Developer Diary

Matchmaking and Related PvP Changes

Hello Wizards!

Ben "Ratbeard" Durbin
Lead Designer
Esteemed PvPers, lapsed and future PvPers, and the PvP curious:

Today we’d like to give you some more details about the changes coming soon to PvP Matchmaking, as well as an update on some PvP-related changes that we will be rolling out in the near future.


Generally speaking, unless specifically mentioned otherwise, the changes we will be discussing here apply to 1v1 Ranked PvP. Players frequently ask, “What about team PvP?” and/or “What about tourneys?” and while we agree that these are important PvP features, our first and primary focus is on getting the fundamentals of PvP into good working order – and that means starting with 1v1.

Matchmaking is handled by a Matchmaking Server. The Matchmaking Server waits for players to attempt to find a match via the kiosk, and this prompts the Matchmaking Server to either add that player to its existing queue or “spin up” a new queue. The Matchmaking Server does not create a queue on any particular schedule, but on an as-needed basis. Once a queue is created, the Matchmaking Server waits for a specified period of time and keeps adding players to the queue until it is time to create matches and prompt players. This waiting period can be set by us, and we will modify it as needed to best serve the players. The waiting time can’t be too long, or players will get frustrated waiting for a match; and it can’t be too short, or the player pool will be too small to create good matches. We can also force the Matchmaking Server to start making matches as soon as it has a certain number of players in the queue, and we can force it not to start making matches (extending the queueing period) if there aren’t enough players in the queue.

When the queuing time period is complete, the Matchmaking Server begins making matches. It starts by sorting all players into a list, following a specified order of sorting criteria (see below), and then it begins making matches by just going down the list – it offers a match to player 1 and player 2, player 3 and player 4, player 5 and player 6, and so on, until it reaches the end of the list. Each pair of players gets their arena prompt at the same time, and, assuming they accept, off they go. Players who don’t accept their match, who don’t click in time, and the “odd man out” are simply moved over to the next queue to continue waiting.

Sorting Criteria

TIER - Tier is the topmost sorting criteria; a broad grouping of characters defined by their access to certain combat mechanics. Specifically, we will group players into Tiers based on whether or not they have access to Crit, and whether or not they have Shadow Rating.
  • Beginner/Novice (Tier 1) - This includes all characters with 0 Crit Rating and 0 Shadow rating.
  • Intermediate/Legendary (Tier 2) - This includes all characters whose Crit Rating >0 OR Level 60+ and whose Shadow Rating = 0.
  • Expert/Exalted (Tier 3) - This includes all characters whose Shadow Rating > 0 OR Level 100+.
As some of you may have noticed, players have some ability to “manipulate” their Tier – but not really in an advantageous way. There’s low level gear available that may give you “early access” to Crit or Shadow and, if you equip this gear when you queue up, we’ll sort you into higher Tiers, as appropriate. In any case, when you queue up, we’ll tell you what Tier you’re queueing for, so you’ll have a chance to change your gear before you queue.

Because the Matchmaking Server is sorting players according to Tier, it is still possible for two players to be matched across Tiers – but there should only be one such match per queue and Tier pair. For example, in any given queue, there should only be ONE match between a Tier 1 and Tier 2 player, and only ONE match between a Tier 2 and Tier 3 player. When this happens, the player from the lower Tier will receive a prompt allowing him to opt in to that match, or to continue waiting in the next queue.

RANK – Players within each Tier are sorted by Rank. Ranks are earned by achieving a certain amount of Rating, and unlock the Rank titles that PvP players are already familiar with: Private, Corporal, Sergeant, Veteran, Knight, Captain, Commander, Warlord. Although we may change the way that Rating is calculated for 5th Age (see below) we will keep the same Rank titles.

One change we are considering is that Rank can only be gained, but not lost; that is to say, within any given PvP season, once you have achieved enough Rating to qualify for a new Rank, you will not lose that Rank even if you lose Rating. This means, for example, that once you have achieved Warlord Rank, you will be sorted and matched with other Warlords. You can’t down-rank in order to get easier matches. However, we’ll likely implement some kind of automatic down-ranking each time we launch a new Season of PvP. (A Season allows us to reset the leaderboards, adjust Rank and Rating inflation, perhaps address some spell balance and other issues related to the meta, but stops short of the sweeping changes that define a new Age.)

When players of two different Ranks are matched, as with Tiers, the player from the lower Rank has the option to take the match or continue waiting in a new queue.

RATING - Rating is an integer value that represents a character's Win/Loss record. For the remainder of 4th Age PvP, we will continue to use the existing Rating system; be advised that 5th Age PvP may transition to a different Rating system (such as Elo).

LEVEL/BAND - The character's level and level band (a group of 10 levels, eg level 1-9, 10-19, 20-29, etc.)

Tier, Rank, Rating, and Level Band are the major sorting criteria and almost all matches will be created (and in all likelihood, determined...) based on these criteria. However, we have defined some additional, more “creative” sorting criteria to serve as a tie-breaker.

STREAK - Streak records the number of matches the character has Won within the last few matches. The Streak criteria very specifically looks at “How good are you playing, lately?” In short, players who have a longer streak of Wins will tend to be matched against other players with a long streak of Wins. For example, if you have won 7 of your last 10 matches, you’re more likely to be matched against other players with a similar streak.

MATCHES PLAYED – This criteria will tend to match players who have played a lot of matches (regardless of outcome) against other players of similar experience. This criteria in particular is likely to be saved from Age to Age with no reset (for as they say, there’s no substitute for experience).

LAPSED AGE - This criteria counts the number of days since any character on the account has played a PvP match. If you haven’t played in a while, you’ll be sorted a little higher in the queue to try to get you a match sooner. Welcome back!

TIME IN QUEUE – We've already addressed a few of the ways that a player can fail to get a match and get sorted into a new queue. This criteria creates a preference for players who have been waiting the longest.


“Boosting” is the act of trying to manipulate the Matchmaking Server or the current queue to try to get a match against a known opponent – or even against yourself – in order to have the other player throw the match so that you gain Rating. Needless to say, we frown on this behavior, and we have implemented several counter-measures to prevent boosting. For obvious reasons, we won't go into detail on these measures.


“Setting” is the act of trying to manipulate the queue so that you are matched against a specific player in the queue, and then “setting” your gear against that specific player to gain an advantage. Note that we don’t have a problem with players setting against a specific “type” of opponent – for example, if you tend to do well against every opponent but Storm, it’s perfectly acceptable to build a Storm-busting set of gear and to equip it in anticipation of such a match. It’s not ok to wait until you know that Wolf Stormblade is in the queue and then change your gear setup to set against him specifically.

To help address the issue of Setting, we will “lock” your gear setup when you queue for a PvP match. Once you enter the queue, you can freely change the spells in your Deck, but you cannot change the equipment you have in your slots. We will save the equipment that you have on when you queue for your match (think of it as a custom Equipment Set), and we will equip that same set of gear before the match starts. (If you do something silly like sell a piece of gear before the match starts, you’ll just have an empty slot. Don’t do that.)


Before we discuss the measures we’re taking to address Stalling, let’s first be clear on our definition of the term. KI considers “stalling” to be playing a match on full defense, using resist and/or healing to keep the match going forever, with no intention of mounting an offense, until your opponent is forced to flee the match. Stalling does NOT refer to “tanking” or slow-playing a match, playing defense until you are able to come back and win the match. But the emphasis here is on WINNING the match - and winning requires attacking.

We liken PvP to a boxing match, and we consider all of the following to be legitimate winning strategies:

Playing like Mike Tyson, hitting hard and fast and going for a knockout

Playing like Leonard or Mayweather, focusing on a frustrating defense while at the same time throwing enough punches to win by decision

Playing like Rocky Balboa, taking tons of punishment until you can “get up off the mat” and knock out your opponent

Ok, so that last boxer is fictional – but our goal is to ensure that every school can win matches by at least one of the strategies listed there. They all have their place. But there is no place in our PvP plans for stalling until your opponent has no choice but to flee the match and give you the win. We consider that griefing, and it will not be supported or tolerated.

If any school lacks the necessary spells in their arsenal to win a match, we will address them on a case by case basis. (Notwithstanding the stalling issue, we consider the careful curation of the PvP meta to be an essential and never-ending task.)

Many players have suggested that we simply “cap” Resistance, but this 'solution’ falls disproportionately on the Ice school, and directly attacks the foundation of thier school identity. Ice is as reliant on high Resist, above the ‘cap’ that other schools can achieve, as Storm is reliant on high Damage, above the ‘cap’ the other schools can receive. Any stalling change built around Resist would have to be a proportional system, not a cap. Changes along these lines are in fact in progress, but under the auspices of the game-wide stat rebalance.

More to the point, it is our belief that the problem isn’t the high Resist, it is the behavior; that it is the intent of the player with high Resist that matters. So our proposed solution attempts to curtail that behavior. For 1v1 Ranked PVP, we will implement the following changes:
  • Reduce the critical multiplier on healing spells
  • Implement a time limit for matches
  • Streamline Diego’s decision criteria to favor aggressive play

Reduced Critical Healing Multiplier

First, let’s be clear that we are currently considering this for PvP only. PvE should not be affected by this change.

This change brings the critical multiplier for healing into even parity with the critical multiplier for damage. Although it is important that healing serve as a sort of “universal counter” to damage dealt, the prior multiplier was too high, and more importantly so much higher than the damage multiplier that it forced matches into a “one shot” paradigm, where if the attacker failed to defeat his opponent in one shot, the defender could reliably count on healing back to full health.

It is important to note that, critical hits notwithstanding, “healing per pip” already greatly outpaces even the best damage-per-pip in the game: roughly 190 healing per pip vs. 130 dpp for Storm shadow-enhanced spells. HPP is so high because a solo player needs to be able to have enough healing to heal back the damage output of the typical 2-mob sigil. A player geared for healing also has incoming and outgoing healing gear (offset by the attacker’s damage gear), and attackers have no way to offset the defender’s gear bonuses. After a deep dive into our item data – while recognizing that the defender must sacrifice other stats in order to optimize his healing setup – we believe that even after reducing the critical healing multiplier, defenders with a healing focus can still “outheal” the best damage setups in the game. (Certainly this is true of ‘typical’ damage vs. Healing setups.)

Match Time Limit

The two most important things to note on this change is that (a) we are open to feedback regarding the optimal time limit for matches and (b) we must enforce some kind of time limit for matches; they can’t simply go on forever. As of this writing, our intent is a 30 minute limit on 1v1 ranked matches.

We realize that some players might enjoy longer matches – and to the extent that such matches offer interesting play-counterplay action, we might be inclined to agree. But from the developers’ standpoint, we will favor shorter matches, so that even players with a limited amount of game time feel as if they have enough time to queue up and finish at least one PvP match in a typical play session. If players know that a PvP match might take hours to complete, they will be discouraged from even entering the PvP queue in the first place. We hope that PvP players will agree that making PvP more accessible to more players will ultimately be best for the future of PvP.

Decision by Diego

At the end of the Match Time Limit, if there is no clear winner, Diego will decide the match. Of note here is that this is a much more strict and streamlined Diego. While we do not want to go into great detail as to how Diego will decide matches, we will say that (in keeping with our boxing metaphor from earlier) he will primarily be looking to determine the aggressor in the match, and awarding the win accordingly. This “new” Diego is hopefully improved over the version you may be familiar with from other PvP formats, but we are open to feedback here as well.


The changes discussed here will be implemented incrementally over the next few updates, and your feedback is critical as each change goes live. All feedback is welcome, but we are particularly interested in the following:

Stalling Feedback

  • If you use a healing strategy, are you able to heal back damage as quickly as the opponent can deal it? What kind of healing gear setup do you use?
  • Is the match time limit sufficient for most matches to reach a successful conclusion without going to a Decision by Diego?
  • If your match goes to a decision, does it feel as if Diego ruled in favor of the more aggressive player?

Matchmaking Feedback

  • Does the wait time in queue feel about right?
  • If you feel you are waiting too long, how long are you typically waiting?
  • Do you feel you are being matched appropriately? (Tier, Rank, Rating, Level)?
As always, please remember that we are as eager as you are to relaunch PvP with a brand new 5th Age – but we caution that “the last days of 4th Age” are likely to be chaotic and messy while we work – together – to find the right solutions to various PvP problems. Frustration is understandable (and probably inevitable) but this is an exciting time! Join us in thinking deeply about PvP problems and flex your creativity. We’re listening!


Ben "Ratbeard" Durbin
Lead Designer

Matthew "Mattnetic" Fahey
Senior Systems Designer II

Leah "Professor Falmea" Ruben
Lead Producer

Dated: September 3, 2020