I am thinking of making a hint system for my Inform 7 game. I almost made an Inform 7 thread asking for extension recommendations, but then I wondered: what is the current craft consensus regarding in-game hints? What do players like and dislike?
If a system is more than just functional—if it is interesting—what does that look like?
Are there games that stand out as having done hints especially well or creatively? Are there bad hint systems?
While I’m not looking for anyone’s take on whether hints ought to exist or not (no stealth bragging allowed here), I am interested in most other takes about hints loved, hated, wished for over the years.
This shouldn’t be a surprise, since I’ve been neck-deep in Infocom stuff: I’m most familiar with Invisiclues. I love their rhetoric and self-referential flavor, but I know there’s a wider world out there. Thanks in advance!
I prefer hints that are gradual but end up being clear. So the first hint about a topic only nudges you in the right direction, the second nudges a bit more… but the final hint had better just be a step by step description of exactly what I have to do. Nothing is worse than going through all the hints about a puzzle and still not being able to solve it!
Nothing necessarily wrong with creative / in-game hint systems, but I prefer the clearly extra-diegetic.
Yes, I think a funneling is best, and at the most narrow point, all should be very clear. By then, I’ve asked and I’ve asked for help, and it is ungenerous withhold the solution after so many entreaties.
Perhaps we can leave the door open for the creatively diagetic!
edit: @VictorGijsbers creatively extra-diagetic, I mean to say! Any character in my game that could provide that kind of help could not be convinced to do so.
Yeah, agreed with all of this – I think the best systems I’ve seen 1) unlock topics bit by bit, so the player doesn’t see the questions/topics in advance of when they’ve encountered the puzzle; 2) have at least three or four layers of hinting that gradually prod you in the right direction (with dependencies flagged early on – like “if you haven’t solved X puzzle yet, don’t read on”); and 3) end with the exact set of commands that are needed to solve the puzzle.
I’m sort of assuming here that that you’re using a menu type system rather than a context-dependent one where you get a single, hopefully-appropriate hint when you type HINT; I know there are partisans of the latter approach but personally I don’t like them. Even when they work well and are relevant to whatever puzzle I’m working on, to me it can make a game feel annoyingly linear.
The challenge for the context system would be guaranteeing that the player never gets spoiled or overclued. I think that would be very hard to test for. Or to get people to test for.
I really am just listening to ideas for now, but I would be very hesitant to do anything that could spoil something for players.
I make my BFF go through and type HINT before every move. And it’s still not perfect, of course. Definitely err on the side of underhinting, I think, because if the player is truly stuck, there’s always the walkthrough.
A vague concept knocking around the shadowy recesses of my mind is building a general constraint resolution engine one of whose purposes could be to provide a highly granular hint system: one would essentially build a graph of all the things necessary to accomplish some task, including acquisition of all the relevant knowledge to think to do those things, and the things necessary to acquire that knowledge. (I suspect in many cases, building this would be a useful exercise in discovering things that were inadequately clued.) And then when you asked the hint system about some given topic, it could dynamically point you toward one of the first steps missing, like “You never did examine the candlesticks on the mantel in the Sitting Room.” or “There’s a door in the North Hallway you didn’t try.”
That’s the hint system I’ve always wanted and haven’t seen (but might well exist out there somewhere – I just checked IFDB and I’m surprised I don’t see any lists regarding “Games with good hint or help systems”).
Edited to note: the above would also oblige maintaining a detailed history of things the player has seen, knows, has tried, has done, of course, to compare to the task-prerequisite graph.
As a caveat, I’d like to point out I hate hints for non-walkthrough material in the walkthrough. If something is an Easter Egg and non necessarily to complete the game, I don’t want it spoiled in a hint system or walkthrough. It’s a personal preference, I’m sure.
The Invisiclues-inspired systems tend to have a separate “for your amusement” section for easter eggs, jokes, etc.
I don’t plan to write a walkthrough of any sort, unless the hint system has accessibility issues (some menus don’t work well with all 'terp/reader combinations). I think that whatever I do will wind up being part of the game’s “text”, so I hope to avoid making separate documents.
This sounds amazing! I’d love to see it in action someday.
What I posited above is definitely a context-dependent system, though one striving both to always be appropriate and to offer the smallest relevant hint. Do you think that would still feel annoyingly linear? It’s my hope that the level of granularity would escape that.
I’m not making an Inform 7 game, but the point I want to make can definitely be done in Inform 7 if the game you’re making makes sense for that. I decided to multi-task my hint system in Budacanta. So yes, it does give help, but it also provides some bonus characterisation through the tone of those hints, and the sort of approaches advised in situations where more than one valid answer exists.
It’s definitely not the only way to go with hints, but I like the sort of hint system that doesn’t make the player feel bad for engaging with it. If you can find a way of making the player smile, laugh or otherwise care more about your game after reading the hints, then it
Song of the Mockingbird’s walkthrough would also be a good place to look for how to make a hint system that adds something to the game that makes it appealing. If you design your hint system so people who don’t actually need a hint still want to use it, then you will have cleared a high bar to making an interesting hint system.
Bad hint systems are ones that aren’t helpful. Better to have the most boring system imaginable that succeeds in providing hints than an interesting one that is useless or actively misleading. (Pointing towards something interesting, harmless but not directly giving the player progress is acceptable in moderation).
Make sure that it’s possible to get to an answer to all progress gates/puzzles/problems through the system (with vaguer hint layers beforehand if one has time and can think of a potentially useful “vaguer” hint. Feel free to have as many layers of these as makes sense for the part of the game for which the hints are being provided. If the final step of each part of the hint system, taken together, would enable a minimally successful walkthrough, you’re in the right place. This is especially important if your walkthrough is in a potentially inaccessible format like PDF.
Make sure that the last hint, at least, works for multiple playtesters.
Don’t put unreasonable restrictions on when the hint system can be consulted.
If a problem/puzzle/block to progress has been encountered, hints should be available for it from (at least) that point onwards. Err on the side of revealing too early rather than too late, if it proves impossible to time the hint correctly on 100% of occasions (obviously, just-in-time is the target for all hints). Being spoilered is not fun, but it’s better than being completely and (if there’s no walkthrough, wiki or other player to help) irretrievably stuck.
Hints should be optional, unless there is a compelling plot, setting or “fair play” reason to park a specific clue in someone’s face. If a puzzle is being portrayed as part of a competition, it’s fine to have an NPC announce the rules of solving the puzzle, but it is not fine for that NPC to give a step-by-step solution; that should be in the optional part of the hint system.
As long as the above are complied with, try for a writing style that fits the rest of the game. Definitely think about who is providing the hints (even if it’s the narrator, it’s possible to give that narrator a character with a distinct style, opinions and/or feelings). It doesn’t have to be the same source through the entire game, as long as the player always knows how to access the hints. A hint of the correct level of clarity is good. A hint of the correct level of clarity that suits the game, character and situation is brilliant.
I like @Zed 's idea of using a puzzle dependency checker for the game to assess which hint set(s) you need to see.
What’s worse, it’s painful to build the hint system early in the game’s development, because every time you change a design element or a goal, it must be reflected in the hint system as well. But if you wait until the end of the design cycle, the hint system will be one of the least-tested and least-exercised parts of the game.
I came here to say something similar. As I was building out my hints, I realized the task was forcing me to re-examine the game design with fresh eyes. Did I provide sufficient clues for all goals? What linear paths exist in the game? Are there possible unwinnable states?
Apart from Zed’s fantasy system, and maybe even it as well, probably every hint system will, at some point, annoy someone, to a degree.
Meself, I can’t stand invisiclues style. But I do mean the style chained to how Infocom dispensed them. Its problem for me is that info is heaped high on fewer points. If what I want is far into the heap, I have to read lugubrious prose (well, lugubrious if you imitate how Infocom delivered invisiclues) about stuff I already know to get to it, and I can’t even be sure it will be there. So in that sense I prefer more numerous clues offered on more points, which points towards Zed’s system.
If you take the invisiclues heap and at least use software to filter to the level the player’s at when they ask for a hint – which numerous games now do, and every Andrew Schultz game – that’s already an enormous improvement.
On the other hand!.. the old analog formatting of invisiclues is hinting in itself. You can read two questions and realise you’re sort of inbetween them, game wise. Or you can find that you’re far in the game but have missed something that comes much earlier. Of course digital hints can reveal this as well, but it won’t be in a context, it’ll be in a continuum of what to do next. But your game may not suit analog hints or certain kinds of fourth wall breaking.
As a last suggestion which will suit few real life applications – but is super fun if you can avail yourself of it – you can do a hint sheet the Scott Adams way, as I did for Leadlight. Sheet attached.
I like in-world hint systems, say a “seer” located somewhere in the game world whom you have to navigate to to pick up advice, more than a “hint command” which breaks memesis and become too tempting to use.
Jim Aiken’s game “prom dress” had a cell phone that would deliver hints to the player, purportedly from the daughters fortune teller.
I do prefer when the hints are integrated into the story rather than a separate menu to consult. In a couple of my games, I had a Think command that would deliver hints that change depending on location.
Oh, the fantasy system is the one where instead of creating an obstacle-resolution graph that redundantly reflected huge portions of the game, the game itself is expressed as such a graph and the hint system becomes a fairly trivial gloss over the existing data structure.
That’s somewhat similar to something I’ve been working on implementing myself.
It started out as a test harness for unit testing: given a relational graph of a puzzle, generate and execute the command line inputs that a player would type when solving that puzzle.
From there it grew into a way the player could enlist the aid of NPCs to solve puzzles they were having trouble with. The idea being that NPCs will help the player with puzzles (or just specific bits of puzzles) in exchange for the player helping the NPC with something, which allows the game to implement a sort of puzzle type exchange: if the puzzle the player is having trouble is Towers of Hanoi (disclaimer: I wouldn’t implement a Towers of Hanoi problem as a puzzle) then maybe the NPC who’s an ace at that kind of thing wants the player to beat them at chess (disclaimer: I wouldn’t implement a chess problem as a puzzle). And so on.
I’m also implementing an in-game hint system, but having NPCs help this way both effectively allows the player to sorta opt out of certain puzzles in a way that “makes sense” in-game, and it basically functions as a walkthrough (of the puzzle) done completely in-game (since the NPCs are executing commands the same way the player would be, the player can just watch and observe all the steps to solve the puzzle).
Also, I found a list of