Tutorial, or not Tutorial

Inspired by the discussion I got into this morning on Twitter, where there simply aren’t enough characters, my strong feelings on tutorial design.

Background

I didn’t design any tutorial for any released game. I worked on some tutorials for a game that didn’t ship (and I don’t think the decision to cancel was based on the tutorial). I had supporting roles in at least one of the tutorials for RuneScape, managed people with more direct roles, and paid plenty of attention to the ones I didn’t work on.
So if you want to disagree on the strength of your experience, go ahead. I encourage you, as always, to look at these issues with whatever evidence you can gather (be it player testing, metrics, psychology or other science, etc.) and make sure you’re approaching learning design in a rational way.

Tenets

The following core statements are intended to be fairly uncontentious, so that you can see where I’m coming from.

1. Players who do not have the skills to play your game effectively will not enjoy it as much as they should. It’s your responsibility as a designer to give them those skills.
2. Players hate to be taught, in an obvious or heavy-handed way.
3. ‘Tutorial’ is a dirty word, conjuring images of the obvious teaching, above. Many studio staff will avoid using the word, but don’t let that distract you from the need to provide a learning experience, per (1).
4. If offered the choice to skip a tutorial, many players will do so even though they don’t have the skills (because of (3) and (2)). Then see (1).

Results

You’re not going to like this. I’m going to have to spend the rest of the article defending this position, even though I’m trying to make sure it flows logically from the premises above.

• You must have a learning experience.
  • Letting people inside the studio claim that the game ‘has no tutorial’ is dangerous: sooner or later it leads someone to dismiss the need for a learning experience, not giving it the budget and attention it deserves. Because of this I recommend getting stubborn, calling it a tutorial (internally) and explaining to everyone why it’s necessary (and why it won’t be terrible). You (and Marketing) can call it whatever you like, player-facing.
• Your learning experience must be as smooth as possible – fun, like real game-play (and a good advert for the rest of the game) – and as far away from teaching as you can manage.
  • This statement isn’t fair to teachers, since modern teaching strives to be interesting, interactive and varied, but you know what I mean. (Hence the scornful italics.)
  • Lots of things can help with this; I’ll give some ideas below.
• As part of making the learning experience smooth, it should react to the player demonstrating skills to avoid unnecessary delays and teaching. However, you should never ask the player whether they want to skip the tutorial.
  • Two near-exceptions: Players who have already played this game, and players who have already played similar games. I’ll deal with these below.
  • Multiplayer short-session games, especially competitive ones (e.g. MOBAs, many FPSs), pose additional challenges, and an unskippable single-player learning experience might not be acceptable to the player. I’m not a specialist in such games, but I’ll have a go at that below, too.

Good Player Stories

If you’re into such things, here are some target player experiences. Let’s assume a fairly typical single-player FPS with single-player campaign.

Experienced

Andrea is an experienced FPS player, who happens not to have played yours yet.

• She starts the single-player campaign. It begins in a relatively safe space.
  • It may or may not look safe. I’d suggest that you may want the environment to be an exciting (i.e. dangerous) one (e.g. we’re in a battleground) but for the immediate surroundings to be obviously safe (we’re in a sturdy bunker). Most FPS players will probably be OK if the starting area doesn’t appear safe, although it should still be safe in practice.
• She immediately starts to move and look around, comfortable with WASD and mouse-looking. The game reacts by letting her proceed at full speed, skipping the instructions and tests for those interactions.
• She is given an objective, and attention is drawn to the on-screen/in-world objective marker. This isn’t tested, per se, but she continues to be exposed to it throughout the game, and will be re-prompted if ever she seems lost or unaware of the current objective.
• She is given a weapon and something to shoot at. She immediately demonstrates that she understands left-click firing, so the game does not explain it.
• She tries aiming down the sights, but this game has right-click alternative-fire. The game doesn’t tell her more until she needs it, but perhaps it logs that she’s experienced what the button does.
• She approaches an obstruction that needs to be jumped or mantled. If she hasn’t already jumped, she now gets an onscreen prompt with the keybind.
• She enters an area under enemy fire. The sergeant tells her to get behind cover. If she hasn’t already used the crouch key, she gets an onscreen prompt about it now. (More so if it has detailed cover mechanics, since she won’t have demonstrated them yet.)
• She finds a crate of grenades (an unlimited supply). The sergeant points to a machine-gun nest: “Corporal, I need a grenade on that MG position, now!” Used to another game, she can’t find the grenade key, so after a moment an on-screen prompt tells her which key grenades are bound to. She can’t continue until she destroys the nest as ordered.

And so on for other player capabilities. Any time she encounters a new skill she’s given an opportunity to demonstrate that she can do it (if that’s possible in context), given guidance if she needs it, and tested to make sure she’s understood.
The initial learning experience is as short as possible (depending on how many player skills are absolutely essential to the core game experience, starting with those that logically can’t be taken from the player character), but learning continues throughout the game. The player continues to have basic tests woven into the experience (e.g. things that must be crouched, things that must be jumped/climbed, things that explicitly should be attacked with grenades), and any time she appears to be stuck on a test, she’s given the onscreen prompt again. Additional player capabilities are unlocked throughout the game, and treated in the same way: allow an experiment, inform as necessary, test.
At no point in this flow is she ever waiting for information she doesn’t need, and forcing her to wait for information she does need is kept to an absolute minimum. For example, you should not lock the character’s movement until you’ve tested looking around or in order to tell the player how to move (you might do so for narrative reasons, but I’d argue that is risky), because in practice many players will immediately want to do both.

Inexperienced

Bartholemew has never played an FPS. He’s probably fallen through a portal from the 18th century, and yours is the first game he’s found.

• He starts a single-player campaign (it could have gone without saying at this stage that your menus need to be accessible to people who don’t know the genre conventions, but that’s a rant for another day).
• In the safe space, the character stands like a lemon, his player worried about giving input that might cause something to go wrong. He gets an onscreen prompt encouraging him to look around with the mouse.
  • Unless your game relies on looking up a lot, don’t stress too much about teaching the full capabilities of mouse look. At this stage the player just needs to be able to steer the character.
  • If you’re trying to keep your instructions in-character, you may need to fudge things a little here. Give the sergeant a reason to cut the character some slack and give a little encouragement (maybe the character is recovering from injury), rather than having him go full parade-ground arsehole on the character.
• Once the game has seen that the player can look around, he gets an onscreen prompt with the movement controls, and is given somewhere to go.

And so on. This is the tutorial most players don’t want, and yet some players need it. We’re still not locking the controls, we’re not giving walls of text, and we’re still following the basic pattern. It’s just in this case we’re ready to give a lot more guidance (including perhaps alternative versions of onscreen prompts, for variety and expansion).

The Recipe

I propose (and I hadn’t thought this hard about it before, so I’m making this up as I go) that the basic recipe for getting the player to learn a skill might go like this:

Unlock the capability and let the player experiment

This may not always be possible, but actually I’m struggling to find a sold counter-example. Clearly you can let the player look around, etc. You could give them cover to play with your cover mechanics, even before they need it in combat. If you need them to use a huge superweapon (or other narrative reason they can’t practice safely) give it the same player interactions as something they can practice.

Prompt with a test to confirm they’ve understood

Note that we’re presenting the need for the skill before we’ve formally introduced the skill itself. This is important for high-skill players, because it reduces the chance that they’ll get frustrated as you present them with something they already understand. It helps low-skill players, because they can take in the skill introduction more easily once they’ve seen a use case for it.
So for example, I see the crouch-obstruction before I’m told I can crouch. I’ve seen enemies that I should throw grenades at before I’m told how to throw grenades.
In some cases you can do this in multiple stages: the use case is seen, then I’m shown how to deal with it (next section), then I’m given a formal test on it. This is may be better in the case of the grenade throwing: I understand there are hard-to-shoot enemies, but the sergeant doesn’t give me the order to destroy that MG until I’ve had some grenade advice.
You’ll need to fine-tune this timing. Giving advice before the order is safest, and some players may deal with the MG before the order arrives. Giving the order and then a brief window in which to destroy the MG may be good for players who can get the controls right but not intuit the objective, but it’ll be bad for players who are waiting through that brief window with an order that they don’t feel equipped for.

If the test isn’t completed promptly, provide guidance and information

The guidance you give is a balance between players getting annoyed because they don’t like to be told what they know, and players getting frustrated because they’re not being told what they need to know.
If possible, start the guidance small and ramp up. Give progressively more explicit advice, or employ more screen/world highlighting and other tools, the longer a player goes without succeeding at the test.

Once the test is complete, allow progress

And test again later.

In Other Games

The single-player FPS isn’t a complicated game to play and the genre has some really strong conventions that you can lean on. I picked it because the stories fit nicely into a blog post. But even so I bet you can think of some that wouldn’t give Andrea (or you) such a smooth starting experience.
Other genres can be trickier, and it’s often about how much information you consider to be vital core experience (or generally how ‘large’ that ‘core’ is).

Strategy Games, Broad MMORPGs

I use ‘broad’ MMORPGs to describe ones with lots of parallel content and player choice, be it professions, skills or whatever (Hi RuneScape!), rather than ‘tall’ MMOs in which everyone is on roughly the same critical path. The latter are much more common (WoW, LotRO, SW:TOR, etc.) and I suspect there’s a bigger market for them (and if you catch me telling you otherwise it’s probably because I’m trying to persuade you to fund a broad MMO and hire me to lead it). They’re also much easier to teach, because you’re free to just level-lock the content, drip-feeding it until the core game is small enough for an ‘easy’ tutorial as described above.
In a broad MMORPG you have to balance keeping the learning experience small with making it an appropriate advert for the breadth of the game. This is really tough, and perhaps a topic for another day (and needs to be tailored to your game anyway). I suggest you err on the side of a small starting experience that shows that breadth is there, but doesn’t require the player to learn it yet (or wherever possible, ever, because the choice not to engage with content is also important to your breadth). You could (and should) have parallel learning experiences for different streams of the game, but you need to be careful where player choice paths overlap and cross.
Strategy games are in a similar boat. In many games there’s a ‘standard’ mode with all techs etc. unlocked, and a huge pressure from players to get them ready for that ASAP. I’m not a specialist in the genre (and I don’t want to spend too much space speculating now), but you might try making each playthrough scale as though level-locked so it can have a tutorial as above (tech trees are good for that), or if that isn’t appropriate try taking the emphasis off the ‘open’ scenario and lead the player through a more structured experience first.
In both cases, it will still be hard to make a tutorial like the one described in the stories above, because there’s simply so much information to go across. But trying your best might really open up your potential market.

Competitive Multiplayer

So, the bad news is that you can’t really separate out the learning experience for competitive multiplayer games (MOBAs, many FPSs, etc.) without getting something that fundamentally isn’t representative of the game, which annoys players and hurts their learning. The good news is that your players will be mastery-driven, and will cut you some slack provided you’re helping them crush their foes in the long run.
Many of the best learning experiences I’ve played in this category (and I’m by no means an avid player of these games) have simply used matchmaking or game categories to put you into early games that are safe to experiment, maybe a little easier or slower, and otherwise allow for learning in a space that doesn’t hurt your standings. Have a boot camp mode, make it compulsory for low-rank players (but make it easy for competent players to get out of it quickly), don’t allow high-rank players (but don’t stress too much because they could alt: you’ll need some other mechanics to make sure it’s not fun to twink or interfere with). Have the bootcamp give frequent onscreen tips (in tiny doses) when it notices that the player isn’t using a skill.
Once you’ve done that, the best you can do is make it easy for players to perfect their craft on their terms. Give them information they can look up, give them a practice mode with dummies or bots.
Or maybe ask someone who’s actually worked on one of these 😉

The Two Exceptions

Above I mentioned two ‘near-exceptions’ to the rule of never offering a choice to skip: a player who has played this game before, and one who has played similar games (or the previous ones in this franchise).
By now you can probably guess what I’m going to say: if you’ve done everything you can and made the smoothest possible learning experience that reacts properly to the player’s existing skills, you can be confident having them all play it anyway.
If players are replaying the game (new game plus, or alt characters), you may want to let them skip. It’s not that I’m not confident in your tutorial, and you shouldn’t be either: I say this because you may want to let them around the early non-branching narrative. The big But is this: make it skippable for new game plus, or when the account shows that this isn’t the first character to do it. If you ask the player, probably even if you’re asking the apparently innocuous ‘is this the first time you’ve played’, you’re going to get some people skip who shouldn’t.
Even after you restrict the skip to confirmed prior players, you should probably still make it optional. You could make it opt-in rather than opt-out, though.
If players have played a similar game, they’ll be able to coast through your tutorial with minimal resistance, because you designed it so well. But ask yourself how similar other games really are. Even FPSs, which have some core controls in common in almost all cases, have a wide range of secondary controls, including what the right mouse button does, which key throws grenades and which does melee, exactly how weapon choice works, and so on.
And you probably don’t want to commit to following someone else’s control scheme, or even one of your old ones. The sequel might need to change something, or you might need something a little different to your competition.

Learning Styles

I haven’t said much about learning styles. There’s a lot of academic writing on the subject, and lots of competing models, some or more of which may have been discredited already. I’m far from a specialist (ask me again in two years).
The important thing to realise is that people do learn in different ways, and that you need to try to cover all bases. I’ve already mentioned experimentation, testing and repetition, so most education theorists probably haven’t given up on me yet.
The other thing to do is to make sure that the guidance you give is multi-faceted. We all assumed text; voice-overs are nice too (but give the text too). Back them up with visual highlights in the UI or the world. Add diagrams if it makes sense. Videos may be useful but can be problematic, see below.
Games are inherently interactive, and are already an excellent medium for teaching. Use that, and supplement it with other stuff as best you can.

Please Don’t:

I’ll sign off soon, because I think I’ve covered all I wanted (and the post is long, as always), but a few things to avoid. As a player, I beg you, please don’t resort to:

Walls of Text

I just don’t want to read that much in a game, and I’m not the only one. I read novels, but I approach games in a completely different frame of mind, and I was surprised by how averse I am to reading their text.
Tutorial text should be as short as possible, be relevant right now, and backed up wherever possible by relevant visuals (such as highlights for UI elements, even diagrams, etc.).

Big Videos

This may be controversial, but I really don’t like games trying to make me watch tutorial videos, and again I’m probably not the only one. Videos show very well, but they don’t offer the chance to experiment or practice, they don’t test, and they generally don’t have repetition (and if they do, you’re obviously teaching me and I’m going to get bored). While they may be a great resource for people who need to see something demonstrated, they’re not good as your primary approach to player learning.
The right context for videos would be as demonstrations, to back up guidance (but probably not to be your front-line guidance, because videos take away the player’s control of pace). Save them for things that are conceptually difficult (e.g. where a player should aim, rather than the mechanics of aiming). Make them really short and focused, to minimise the time away from player control.
Remember the pattern above. The problem with both text and video (or voice-over), is that it only provides one part of the pattern, so make sure you break it up fine enough to fit in that slot.