Out by Midnight

Over the past ten years or so of IFComps, Art DiBianca has firmly established a niche of puzzle-box-style limited parser games. No frills, minimal story, just a small set of well-honed puzzle mechanics and maybe a map to explore. Out by Midnight sees DiBianca applying his signature style to an optimization-based treasure hunt in the vein of Captain Verdeterre's Plunder or Sugarlawn.

After playing a few games that felt like their reach was exceeding their grasp in various ways, it was nice to sit back, relax, and go on a ride with a veteran author doing something that is firmly in his wheelhouse. I think this could be an instant new favorite for some people. Unfortunately, I don't quite have the kind of brain that loves to latch on to this kind of thing. I enjoy the exploration phase, but once the new discoveries slow to a trickle (or dry up completely), I start to lose interest in the optimization part. (I've enjoyed replay-based optimization mechanics in games with a little more story/flavor, like Forsaken Denizen; when there's a narrative I'm invested in, little tidbits of new stuff can keep me going for quite a while. But that's not what Out by Midnight is about, and I respect that but it does make the game not the best fit for me personally.)

So although, as is common for DiBianca's games, meeting the basic win condition turns out to be just the tip of the puzzle iceberg, I called it quits after getting there. Out by Midnight is a tightly constructed, nicely polished experience, a game that knows what it wants to do and does it without a hitch. It's not my new obsession, but it might be yours.

I always look forward to Arthur DiBianca's games, which tend to be tightly organized around a single concept and to be minimalistic in verb usage and text.

This one was a timed optimization puzzle (turn-based time, not IRL timed), where you wander around looking for 10 objects in a map. Most objects have duplicates. Each room has 0-2 interactable objects or doors that take a fixed amount of time to open. There are shortcut items as well including scissors, wd-40, paint remover, etc.

So it becomes a challenge to find all 10 minutes before the end of the day.

I get 5 easily, and started mapping it out, but there doesn't seem to be a lot of reward for it. There's not much story or surprise. I did like the ambient messages. If this were a first-time author's game, I'd say it was amazingly well-implemented and interesting. I guess that Arthur DiBianca just has so many good games that, compared to his others, I'd probably place it 10th to 12th. (just looking over his past list to verify this, there are so many fun ones on there like Skies Above or Variety Box). Out of parser games I've played this competition though, I'd say it's definitely above average.

At this late date, is there juice left to squeeze out of the Verdeterre-like? A baker’s-dozen years one from the inaugural entry in the subgenre, we’ve seen at least five or six other runs at the collect-treasures-under-time-pressure premise, from swapping out the pirate milieu for sci-fi (Dark Star) or Egyptology (Curse of the Scarab), making the map much bigger and more complex (Sugarlawn) or smaller and more manageable (Museum Heist). I don’t think we’ve seen a limited-parser take, though, so it makes sense that Arthur DiBianca would throw his hat in the ring.

The signature elements of the author’s recent games are all on offer here: there’s a slick map to make navigation a breeze; actions are limited to movement, observation, a general “use” verb, and various meta commands; and there’s a post-game achievement system that kicks once you’ve managed a basic win and encourages you to take on trickier challenges. It makes for a slick, easy-to-get-into experience, and the setup – you’re on a scavenger hunt in a spooky manor – likewise is immediately intuitive. Since this is an optimization game, your eyes will glaze over after the fourth or fifth time you read the same bit of prose, but the writing is certainly fun enough when you first encounter it (I liked the wry note, upon checking out the dilapidated surroundings, that “it would be nice if ‘dead plant’ were an item on your scavenger list, because it would be so easy to get”), and unobtrusive enough to get out of the way thereafter.

And there’s a whole lot of thereafter, as even getting to the basic winning condition requires a good amount of iteration. There are ten categories on your to-collect list, some of which can be satisfied by more than one item you can find in the mansion, but you need to collect all ten and leave the mansion before midnight. The game starts at 7, but time goes fast since working through the house’s various challenges – picking locks, untying knots, cracking a safe the old-fashioned way – can make 15 or 20 minute elapse in a single action. It quickly becomes clear that many of these barriers have workarounds: there are keys and tools that can make a task that otherwise would eat up 17 minutes take only 2 instead. But with something like a dozen keys and 7 or 8 tools to keep track of on top of stuff you need to collect, it takes a bunch of resets to run down all the permutations and figure out whether the lubricant, the paint remover, the grabber, or the magnet on a string would help you speedrun any particular “puzzle.” The game does offer some support – that map is super helpful, and there’s a RECALL command that gives you an instant download on everything you’ve learned about a room and its contents – but you definitely will need to take a bunch of notes, and might even want to consider a spreadsheet.

Now, I’m the kind of sicko who usually gets excited when I get to fire up Excel alongside a piece of IF, and have generally gotten most of not all of the challenge-mode achievements in the author’s previous games (even the sokoban one!) but I confess that while the bones of the design here are good, I wasn’t sure I’d have the motivation to power through to even the basic ending. Partially this is because the difficulty curve is steep: blundering your way through the house sufficiently fast to check off six or seven categories isn’t too hard, but close only counts in horseshoes and hand-grenades, not haunted houses. I don’t think getting to eight or nine, much less all ten, is possible without wrapping your brain around a fairly complex web of dependencies and choices: you need tool X to get treasure Y quickly, but that means finding key Z which is all the way on the other floor; or maybe you could go for treasure A for that category instead, which requires tool B which you know you’ll already need to get treasure C, so it might wind up being faster even though you’ll need two different keys to get to A. Or actually might it be the case that all this rigmarole is more trouble than it’s worth, and grabbing treasure Y the slow, no-tool-or-key way, will wind up being faster at the end of the day than jumping through all these hoops?

But the overall challenge level wasn’t the only factor for me – there’s also a real dearth of dopamine-hits along the way. In a typical Verdeterre-like, as you go through runs, you’ll find new treasures, solve traditional IF puzzles to get them, and unlock shortcuts. There’s a regular drip of discovery with your increasing knowledge opening up new options and opportunities: progress is usually binary (did you solve the puzzle, yes or no), with the impact on your score more granular (maybe the new treasure is worth $10, but maybe it’s worth $100!) – and at the end of that process, you sit down to chart the optimal path.

Here, by contrast, you can brute-force every puzzle the first time you encounter it; you’ll see everything there is to see in a room on your initial visit. Sure, much of the middle-game does involve tracking puzzle-like interactions like seeing which tool works on which barrier, but you’ll always know what’s on the other side of the barrier in advance: there’s no narrative reward, just the knowledge of how much time you can now save. And because you need to get one of each category of item, with all of them counting exactly the same, there’s no thrill from finding something especially exciting: it’s the reverse of the usual Verdeterre approach, with progress entirely granular (+24 to the timer to get an item rather than +41) and results completely binary (do you have one of these treasures Y/N).

For all these complaints, after a good night’s sleep I did put on my big-boy pants and despite the comparatively-arid mid-game, I compiled a full set of notes, charted out a plan of attack, and implemented it sufficiently well that I also wound up with the bonus achievement for winning extra-fast. Optimization is fun, and it’s satisfying to gain mastery of the complex web Out by Midnight offers up! But I don’t think I’m feeling up to going for any other achievements; for all that this is a heavily-mechanical subgenre, part of what made the original Captain Verdeterre’s Treasure so successful was its demented energy, with the onrushing waters and the abusive rat providing some narrative adrenaline to the bean-counting and the aforementioned regular unveiling of new stuff promising that doing the work would lead to even more fun. Out by Midnight is cleanly designed (and so far as I can tell, completely bug-free), but a bit too staid to reach my personal Verdeterre-like pantheon.