In WoW, players are split into strata (Pvp Tiers) 10-20 / 20-30 / 40-50 / 50-60...

This causes the playerbase to split. It creates a rift, players of uneven levels cannot be matched together for gameplay purposes. For example a lvl 1 player would suicide (1shotted) if he dares to attack an AFK druid from his thorn buff (30 min duration) before the autoattack lands.

Another game suffering from same issue is Clash Royale. If players power creep with additional level cap increases. And when more players reach the top strata, the other strata will become isolated.

Example: Low strata 0.1% playerbase. Top stratum 99.9% playerbase.

Matchmaker Constant Bickering: I was matched with a player N levels above me. This is unfair.

  1. If the game was designed today from start, how would game design solve the issue before it ever began.

  2. Is it possible to create "Power gaps" in player base, without splitting it?

  • 1
    \$\begingroup\$ Your problem is that player power scales too much in these games. Fix that, and you fix a whole swathe of issues: griefing, camping, grinding, scaling players to party's levels, areas becoming obsolete, etc. \$\endgroup\$ – Miles Rout Jul 18 '17 at 22:25

Generally such systems use [player_level] +/- Y, where Y is some value that is roughly "fair." E.g. 5.

So a level 15 player is a decent matchup against both a 10 and a 20, even if a level 10 isn't a fair matchup to a level 20. This gives a "floating bracket" around which to match players together.

How wide your bracket is would depend on your game, how important +1 level is, and other factors, which would probably require some testing. There's no easy way to know that the lower-level player has a decent chance against the higher level player as there are a lot of factors: max hit points and damage output are only one. Status infliction abilities, max mana, gear, etc. etc. can all influence the outcome.

Another option (which works best when most players are "top tier") is to use a Chess Rankings-like system. When two players go up against each other, determine their ranking difference. If the higher-ranked player wins, the point change on both is small (and the greater the difference, the smaller that change is). If the lower ranked player wins, the point change is large (and the greater the difference, the larger that change is). For evenly-ranked players the point change isn't very large, but it would be the same for both players (i.e. it doesn't matter who wins, the point shift is still X). The EFC Grading page goes into detail on how to calculate the point shift, but you're free to come up with another algorithm.

So even at the "end game" where everyone is level 100 (the cap in your game) players will have different skill levels and this rating system will match up evenly skilled players. You can also allow any two players to face off against each other, regardless of their rating (although you may wish to give them the option to say "this game is unranked" so they aren't wagering points1).

For lower level players I would suggest a floating bracket with a hidden ranking number underneath. The level is more important during this time, while also having the system track wins and losses in the ranking system, then when the player hits the point at which they're allowed to fight players at the level cap, their ranking can unhide and be a better determination value. By tracking their ranking up to this point you don't dump them into the Big Kids pool with a default ranking score: afterall, you could have data about their performance up until now, so you may as well track and use it.

1Off-topic: back in highschool one of the math teachers ran a chess club during lunch. Against one friend I almost always won games that were ranked against him and almost always lost when unranked. Irked him pretty bad, as he was ranked quite a bit higher than I was.


Make your PvP gameplay level-agnostic by using different game mechanics in PvP which negate level advantage.

When you designed your game mechanics properly, you should have a progression formula which tells you how much hit points, DPS, etc. a level X character with level X appropriate equipment has on average. You will need such a formula in order to balance your PvE content properly.

Now you just have to use your power-by-level formula to adjust the character stats in PvP up or down depending on character level. That way even a level 5 character is mechanically on-par with a level 87 character. The combat will be decided by strategy, tactics and mastery of the game mechanics, not just on who spent more time grinding.

If you are worried that this will go completely against the idea of character building in an RPG: Remember that characters can still be above or below the power-by-level curve due to smart character building decisions and luck with equipment drops. So characters will still have different power in PvP. Just that they won't depend as much on level. Also, higher level characters will usually have access to more abilities, which gives them more flexibility (even if those abilities aren't mechanically stronger due to above described level normalization).

This is essentially how most non-RPG competitive multiplayer games with character progression work: Reward progress with access to more weapons and abilities, but don't just make characters nominally stronger.

Possible pitfall: If your game allows low-level characters to use high-level equipment or receive high-level buffs, their power will greatly exceed the average power on their level. That can make them extremely overpowered in PvP. The easy solution would be to simply disallow that (equipment gets level requirements, buffs received in PvE don't work in PvP). The harder solution would be to also have level-scaling rules for equipment and buffs in PvP.

  • \$\begingroup\$ Note: this approach is called "normalization." Some mechanics (such as allowing the player to choose where to put skill points) can make normalization hard, or even impossible. Additionally, players may dislike normalization, particularly high level players who want to play low level content and just smash right through it. I am involved in the active development of a game where normalization was A Standard Thing and I thought it was a great boon, but actual players hated it. It's since been removed and relegated to (ultra rare) event challenges. \$\endgroup\$ – Draco18s no longer trusts SE Jul 17 '17 at 19:08

For a current project, the approach I'm considering is to have lateral level progression. A low level maximum, and progression for players actually places them at a lower level. I haven't had a chance to implement or test this yet, so it's more a thought experiment at the moment.

Player experience is divided into three parts.

  1. New to Game
  2. New to Aspect (class, mechanic, etc.)
  3. Mastery of Aspect

And apply a set number of levels for each part, I'm starting with 10 for each.

Level 1-10 are content for getting users used to the game both from a mechanics and story, as well as a social aspect (knowing terminology, how social features work, etc.)

At this point they can choose a specialization, in the form of a class for my use.

Then levels 11-20 is introducing different mechanics of that specialization. So all mechanics are available at level 20.

From levels 21-30 they're given customization choices for the specialization. This allows them to master that specialization.

At any level between 21-30 they can choose to change to a new specialization and this brings them back to level 10 or possibly 15 with some of the new mechanics of the new specialization.

With this, stat difference between the core of players, level 10 - 30, can be kept within a smaller range, and advantages for players may be found more in what specialization they have and how they use it.

This also allows for expansion of the game without having to bump the level cap with each addition.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.