~chron

As you might have guessed I’ve been having quite a bit of fun with the new armory features I mentioned last week. I toyed around with doing some reports based on the data behind the Opponent History page, but for whatever reason it seems like it’s missing a lot of games compared to the Match History page. The Match History page has better data anyway, I was just hoping to save myself some HTTP requests since one for each game adds up pretty quick.

One problem with the new pages is they don’t seem to be updated very frequently – it’s possibly done with the weekly reset, since neither of my teams seem to include any of the games we’ve played this week – my rogue in particular played close to a hundred games. I’ll definitely do some updated stats once this week’s games are included.

So anyway; the questions I wanted to answer with this investigation are:

• What compositions do we have trouble with?
• What classes in particular are we losing a lot against?
• What are our win/loss ratios like on different maps?

I could hazard a guess at all of these of course, but you can’t beat hard data! I’ve elected to include every game at the moment, but it would probably be more useful to cut off games below a certain rating threshold – the two teams I’m playing in actively have been used the entire season and there’s been a degree of discovery with the comps. In fact, both of my partners are new to their specs this season, so it would probably make sense to ignore some of the “training” games.

That said, here’s the data for my rogue/rshaman team. The columns are, in order: wins, losses, total games, win ratio, average match length (seconds).

Total games played: 266 (145 wins, 121 losses, ratio: 0.545).

                 Hunter/Priest =>   5   0   5  1.000  190.20
                  Druid/Shaman =>   2   0   2  1.000  231.00
                  Mage/Warlock =>   1   0   1  1.000  192.00
            Death Knight/Rogue =>   1   0   1  1.000  190.00
                Priest/Warlock =>   1   0   1  1.000  139.00
                Hunter/Paladin =>  15   1  16  0.938  323.69
            Death Knight/Druid =>   3   1   4  0.750  130.50
               Paladin/Warlock =>  10   4  14  0.714  222.21
                 Hunter/Shaman =>   5   2   7  0.714  257.86
                Paladin/Shaman =>   6   3   9  0.667  334.44
                  Rogue/Shaman =>   4   2   6  0.667  255.50
                Paladin/Priest =>   2   1   3  0.667  201.00
                Shaman/Warlock =>   3   2   5  0.600  232.60
                 Paladin/Rogue =>   3   2   5  0.600  265.80
                    Mage/Rogue =>  17  12  29  0.586  156.31
           Death Knight/Shaman =>   4   4   8  0.500  264.00
                   Druid/Rogue =>   3   3   6  0.500  189.50
           Death Knight/Priest =>   3   3   6  0.500  183.50
                  Druid/Priest =>   2   2   4  0.500  289.00
                 Druid/Warlock =>   1   1   2  0.500  316.50
          Death Knight/Paladin =>  20  21  41  0.488  265.54
                    Druid/Mage =>   5   6  11  0.455  218.64
                  Druid/Hunter =>   4   5   9  0.444  251.11
                  Mage/Paladin =>   8  11  19  0.421  163.95
               Paladin/Warrior =>   7  10  17  0.412  322.24
                  Priest/Rogue =>   6  12  18  0.333  204.78
             Death Knight/Mage =>   2   4   6  0.333  155.33
                 Rogue/Warlock =>   1   2   3  0.333  181.67
                 Druid/Paladin =>   1   5   6  0.167  350.50
                   Mage/Shaman =>   0   1   1  0.000  294.00
           Death Knight/Hunter =>   0   1   1  0.000  192.00

They’re ordered firstly by win ratio and then by total games played. Here’s a look at the individual classes:

                        Hunter =>  29   9  38  0.763  273.34
                       Warlock =>  17   9  26  0.654  222.42
                        Shaman =>  24  14  38  0.632  273.13
                       Paladin =>  72  58 130  0.554  267.81
                         Rogue =>  35  33  68  0.515  190.49
                        Priest =>  19  18  37  0.514  206.38
                          Mage =>  33  34  67  0.493  171.21
                  Death Knight =>  33  34  67  0.493  237.85
                         Druid =>  21  23  44  0.477  242.68
                       Warrior =>   7  10  17  0.412  322.24

Warriors win at drawing games out, who knew? Finally here’s a breakdown by map:

                 Nagrand Arena =>  40  22  62  0.645  274.13
                Dalaran Sewers =>  33  26  59  0.559  229.44
            Ruins of Lordaeron =>  25  22  47  0.532  230.77
            Blade's Edge Arena =>  34  35  69  0.493  224.87
             The Ring of Valor =>  13  16  29  0.448  202.17

So, some quick observations:

• Warrior teams are hard but there aren’t a lot of them and they only really have one viable comp.
• There are a LOT of DK/Pally teams out there, and Mage/Rogue is pretty highly represented as well. I should probably do some graphs of appearances of a certain comp versus rating because that gives you a good idea of what you need to be able to beat to progress up the ladder.
• Mage games, Rogue games, and Mage/Rogue in particular, are over really quickly.
• Priest/Rogue is the “popular comp” that we have the hardest time with. Especially when they both abuse engineering, which seems to be everybody this week.
• Org arena is really crappy, I’m ecstatic they took it off the live realms.

There’s a lot of other data you could look at – damage done or taken versus different compositions, how many killing blows the healers manage to steal (spoiler: Azuba has over a billion damage done so you better fucking watch out). This is the sort of stuff I always wanted to do with ArenaHistorian but never quite got around to!

chronic Arena, Data Mining, Rogue, Shaman

On my rogue I’ve been doing 2s with Azuba as resto, so we’ve been talking a fair bit about shaman gearing and different spec possibilities recently.

I ran a similar report to the paladin one I did the other day with the current top 200 Reckoning 2v2 teams – restoration shaman only. I won’t do any analysis since I don’t know that much about shaman, but I thought I’d post the output here anyway in case people were interested.

Here are the results!

As a bonus there are Google Chart URLs with each of the statistics if you prefer to consume your data in pie form. Here’s the “weapon equipped in main hand” one for your viewing pleasure:

Note that I have completely excluded “PvE geared” (ie, no Medallion) shaman from the report. Here’s one more pie, the partner class. Looks like shaman have a few viable choices:

chronic Data Mining, Shaman Arena, armory.rb, gchart, pvp, Shaman

I’ve been thinking recently about re-gemming for survivability on my Holy Paladin, since I’m only at about 550 resilience in my current gear. It’s not every fight that I feel like I need it, but I’m pretty happy with my throughput (healing is only really a problem if I get CCd or school-locked) and mana usage. I have about 18.5k mana and either JotW or drinking (depending on the comp) in concert with Divine Plea mean we very rarely lose on mana to a team that isn’t aggressively mana burning.

(The deadly libram will also be a great throughput upgrade, I just wanted to replace all my blues first!)

So, as always, let’s go to the Armory and see what high-rated Paladins are doing. My sample this time around is Paladins from the top 200 2v2 teams on Reckoning, and I’ve only included people wearing pvp gear (medallion or human with >100 resil) at the time of the survey. I’ve also explicitly excluded meta gems since there isn’t really any contention there: int + manarestore all the way.

Here’s the initial results. I’m assuming the strength gems are from people caught in their ret gear (but still as holy spec), but I didn’t actually look them up. I’ve also excluded the quality of the gems, since it didn’t really seem important for my purposes. The data set consists of 75 characters between ~2750 and ~2140, with 784 total gems.

                                         Intellect => 149 | 20.93%
                                       Spell Power => 113 | 15.87%
                                 Resilience Rating =>  96 | 13.48%
                         Spell Power and Intellect =>  54 | 07.58%
                 Spell Power and Resilience Rating =>  51 | 07.16%
                           Spell Power and Stamina =>  49 | 06.88%
                     Resilience Rating and Stamina =>  44 | 06.18%
            Spell Power and Critical Strike Rating =>  41 | 05.76%
                             Intellect and Stamina =>  29 | 04.07%
                            Critical Strike Rating =>  19 | 02.67%
                                           Stamina =>  18 | 02.53%
              Spell Power and Mana every 5 seconds =>  18 | 02.53%
                Intellect and Mana every 5 seconds =>  10 | 01.40%
                                      Haste Rating =>   5 | 00.70%
                Critical Strike Rating and Stamina =>   5 | 00.70%
                      Spell Power and Haste Rating =>   3 | 00.42%
                                          Strength =>   2 | 00.28%
                              Strength and Stamina =>   2 | 00.28%
                                        Hit Rating =>   2 | 00.28%
             Haste Rating and Mana every 5 seconds =>   1 | 00.14%
               Strength and Critical Strike Rating =>   1 | 00.14%

We can do some basic categorization of the gems, too. I’ve divided them into throughput (spell power, crit, haste), longevity (intellect, mp5), utility (hit rating) and defense (stamina, resilience). I realise these categories aren’t entirely accurate because most paladin stats affect multiple areas (for example, via Illumination), but it’s a good starting point.

                                        throughput => 181 | 25.42%
                                         longevity => 159 | 22.33%
                                           defense => 158 | 22.19%
                                defense/throughput => 105 | 14.75%
                              longevity/throughput =>  73 | 10.25%
                                 defense/longevity =>  29 | 04.07%
                                             other =>   2 | 00.28%
                                           utility =>   2 | 00.28%
                                     defense/other =>   2 | 00.28%
                                  other/throughput =>   1 | 00.14%

Then maybe it’s useful to see totals for each category, if you count a “pure gem” as 2 points for its category and the hybrid gems for 1 point in each:

                                        throughput => 541 | 37.99%
                                           defense => 452 | 31.74%
                                         longevity => 420 | 29.49%
                                             other =>   7 | 00.49%
                                           utility =>   4 | 00.28%

From looking at that it seems most people are balancing their gemming across the three major categories, and browsing through the individual results seems to support that. Let’s see one last view over the data, since this is starting to get long.

This table contains a row for each paladin, with a column for each of the three major categories defined above – the cells show what proportion of that paladin’s total gems are used on that category. Hybrid gems are split across both categories as in the previous report, and strength/hit gems are ignored although they do still count in the totals.

          name throughput  longevity    defense
    Varluneyna       0.33       0.33       0.33
        Heimen       0.45       0.35       0.20
         Jeuce       0.38       0.50       0.13
        Haydee       0.39       0.11       0.50
    Unicornica       0.50       0.50       0.00
       Wcstyle       0.60       0.10       0.30
      Mchaggis       0.45       0.23       0.32
        Aiiden       0.44       0.38       0.19
        Braves       0.36       0.45       0.18
          Gina       0.50       0.22       0.28
        Dercas       0.43       0.43       0.14
     Mournhold       0.71       0.29       0.00
       Beldara       0.39       0.28       0.33
         Oryun       0.59       0.14       0.27
      Dimarius       0.56       0.44       0.00
      Tygrilol       0.50       0.17       0.33
       Sasorii       0.38       0.63       0.00
   Solidarityx       0.46       0.46       0.08
        Ieetoh       0.10       0.00       0.20
       Reepent       0.25       0.50       0.25
      Harumaru       0.55       0.45       0.00
      Magicpie       0.39       0.39       0.22
        Pummie       0.44       0.56       0.00
     Holynight       0.44       0.33       0.22
   Redemptionz       0.44       0.39       0.17
     Derreidos       0.41       0.59       0.00
     Femaleftw       0.44       0.28       0.28
     Diplomacy       0.50       0.50       0.00
           Qiu       0.65       0.15       0.20
        Gellin       0.45       0.55       0.00
        Stylec       0.25       0.50       0.25
         Aurog       0.35       0.45       0.20
         Kaara       0.40       0.20       0.40
      Mcjordie       0.50       0.35       0.15
        Arasof       0.50       0.06       0.44
  Bubbliciouss       0.61       0.28       0.11
   Eskimopally       0.50       0.25       0.25
         Zyrxy       0.38       0.50       0.12
   Lostprophet       0.56       0.17       0.28
       Valefor       0.40       0.35       0.25
     Sylvanisa       0.44       0.33       0.22
       Luncarn       0.30       0.70       0.00
     Feladence       0.50       0.50       0.00
    Smokealarm       0.39       0.33       0.28
         Ermad       0.28       0.72       0.00
   Appocalypse       0.50       0.11       0.39
        Macabe       0.38       0.63       0.00
        Fiaraa       0.50       0.35       0.15
         Trawn       0.56       0.19       0.25
   Moomoobeezy       0.38       0.38       0.25
        Oculis       0.19       0.56       0.25
   Espectrolol       0.50       0.30       0.20
      Romulisa       0.50       0.28       0.22
         Clint       0.38       0.42       0.21
       Kehrsyn       0.40       0.00       0.60
        Sayami       0.29       0.71       0.00
         Jdubz       0.31       0.38       0.31
    Shadgavena       0.55       0.14       0.32
         Serjj       0.65       0.25       0.10
      Judgmant       0.36       0.64       0.00
   Crispybacon       0.30       0.30       0.40
          Guli       0.31       0.44       0.25
  Annihilacion       0.64       0.36       0.00
       Clapton       0.44       0.44       0.11
         Zatin       0.44       0.56       0.00
        Tranze       0.38       0.63       0.00
      Brianlee       0.23       0.77       0.00
       Makitoo       0.45       0.55       0.00
          Fron       0.64       0.23       0.14
       Zenakuu       0.44       0.39       0.17
         Duner       0.36       0.64       0.00
          Evir       0.78       0.11       0.11
        Zorgat       1.00       0.00       0.00
       Svendor       0.63       0.38       0.00
          Mute       0.60       0.25       0.15

They’re sorted by rating, if you’re wondering.

chronic Data Mining, Paladin armory.rb, gems, Paladin

I like reading the forum threads on AJ – there are a lot of good players on there, and when a few 2400 rated people agree on something you can generally assume they have the right idea. Still; it’s a small sample size – and a self-selected one at that – so I find it useful to sanity-check my decisions based on what successful people are actually doing.

So we turn to the official Armory. I’ve mentioned my data-mining projects in the past, and my current focus is on the Reckoning 2v2 bracket. I wrote a small script to have a look at all the holy paladins (you’re holy if you have more points in the holy tree than any other tree) in the top 300 teams so I could get some “real” data about what kind of gear people are using, how they gem, how they spec, all that good stuff.

Rather than attempt a hideously wide HTML table, here’s the output of the script as a text file. The second half is the statistics, so be prepared for some scrolling. There’s also a CSV file of just the non-PvE characters.

First thing to note – a number of those characters are flagged by the script as “PvE” (the exclamation mark) which discounts them from many of the statistics. I consider a character to be wearing PvP gear if she has any of the various medallions of the horde/alliance equipped or if she’s a human with at least 100 resilience across all her gear. I realise that isn’t a perfect test, especially in a world where paladins can afford to wear a lot of PvE gear, but it works well enough for my purposes.

Second thing – if there’s more than two people on the team, the script will take the two with the highest number of games played so far this season. In practice most of the teams just have two serious players, but I’ve probably missed a few people out because of this.

Representation overall is as good as you could hope for – 101 paladins out of the 300 teams is something like 17% which is absolutely crazy for a single spec. Paladins are the dominant 2s healer, no doubt about that.

In terms of partners for a Holy Paladin, Death Knights are far and away the most popular. Unholy (specifically Shadowfrost or 0/20/51) continues to be the dominant spec (46.5%), though the numbers of Frost and Blood DKs are slowly creeping up. Having said that, Paladins still have a fairly wide variety of potential partners at the moment. To some extent its just a case of putting any two overpowered classes together for even more overpowered-ness, but Survival Hunter, Fury Warrior, and to a lesser degree Rogue, Warlock, and Feral Druid are all very viable options.

The holy trinity of Paladin PvP specs is well established, with the hybrid repentance spec by far the most popular now (42.25%). Interestingly there’s quite a bit of “long tail” variation – about 24% of the paladins are NOT using one of the cookie-cutter specs, even just looking at point-distribtion alone. The faction ratio is pretty balanced at 50.7% alliance.

In terms of gear, the average resilience is ~565, which speaks to the strengths of Paladins’ innate defenses in high-burst environments. A similar thing happened in Season 1 and I will be curious to see how further LK seasons shift this balance.

I am on there, but pretty near the bottom :<

chronic Arena, Data Mining, Paladin