Given that I haven’t updated my blog in more than a year (oops) likely the first place you find out about this won’t be here if you managed to find your way here from one of my other socials. BUT, I’m giving a talk at GDC next year!

I wanted to focus some of the discussion on topics people found interesting, so I asked folks on twitter for topics they were particularly interested in covering.

Players run WoW dungeons and raids many times. How do you figure out whether a mechanic has legs and will remain fun over time?

How do you allow for players with different skill levels to play in the same group and have fun with each other?
— Lowbo | 就是萝卜 (@itslowbo) December 7, 2023

These questions stood out as maybe not directly related to what I’m covering, but definitely worth discussing. I thought I’d tackle Replayability first.

Replayability

To be completely honest, from the point of view of an encounter designer, or more broadly a combat designer focusing on enemy behaviors, I think it’s very difficult to identify a mechanic as having legs. Replayability is born of the holistic excitement of a play experience, and not any individual mechanic on its own, in my experience. It’s a resonance between all of the various aspects of a game humming together in dynamism.

But, when limited to that encounter designer’s point of view, I do think there are things you can evaluate to try to encourage replayability:

First - identify the aspects that contribute to your game’s game state that are “chaotic”. For a multiplayer game, that’s fairly obvious (other people) - but as a random example, for a single-player game it might be the physical state of the character. The thing that makes any given jump tricky in an ice-level in a platformer is that the player can’t have quick reactive control over their momentum, so that’s might be a chaotic aspect you could play with.
Second, identify the type of replayability you are worried about.

Often, roguelikes and high-difficulty encounters in games are focused on replayability of the form “repeated unsuccessful attempts to defeat this encounter should continue to be satisfying”. The idea of “failure is fun”.

On the other hand, low-difficulty repeatable content is often focused on other types of replayability - maybe social replayability that comes from the experience of doing something with different groups of people, or simple variance, where individual attempts of an encounter have radically different ability sets or mechanics.

It’s extremely important to identify which you are going for because they are often diametrically opposed! The random mechanical variance that might work in a piece of casual content is often, in my experience, despised by players challenging extreme difficulty content. They want the ability to break an encounter down mechanic by mechanic and have an enjoyable learning experience. Randomly swapping what they are trying to learn out from under them does not, in my experience, accomplish replayability but instead increases frustration.
Carefully balance how and when the player will shift from reactive play to learned play. The core difficulty here is not making a mechanic interesting on future playthroughs but instead making solving the mechanic rewarding on future playthroughs. Responding dynamically to a chaotic situation can be extremely taxing mentally for players, so allowing a dynamic situation to quickly “fold down” into a learned process can help balance approachability with replayability.
Avoid purely binary outcomes. For instance, contrast a situation where a boss goes invincible and forces the player to dodge a complex series of mechanics vs. one where they remain attackable during the sequence. When the boss can’t be attacked, there’s no headroom for a player once they’ve survived the mechanic - either you die or you don’t. When the boss can be attacked and the player is forced to multitask between doing damage to an enemy and avoiding the enemy’s attacks, you essentially have a sliding scale of player performance, all of which would be considered “success” from a purely mechanical point of view, from “I’m 100% focused on avoiding this attack and not trying to do any damage at all” to “I’m able to dodge this mechanic entirely while continuing to do as much damage as normal”. This isn’t just relevant for MMOs - even action platformers like Mega Man, or 3D action games like Dark Souls have boss mechanics where great players can track the boss and deal damage during them and less skilled players have to focus entirely on dodging, increasing the overall length of the fight.

To assist with this, it can often be really helpful to have useful player-visible metrics - either on the game level or within the encounter itself. The obvious ones are things like damage meters or fight timers, but one common more organic example seen in MMOs is being able to push a boss encounter to the next phase before a certain ability goes off. It’s easy to see that as a design flaw, but in reality it can serve as a useful progression marker for a group who has already beat the content and is trying to measure their progress this week against last. Players optimizing their gameplay performance has bred the entire speedrunning scene and leveraging that instinct in your designs can give players a taste of that progression in normal play.
If you have different classes like in an ARPG or MMO, or different hero characters in a game like LoL or Overwatch, try to make your mechanics clash with the differences in hero stats and abilities in a way that will substantially remix the way an encounter’s mechanics play given a group composition or player choices in character creation. For example, when trying to design engaging “Tank Mechanics” (basically, mechanics that are intended for tanks to do and which force a group to bring a certain number of tanks to a raid encounter), I would often try to find mechanics that would feel very different to the group based on which tank classes they had selected. In WoW this had tons of different angles it could be approached from - different tanks had different damage mitigation strategies, different movement styles, different sets of defensive cooldowns, etc. But, in addition to the differences for the Tank the best Tank mechanics would change how everyone else would react to the tank mechanic based on the class of the Tank as well. Of course, this tends to be more palatable to players on harder difficulties - you often don’t want to target challenge as your avenue of replayability in more approachable content.

Note this is also more useful in a game where you are encouraged to run content with lots of different players. For instance, this kind of variety tended to work better in Mythic+ Dungeons in WoW, which only required 5 players and was basically focused around pick-up groups, vs. Mythic Raids, which required 25 players - the substantial organizational headache tended to force players to form more or less static groups, where they weren’t as likely to see the variance that came from different tanking compositions.
Finally - and this is most important way but also I can only provide the how here and not the what - but gather the data as a player and from your players.

Keep a list of things that worked in the past and that you can remix or introduce. When you’re designing, implementing, and playtesting a mechanic, are you having fun? If your game hasn’t been released yet, look at other examples from your genre, or that your audience would be familiar with. The more you can empathize with your player’s point of view, the easier it will be for you to identify probable reactions to things you make.

Finally, if you are making a game where replayability is a core goal, I would strongly recommend spending a lot of time doing research in “session” games, aka games where you are intended to do many playthroughs specifically for mechanical reasons (i.e. not VNs or choice-based RPGs - totally cool but a different source of replayability imo). Roguelites/Roguelikes, 4x (Civilization etc.) and Tactics games like XCOM are all great examples. PvP games can also be a big help, though it can be a little harder to directly apply the lessons. But standouts from these genres are great examples where you can experience how to achieve replayability quickly and on your own. Deriving lessons from those games, in my experience, directly contributed to my ability to identify and create replayable mechanics in my work.