Let's Fix: Masteries

Masteries

What are they?

League of Legends players can edit their Runes and Masteries and Runes to customize any champion as they see fit. They provide maybe a ~5% boost to a champion's early game power, waning to somewhere around 2% later on. Since everyone at level 30 has equal access to masteries (and to a lesser extent the runes) there's not really a problem with balance. If the Offensive page is OP then everybody uses the Offense page and nobody gets a real advantage.

Luckily they are balanced enough that even that is not an issue. Your choice of Masteries is usually a reflection on your personal preference of playstyle. Some champions are a bit limited in their selection: an AD carry for example will run 21/9/0 with little variation. Others however get a bit more choice in the matter. As Warwick, I might run down the Defense tree if I plan to try to outsustain the other guy and transition into a tanky initiator, or I could run a more Offensive build if I plan to gank and go all in for the kill.

The Problem

The first big issue is that the current Season 2 Masteries are incredibly fiddly. Wide swath choices mean a lot (21/9/0 vs 9/21/0 can result in very different jungle paths or lane behaviors), but individual point assignments feel unimportant and nitpicky. They aren't meaningful choices. Brute Force provides 1 AD, while Butcher grants +2 damage to a minion or monster per auto attack. Over the course of the game this could perhaps result in one fewer hit on a jungle monster or one more last hit in lane. Over the course of 100 games it might result in a single champion kill that would have otherwise escaped at 5 HP. When presented with 30 points each point should be relevant.

The options presented are also boring and min-maxey. For an AP Carry there are a grand total of 6 Masteries spanning 17 points in the Offense tree which read "deal more magic damage". Why force the player to spend more than half their mastery points to say that?

Rarely do you see any builds with 30 points in a tree, the vast majority instead straddling 21s and 9s. This is because The only real choices are the 6-7 "final" abilities (those available at points 9 and 21). Do you take Executioner and Veteran's Scars, or Juggernaut and Runic Affinity? Being generous, there's maybe 8 different Mastery builds to choose from, with a few rogue builds out there besides. All the other 28 points are just busywork to get you to there. That is, AD Executioner is a 12 point Mastery which includes things like +6 Armor Pen and +1/2 AD/Level. No choices here.

Any of the real decisions are often very mathy. I can run the calculation to find out if .5% more damage is equivalent to 1/8 AD per level for Caitlyn or Ashe, but the answer is always "it's close enough that it doesn't matter".

As a minor issue, Summoner Spells do not need associated Masteries. Choosing that spell is customization enough, there's no need for a second step which only serves to blur the power levels and make you feel bad if it's out of your tree.

The Shape of the Solution

The fix should be first and foremost intuitive. It should create meaningful choices, such as "deal more damage", "increase lane sustain", or "be a jungler who concentrates on ganking". While customization is important, the real decisions which players remember occur during the game, not before it.

A player should never feel like he or she can't proceed because of an incorrect initial loadout, but he or she should feel like their Masteries give them the edge which synergizes with their preferred build or playstyle.

The end result should provide no more than a ~5% boost to champion power in any direction, but balance comes through playtesting not theorycrafting so that's not being considered at this stage.

The Solution 

First, we reduce the max Summoner level from 30 to 18. It's 18 in the game, why the unnecessary distinction?

Then, alter the existing Masteries pages dramatically to mimic champion abilities. That is, in each tree there are 4 columns - 3 basic Masteries and 1 "ultimate". We can keep the tree names for lack of better terms (though I could imagine something like Carry/Support Tank/Jungler, but let's leave that for now).

Each basic Master column would have 5 ranks while each ultimate column would have 3, just like in the game. And you could rank them just like in the game. For example, you can't reach rank 5 in a column until level 9 since you need to alternate level. The first rank of an ultimate in a tree would be available at once you've invested 5 points elsewhere in that tree, the next after 6 additional points, etc.

Above we have a crude representation of how this would work. You are investing points in your tree and have (at least) one more point available to spend, hence the red plus signs where they can be spent. You're putting a bunch of points into your Utility tree, so you can rank its Ultimate column one more time, or you could continue to rank your Offensive basic abilities instead.

What might these columns look like? The offensive tree is probably the easiest to imagine. One tier would be your basic AD stuff, the next AP, the third some sort of utility-based thing. Perhaps for jungling, perhaps for last hitting, perhaps for summoner spells. Possibly something aligned for people who want to roam the map rather than lane, granting things like move speed or brush interactions.

The ranks could present new additions without branching, too. For example, say we look at the first rank AD column in Offense. It grants +1 AD. The second rank grants an additional +1AD, but also +1% Attack Speed. Third, you get another +1AD, another +1% Attack Speed, and now you get +2 Armor Penetration. Fourth rank gives access to +1 AD/Level, 5th gives increased Critical Strike damage. Meanwhile, the AP column grants similar equivalents.

The Defense tree could have columns for Tankiness/Sustain/Team Support. The first would include things like health and resistances, the second grants regeneration and the like, while the third provides auras and initiation power. Currently the Defense tree has choices like Armor vs Magic Resist, but those can be handled by Runes without Masteries needing to but in.

Utility would be more focused on things that aren't damage or defense.  This is a wide net, including things like GP10, wards, experience gain, etc. I'll admit I'm not sure what the 3 trees would look like, but at a glance I might offer Offensive Support/Defensive Support/Map Control Support. The first two would dictate how you'd like to play your lane (strong poke, sustain, ability to protect your ally), the third would be a favorite of roamers and warders. There could be an ability in the first tree with grants 1 gold every time you damage an enemy champion, one in the second tree which grants 1 gold for nearby creep deaths, and something in the third which grants gold for buying wards or traversing the map.

Ultimates in each tree would be usually where the brunt of boring power comes from, since they're not decision based and so don't need to be too clever with them. At a glance I might suggest +5% damage, +5% HP/resists, +5% move speed.

Conclusion

At first glance this model seems less customizable. But that's not true. As a Carry I might push all 18 points into Offense to just deal more damage. However, I might opt for some survivability and invest in in the +5% HP, requiring 6 points in Defense. I could choose in that tree to go 3/1/1/1, or I could go 0/2/3/1 instead, depending on preference. Before I had ~8-10 loadouts to choose from, based on the available end-points. Here, even ignoring column choices, there are 16 possible configurations for your Ultimates. Multiply that by 9 (3 basic columns per 3 trees) and you get a rough idea that the number of meaningful loadouts is over 100 for even a conservative estimate. The Masteries as they exist now present only the illusion of choice. Strip that away and you find one optimal path and many suboptimal paths.

This model feels much more natural to League of Legends and less like a holdover from World of Warcraft. It matches player expectations from in-game experience and it flows naturally. Each individual point is an important decision rather than a fiddly one.

The current system serves its purpose well enough and is unlikely to change due to inertia. When it comes down to it, the fiddliness and false choices aren't enough of a problem to merit a rework. However, should a rework occur, I vote for this because of the advantages to both choice and intuitive appeal.