, I suspected originally that the Average Rating was just reflected based on what Tier an Ally fell into based on capabilities, but more often than not, it comes in close proximity. During the Season, it was quite askew because Matches seemed broader. My thoughts are that the default is War Rating, but Ally Rating is a secondary factor as well, specifically the Average Rating.
  • DNA3000DNA3000 Posts: 18,558 Guardian
    My observation is that it takes both Average Rating and War Rating into account, but the primary factor is War Rating. Now, I suspected originally that the Average Rating was just reflected based on what Tier an Ally fell into based on capabilities, but more often than not, it comes in close proximity. During the Season, it was quite askew because Matches seemed broader. My thoughts are that the default is War Rating, but Ally Rating is a secondary factor as well, specifically the Average Rating.

    I have no idea what average rating is. There is alliance rating which is the sum total of all individual player ratings, and there is war rating which is the cumulative rating given to an alliance based on their war win/loss record. Nothing in the game I'm aware of corresponds to average rating. The only rating calculated from averages is alliance prestige, which is the average prestige of all members of the alliance.

    There's now ample evidence that the game currently matches based on war rating and alliance rating, and there was direct evidence during season one that when there was no match possible based on war rating the game sought the closest possible match based on war ranking. If you have additional data to contribute, please post examples of what you are describing as "average rating matching."
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    My observation is that it takes both Average Rating and War Rating into account, but the primary factor is War Rating. Now, I suspected originally that the Average Rating was just reflected based on what Tier an Ally fell into based on capabilities, but more often than not, it comes in close proximity. During the Season, it was quite askew because Matches seemed broader. My thoughts are that the default is War Rating, but Ally Rating is a secondary factor as well, specifically the Average Rating.

    Nothing was askew in the lower tiers where there is a larger pool to choose from. 3mil v 3mil (or similar) 8mil v 8mil (or similar) for the entirety of the season and beyond.

    An it is just not a good system.... a fair system would not sperate into alliance rating and therefore allow he best 8mil ally to get same rewards as the best 3 mil ally....
    In wat world would a 3 mil beat an 8mil.... very very rately....
    An averge 8mil get same rewards as an averge 3mil???
    How is that fair???
    As i said work harder and spend time to grow only to get same or lesser rewards than weaker newer people,..
    Seems very backward
  • GroundedWisdomGroundedWisdom Posts: 36,193 ★★★★★
    If you check the Stats, there is a metric that shows the Average Player Rating. My observations are that it's consistently similar, though not always.
  • DNA3000DNA3000 Posts: 18,558 Guardian
    If you check the Stats, there is a metric that shows the Average Player Rating. My observations are that it's consistently similar, though not always.

    Average player rating is alliance rating divided by the number of players. I always considered it to be the average of the player rating, not an actual alliance stat. I'm pretty sure the game doesn't match based on that, because if it did then alliances with fewer than 30 members would get matched against alliances significantly higher in alliance rating than they are, and I haven't seen specific reports of that happening.
  • Maat1985Maat1985 Posts: 2,237 ★★★★
    @GroundedWisdom this would essentially gove the same result. Except allowing for alliances that have only 20 memebers to be matched with stronger allies that have 30memebers. But i a 30 v 30 matchup the result would be the same.
    Still as i said i have a 490k account rating but only 1 4/55 champ.
    While someone else may have a 200k rating and 4 or 5. 4/55 champs.
    And alliance may be full of one or the other.

    A big reason why strength of anykind should not be considered
  • GroundedWisdomGroundedWisdom Posts: 36,193 ★★★★★
    DNA3000 wrote: »
    If you check the Stats, there is a metric that shows the Average Player Rating. My observations are that it's consistently similar, though not always.

    Average player rating is alliance rating divided by the number of players. I always considered it to be the average of the player rating, not an actual alliance stat. I'm pretty sure the game doesn't match based on that, because if it did then alliances with fewer than 30 members would get matched against alliances significantly higher in alliance rating than they are, and I haven't seen specific reports of that happening.

    I'm just saying I've observed similarities when Matching. Have a look for yourself if you like. I do agree that Rating plays a part up to a point, but I think War Rating is always the default.
Sign In or Register to comment.