News Realms Remove Ads - Go Premium! Sign In or Register

Bluetracker

Tracks Blizzard employees across various accounts.

ARENA PLAYERS Been trying to brainstorm what a fun ranking system could be for arena. Goals are for it to have forward progression but still have the best players come out on top. The Pitch ⬇️

ARENA PLAYERS Been trying to brainstorm what a fun ranking system could be for arena. Goals are for it to have forward progression but still have the best players come out on top. The Pitch ⬇️

  • Iksar
    Posted 3 years, 9 months ago (Source)
    ARENA PLAYERS Been trying to brainstorm what a fun ranking system could be for arena. Goals are for it to have forward progression but still have the best players come out on top. The Pitch ⬇️

    • Iksar
      Posted 3 years, 9 months ago (Source)
      Pitch uses 3 numbers. A rating (what is displayed to you and you are ranked on), your average wins per run, and your 'class score'. Arena Rating = AvgWins X ClassScore AvgWins = Your average wins per run ClassScore = Your highest win total on each class added together

      • Iksar
        Posted 3 years, 9 months ago (Source)
        Example 1 - Very Casual, Plays lots of classes Player plays 10 runs. Gets 3 wins every time. Plays 1 run of each class. AvgWins = 3 ClassScore = 30 (3X10) Arena Rating = 90 (3X30)

        • Iksar
          Posted 3 years, 9 months ago (Source)
          Example 2 - Average Engaged Player, Plays limited range of classes Player plays 30 runs. Gets 4 Wins every time. Only plays 5 of 10 classes.

          • Iksar
            Posted 3 years, 9 months ago (Source)
            EXAMPLE 3 - Hardcore and Skilled Player Player plays 60 Runs. Gets 7 Wins every time. Only plays 8 of 10 classes. Avg Wins = 7 Class Score = 56 Arena Rating = 392

            • Iksar
              Posted 3 years, 9 months ago (Source)
              EXAMPLE 4 - The best players Player plays 100 runs with an average of 8 wins per run. Gets 12 wins on 5 classes but has a best run of 10 on the other 5. Avg Wins = 8 Class Score = 110 Arena Rating = 880

              • Iksar
                Posted 3 years, 9 months ago (Source)
                One of the reasons I like this system is that it encourages class diversity, has mostly forward progression, but the best players will still ultimately have the best ratings so long as they choose a diverse subset of classes.

                • Iksar
                  Posted 3 years, 9 months ago (Source)
                  We'd probably give you 10 points per win so ratings are in the 1000's rather than the hundreds. This way they'd look more similar to BG/Merc numbers.

                  • Iksar
                    Posted 3 years, 9 months ago (Source)
                    I'd also consider giving 'bonus points' for 12-win runs. Meaning your score for an individual class could go higher than 12. This way there is still minor incentive to play classes you've gone 12 with.

                    • Iksar
                      Posted 3 years, 9 months ago (Source)
                      We'd find a way to display your score with each class in the client so you can track it. A benefit is that if we just use a rating that lives in our DB, we can do live updated rankings like we do with constructed/bg.

                      • Iksar
                        Posted 3 years, 9 months ago (Source)
                        Reset time is probably 2 or 4 months. Less important to figure out those details now.

                        • Iksar
                          Posted 3 years, 9 months ago (Source)
                          The reason I don't like using a traditional rating system is that there is no skill matchmaking in arena. This means that players would go 5-3 in a run and lose rating sometimes, or go 1-3 and gain rating. Can get confusing.

                          • Iksar
                            Posted 3 years, 9 months ago (Source)
                            I like having a rating that is more like a score because it's very clear what you need to do to increase your rating. "Oh I need to play these classes to a higher class score!" "I need to increase my average wins per run, right now it's 5.1 so getting 6 will help me!"

                            • Iksar
                              Posted 3 years, 9 months ago (Source)
                              Let me know what you think or tag an arena player to get their opinion. Also open to other suggestions, another way to rank players or the opinion that ranking is unwanted.

                              • RidiculousHat
                                Posted 3 years, 9 months ago (Source)
                                @IksarHS what if i tagged them 9 minutes ago already

                                • Iksar
                                  Posted 3 years, 9 months ago (Source)
                                  @RidiculousHat blame twitter

                                  • RidiculousHat
                                    Posted 3 years, 9 months ago (Source)
                                    @IksarHS this is why the pros use twitlonger btw. but i like it this way. more authentic

                                    • Iksar
                                      Posted 3 years, 9 months ago (Source)
                                      @RidiculousHat I enjoy the jankiness or twitter and the mental energy I receive from all the updoots on my phone. I'm like one of those 'top 100' websites that makes you click on a new page for the next thing on the list.

                                      • RidiculousHat
                                        Posted 3 years, 9 months ago (Source)
                                        @IksarHS ...dean did you just compare yourself to reddit

                                        • Iksar
                                          Posted 3 years, 9 months ago (Source)
                                          @RidiculousHat no more like https://t.co/8vPPJeQzgz

                                    • Iksar
                                      Posted 3 years, 9 months ago (Source)
                                      @RidiculousHat wait are you insinuating i'm NOT a twitter professional?!

                              • dannfullergames
                                Posted 3 years, 9 months ago (Source)
                                @IksarHS Interesting idea, need to digest a bit for opinions. It would be nice to get the last two leaderboards corrected (and avoiding what caused the errors for future ones) in parallel with contemplating a replacement idea, though.

                                • dannfullergames
                                  Posted 3 years, 9 months ago (Source)
                                  @IksarHS It feels like there might be a point where it makes sense to “park”, and continuing to play is likely to hurt ranking than help. Incentive to not play feels bad.

                                  • dannfullergames
                                    Posted 3 years, 9 months ago (Source)
                                    @IksarHS How to address player who plays the worst class and high rolls the draft and goes 11, and locks in an 11 average for that class?

                                    • Iksar
                                      Posted 3 years, 9 months ago (Source)
                                      @dannfullergames I think the answer is basically.... good for them. The best players won't have any issues getting 10-12 win runs on every class over the course of the season. The top players *should* be based primarily on avg wins per run.

                                      • Iksar
                                        Posted 3 years, 9 months ago (Source)
                                        @dannfullergames Because the best players will all have 100-120 class score relatively easily and be very close together.

                              • Tabayrak
                                Posted 3 years, 9 months ago (Source)
                                @IksarHS I love this... you know we could talk about this with duels too.

                              • seewhykai
                                Posted 3 years, 9 months ago (Source)
                                @IksarHS Only "ranking" should be leaderboard rank with average (or wins total). There can be a Class score that tracks various win milestones for each class within a season with 120 (10x12) as the max.

                                • Iksar
                                  Posted 3 years, 9 months ago (Source)
                                  @seewhykai Reasonable, and simpler. One of the goals I wanted to accomplish was to have it be a number that mostly goes up over time. If we just have avg wins your number is likely to stay the same all season. Also would have to pick a number of runs required to get a rating.

          • Iksar
            Posted 3 years, 9 months ago (Source)
            (i hate typos) EX 2 would look like this... Avg Wins = 4 Class Score = 20 (4X5) Arena Rating = 80 (4X20)

    • IronicallyEpic
      Posted 3 years, 9 months ago (Source)
      @IksarHS Overall this seems pretty good to me, however it doesn't reward players with the same win average over a larger number of games. (Unless I'm missing something which is entirely plausible)

      • Iksar
        Posted 3 years, 9 months ago (Source)
        @IronicallyEpic This is loosely true. You'd think that a player with a larger sample size would probably have a higher class score, since class score is high watermark and not average.

    • AMPKC
      Posted 3 years, 9 months ago (Source)
      @IksarHS @adwcta @merps4248 relevant

      • Iksar
        Posted 3 years, 9 months ago (Source)
        @AMPKC @adwcta @merps4248 oh no not THESE GUYS https://t.co/0OMrOXZSLV

        • Iksar
          Posted 3 years, 9 months ago (Source)
          @AMPKC @adwcta @merps4248 https://t.co/gcDgfJBtx6

    • seewhykai
      Posted 3 years, 9 months ago (Source)
      @IksarHS I really hope this ranking system will just be superficial and NOT affect matchmaking. I'm not sure how long a ranking season spans (~2 months like current leaderboards?) but I think there are some issues with the Examples criteria.

      • seewhykai
        Posted 3 years, 9 months ago (Source)
        @IksarHS Example 1 seems more like the average player to me. A Casual to me plays under 10 runs every 2 months and would average under 3. Example 2 description seems fine, though 30 runs seems more than I'd expect.

        • Iksar
          Posted 3 years, 9 months ago (Source)
          @seewhykai I wouldn't focus on the names of the player types or the # of runs in the examples, they aren't important. Just wanted to give examples people might understand.

          • seewhykai
            Posted 3 years, 9 months ago (Source)
            @IksarHS Well the averages seem a bit high to me to have 4 distinct groupings. I don't have access to data on than leaderboards which has been around since 2017 and from what I hear from the community.

            • Iksar
              Posted 3 years, 9 months ago (Source)
              @seewhykai They are just examples of how the numbers would work, that's all. The player-types are irrelevant. I probably should have just left them off to avoid this conversation :o

      • DreadsGaming
        Posted 3 years, 9 months ago (Source)
        @seewhykai @IksarHS I think that is one of the big benefits of this type of a system, it keeps the arena feel of not having skill-based matchmaking

        • Iksar
          Posted 3 years, 9 months ago (Source)
          @DreadsGaming @seewhykai The intention is to keep mmaking as is, though I do worry that keeping mmaking the same is always going to lead to a low-population of players. When you and Kai are playing at 0-0 you are potentially playing against people who just started playing card games. Seems wrong to me.

          • seewhykai
            Posted 3 years, 9 months ago (Source)
            @IksarHS @DreadsGaming I think there are several ways to encourage and inform players about Arena which are not merely to use up tickets or get free packs etc. Many "rules" or basic info regarding Rotations, Leaderboard Seasons, appearance rates are in news blog article posts or hidden in forum posts.

            • Iksar
              Posted 3 years, 9 months ago (Source)
              @seewhykai @DreadsGaming Getting people to try isn't really the problem, it's getting them to try a second run. Not very encouraging to get slammed 3 times in a row by superior players.

            • seewhykai
              Posted 3 years, 9 months ago (Source)
              @IksarHS @DreadsGaming Even high level infinite leaderboard players do not know most this info other than maybe what is in Rotation. The Arena filter on the Card Library page helps when it is correct at least, but there needs to be more info displayed in-game.

              • seewhykai
                Posted 3 years, 9 months ago (Source)
                @IksarHS @DreadsGaming On another note, it feels very wrong when I am sub 3 wins 2 losses and get matched up against a player that t1 coins hp into hp etc only to lose due to card quality. I'd argue it is a much worse feeling as it was not due to skill or experience.

                • Iksar
                  Posted 3 years, 9 months ago (Source)
                  @seewhykai @DreadsGaming I think this is more of a question on what you think deck power disparity should be. If decks have vastly different power, it makes drafting a great deck highly varied and exciting to build something awesome (or disappointed to build something awful).

                  • Iksar
                    Posted 3 years, 9 months ago (Source)
                    @seewhykai @DreadsGaming If the power level of all decks was somehow the same, drafting a deck is a lot less exciting, but the outcome of each game is more dependent on player skill.

                    • Iksar
                      Posted 3 years, 9 months ago (Source)
                      @seewhykai @DreadsGaming Personally I am strongly in favor of having high disparity in power level among each deck. Sometimes you draft a bad deck and do poorly, sometimes you draft an amazing deck and do well. Makes the experience more textured with highs and lows.

                      • Iksar
                        Posted 3 years, 9 months ago (Source)
                        @seewhykai @DreadsGaming The best players still win out in the end because they do better on average, which to me, is good enough at sustaining competitive integrity. I know many players don't feel this way, though, and would rather decks be much closer together in power.

                      • seewhykai
                        Posted 3 years, 9 months ago (Source)
                        @IksarHS @DreadsGaming While I agree, this is more of a card design issue which I am sure the Arena community has lots to say about🤡

                        • Iksar
                          Posted 3 years, 9 months ago (Source)
                          @seewhykai @DreadsGaming Yeah, it does make me sad that whenever there is a very strong card in arena the reaction is.... take it out! No strong cards in arena!

    • VideoBlobby
      Posted 3 years, 9 months ago (Source)
      @IksarHS Why talk about incentivizing class diversity like it’s a problem with the attitude of the player base? I think most regular arena players would love to play all the classes, but the lack of periodic and effective balancing is what prevents that.

      • TrumpSC
        Posted 3 years, 9 months ago (Source)
        @VideoBlobby @IksarHS thanks for tagging me @osmajavi I agree a lot with @videoblobby an arena leaderboard/ranking system would be really cool, I do think more important is that there ideally be a less than 5% winrate difference between best class and worst class.

        • Iksar
          Posted 3 years, 9 months ago (Source)
          @TrumpSC @VideoBlobby @osmajavi win rate balance is extra hard in arena because no skill matchmaking the best players determine for themselves what the best classes are, they play those classes at a higher rate than players who aren't in the know, then the win rate of those classes skews higher

          • Iksar
            Posted 3 years, 9 months ago (Source)
            @TrumpSC @VideoBlobby @osmajavi Class win rates are really much closer together if you only look within buckets of similarly skilled players, hard to communicate that to everyone, though :/

            • kolstIRL
              Posted 3 years, 9 months ago (Source)
              @IksarHS @TrumpSC @VideoBlobby @osmajavi I struggle to believe this conclusion - I find high level Arena players will often have one or two pet classes.. similar to pet heroes in BG's - that they'll do better with. So the bad classes get disproportionately played by people that outperform with the class. Selection bias.

              • Iksar
                Posted 3 years, 9 months ago (Source)
                @kolstIRL @TrumpSC @VideoBlobby @osmajavi Just looking at class win rates only with players at X rating it's very clear to see. Also clear to see the population of the best classes is much higher for good players than it is for bad players. Bad players just don't know.

                • Iksar
                  Posted 3 years, 9 months ago (Source)
                  @kolstIRL @TrumpSC @VideoBlobby @osmajavi Early in arena it was clear for a long time that mage was probably the best, warrior was probably the worst, and everything else was somewhere in between. Nowadays with microadjusts the only people that know what the best classes are are following sites like hsreplay closely.

          • VideoBlobby
            Posted 3 years, 9 months ago (Source)
            @IksarHS @TrumpSC @osmajavi So it’s the players’ fault that some classes seem better because they are taking the classes that are better? Doesn’t this again locate the problem with the player base in a weird way?

            • Iksar
              Posted 3 years, 9 months ago (Source)
              @VideoBlobby @TrumpSC @osmajavi It's not anyone's 'fault', the reality is just that good players win more games than bad players. If good players only pick mage and bad players only pick shaman, mage is always going to have a higher win rate. It doesn't matter what the power level of the two classes is.

          • TrumpSC
            Posted 3 years, 9 months ago (Source)
            @IksarHS @VideoBlobby @osmajavi Is noting "I seem to get bad cards when I pick a low win rate class and I seem to get good cards when I pick a high win rate class" all in my head? 🤔 I'm speaking as someone who tends to purposely choose varieties of classes too. It seems high win rate class = better cards

            • Iksar
              Posted 3 years, 9 months ago (Source)
              @TrumpSC @VideoBlobby @osmajavi This is what microadjustments do, yes. My point is that the engaged community will go to a website and see hmm druid is 52% win rate, I should play Druid. Then because better players play more Druid on average, it goes to 53%, then keep perpetuating itself.

              • Iksar
                Posted 3 years, 9 months ago (Source)
                @TrumpSC @VideoBlobby @osmajavi I know most disagree, but I personally think it's fine (good, even!) to have some classes that are weaker than others in a draft format. Especially if there is incentive for me to do well on a 'weak' class. Pushing every class and every deck to the same power level is 🥱🥱🥱

                • Iksar
                  Posted 3 years, 9 months ago (Source)
                  @TrumpSC @VideoBlobby @osmajavi But FWIW, this is what we'll do if the community wants it. Just a personal pref.

                • TrumpSC
                  Posted 3 years, 9 months ago (Source)
                  @IksarHS @VideoBlobby @osmajavi It becomes more interesting when the class cards appear more often too. It is yawn when classes are "pushed" because everyone gets Deeprun Engineer to fetch Goliath.

                  • Iksar
                    Posted 3 years, 9 months ago (Source)
                    @TrumpSC @VideoBlobby @osmajavi I agree with this. There are some safeguards in place in microadjusments so that each card has a maximum value it can be adjusted upwards, but even with a modest maximum value when those cards are neutrals the format can feel a little samey.

                • VideoBlobby
                  Posted 3 years, 9 months ago (Source)
                  @IksarHS @TrumpSC @osmajavi We don’t expect perfect balance, and we understand what you’re characterizing here— it’s how metagame works for most games. We just think a ~20% skew between highest and lowest winrate class is extreme. Such a wide gap can’t be attributed to player base.

                  • Iksar
                    Posted 3 years, 9 months ago (Source)
                    @VideoBlobby @TrumpSC @osmajavi Not trying to be argumentative, but my point is this isn't how metagames work for most games at all. Most games have skill matchmaking so it doesn't matter what high and low skill players pick, they'll always match against someone of their skill level.
Tweet