Reviews by Mike Russo

Having seen me heap 1,500 words of scorn on The Fortuna for the various crimes arising from its use of AI tools, the reader with a vicious streak has perhaps been waiting for me to arrive at Cheree, in anticipation of another evisceration. This is another game that’s built around AI tools, though in a quite different way: instead of being structured like a traditional text adventure, instead it’s built as an extended chatbot session – the conceit is that you’re conversing with the ghost of a young woman who was murdered in Victorian times, and helping her to recover her lost memories. It’s by far the longest game I’ve seen in this format – my playthrough took maybe two or three hours? – and while there are a few traditional puzzles mixed in, the gameplay goes pretty much as you’d think: you’re just talking to a chatbot. Yet despite that, and my by now well-established anti-AI bona fides, I think it’s actually rather good?

(The reader with a vicious streak will be disappointed, but at least the game tangentially touches on the Whitechapel murders, so perhaps that will provide a measure of appeasement).

Explaining why requires talking through the game’s structure and plot in some detail, so let’s dig into that: as mentioned, the eponymous Cheree is dead, doesn’t fully remember the circumstances that led to it, and has come to the player character – turns out you’re a powerful medium, don’t you know – for help. This involves a methodical investigation of locations significant to her, which fortunately she can whisk you to via the powers of astral projection. Sometimes familiarity will spark a memory and she’ll share a few sentences of reminiscence. Then by asking probing questions you can help her remember more and more, which sometimes triggers a cryptogram puzzle. These puzzles are blatantly there for pacing reasons – they aren’t especially difficult and lack any in-game explanation – but they are successful at breaking up all that talking, and don’t outlast their welcome. Anyway, once solved, each will give you a clue; collect all nine clues and you reach the endgame.

The story that unfolds is pure pulp, but it’s well-made pulp; as mentioned, it feints in the direction of the Ripper killings, but doesn’t go too far in that direction and generally treats the subject tastefully. Instead, the story is more of a domestic gothic, featuring a gloomy, religion-obsessed patriarch, a pushy mother, a jealous sister, and her dangerous ex-military fiancé. The present-day section of the story also gets more complex with the introduction of one of Cheree’s still-living relatives who astral-projects her way into your little tete-a-tete. There are lots of twists and turns, and while it perhaps goes a bit over the top in the ending, it succeeded in keeping my interest throughout, and even surprised me once or twice.

Of course, rendering a plot like this as a chatbot rather than a more traditional form of IF is a very risky maneuver, since chatbots can be fragile beasts. The author has been very canny about this, however, disarming some of the most obvious traps through clever narrative and systems design. Like, one of the ways things can go wrong is if the player references something that the character could plausibly know about – often, the bot will either not understand, or err on the other side and start spouting Wikipedia summaries. This came up for me when Cheree made some comment about how it’s better to be better off, rather than better off than others; I told her that Thorstein Veblen would disagree. She responded by saying oh, she’d heard of him and his most famous book. From a systems perspective, this hits the sweet spot – it’s identifying the reference without going into too much detail – and it makes sense diegetically too, because as a ghost who’s been hanging around for more than a century but who can’t physically open up books to read them, she should be familiar with a bunch of stuff but not able to discuss them in depth. Sure, this isn’t infallible – at one point I asked her to go to a lighthouse, which she interpreted as reference to the Woolf novel – but it’s really quite well done.

Another common flaw is the way dialogue can sometimes become one-sided, leading the player to mechanically type in one thing after another in hopes of progressing. Here, Cheree has some agency, doing things like suggesting the player move on if the conversation isn’t progressing, and occasionally quizzing the player on some piece of trivia if they’re quiet for a while. There aren’t any consequences to guessing right or wrong, and yes, it’s clear that this is a game mechanic to sustain engagement, but it’s still fairly successful at accomplishing those goals while once again making sense in the world: if you were a bored ghost with a century to kill, you also would accumulate a ton of random knowledge and be bad at making small talk!

Voice can also break down in AI-driven conversation systems, as chopping a bunch of training text up and putting it into a blender can either lead to pure oatmeal or a wild swing between different tones. There’s a little of that here – Cheree’s set-piece narration of her memories is written in a faux 19th Century novel style, while she gets more informal in conversational back and forth with the player. Once more, though, her unique biography provides a plausible excuse: she’d presumably revert to the language of her living years when recalling them, while outside of that context she’d be more likely to use the colloquial language she’s picked over the decades since.

The final guardrail is the relationship system – progress in some places is gated by Cheree’s level of affection for you, which I assume means that players intentionally trying to mess with the game will hit a wall and have to behave in a way more appropriate to the narrative conceit in order to move the story forward. This seems like another smart, plausible safeguard, though I confess I can’t comment on the execution since I didn’t try to test the edges of the simulation.

Cheree is not without blemish, though. One clear misstep is that there’s far too much empty space. I can see how limiting the potential locations just to those where clues may be found could have made the game feel too mechanical, but the author overcorrected by providing lots of places that are just there to establish Cheree’s love of sightseeing – the graphics of some of these scenic overlooks are nice, sure, but in a game that’s already fairly long, I got bored of the filler. And actually, while we’re on the subject of the graphics, I was not a fan of the 3D models used for Cheree and her relative – they’re relatively low-poly, but what’s worse, Cheree is depicted in full loligoth style and often walks straight into the screen, meaning the camera clips into her crotch, while the relative is dressed in her underwear and spends a lot of time gyrating around in the corner regardless of what’s happening in the plot. Look, I’m not averse to a bit of sex appeal, but if that’s the remit, for me personally seeing a nubile young woman cavorting in her underthings while her friend talks about finding where her murderer dumped her body does not achieve the goal.

This rather cringey aspect of the visuals also combines poorly with a much cringier aspect of the narrative, which is the romance plot. Yes, as you’re building a positive relationship with Cheree, she starts to develop a crush on you, and the other girl isn’t averse to some light flirting either. On a certain level, I can understand her being a bit of a horndog – girl’s been dead a century! – but like many video-game romances, this one suffers from feeling like it goes way too fast. It also creeped me out because the chatbot format made me default to playing the game as myself: Cheree sometimes asks the player personal questions, I’m guessing as another way to keep up engagement, and I tended to respond on my own behalf rather than making stuff up. This means the whole time she was coming on to me, she knew I was married, which again, super awkward! The game does flag that it has a romance element, in fairness, but I think it would have been helpful to provide a little more instruction up front about how the player could engage with this aspect – as well as tuning Cheree to make it easier to put her in the friendzone, because she was persistent.

For all that these are real issues, though, they don’t have anything to do with the game’s fundamentals – another game using similar development tools and approaches could easily avoid them. As such, I think Cheree works on its own merits, but also as a positive example of why the IF community might want to engage with AI tools – this game would look radically different if the author had attempted to make it with traditional techniques. Notably, the author’s note indicates that it doesn’t use one of the existing LLMs or a firehose of scraped training data. Cheree is clearly hand-tuned, with human creativity creating strict boundaries for the AI sandbox; add in the smart design and narrative choices that mean the game plays to AI’s strengths rather than its weaknesses and you’ve got a recipe for success.

I’ve tried to write this metaphor three times now, but it keeps getting away from me. Part of me would love to keep tuning it until I get it right, but another, larger part of me would love to clear my review backlog before the Comp kicks off. So we’re going to go with draft number four, no matter how it turns out.

Imagine you’re looking at a cake. And not just any cake – this one’s colored mauve, let’s say, and has little pitchforks of spun sugar crammed into it at odd angles. Now that you look at it more closely, in fact, the whole thing is asymmetric, the various tiers all stacked off-center in a compellingly lurching way, though the structure is impressively solid. You cut a piece, and the cake itself is a marbled red and black with an anarchy sign somehow baked in, the verdigris frosting veining it almost seeming to pulse. Filled with excitement (and maybe a little trepidation) you fork a piece into your mouth – it’s … well, it’s good, and there’s a hint or two of some nonstandard flavoring in there, but actually, it mostly tastes like vanilla cake? Not that you dislike vanilla cake or anything, and again, this is a pretty good example of the form. But wow you were expecting something else.

Okay, I’m not too unhappy with that. I’m going to dive into a more direct discussion of Hinterlands colon Delivered exclamation point now, but keep that cake in mind.

Delivered! is the second in a series of Hinterlands games (the first, Marooned!, was a one-move game with a lot of jokes and even more gore), and its setup promises a slice of blue-collar sci-fi – a driver for the interstellar Parcel Express service, the protagonist gets caught up short on fuel and has to stop off at a godforsaken rock to gas up. Unsurprisingly, what starts out as a straightforward errand quickly winds up going off-track, but the twists and turns are anything but predictable, involving stealing a golden bucket from a sacred well, befriending then betraying some space moonshiners, drug abuse (like, using it for something other than its intended purpose, though yeah, people also use it to get high), mail fraud, freelance espionage, cross-dressing, blasphemous pyromania… oh, and here’s what X ME gets you:

"You are a more or less average specimen of Outer Lumpan heritage. Your leathery skin is pale blue and completely hairless. Your large spherical eyes are solid black. Your mouth is a lipless slit below an eggplant-shaped nose. Your small rounded ears protrude from the sides of your head in a way that is (apparently) comical to anyone who isn’t from Outer Lumpus. On top of your bald head are several dark blue nodules, which your dermatologist assures you are perfectly normal. Your build is tall and lanky, but you slouch (as your mother is always so keen to mention)."

In other words, this isn’t just blue-collar sci-fi, it’s Heavy Metal sci-fi, full of louche characters, oddball aesthetics, and shaggy-dog plots. This subgenre is pretty thin on the ground these days, so it makes for a refreshing change of pace, and Delivered! is a well-implemented take on the theme – the protagonist even after his triumphs remains firmly a loser, it feels like every single character you meet is running a scheme or scam (or is a gigantic dupe), and the setting boasts a scuzzy detail to go alongside every exotic one. And crucially, there’s some actual worldbuilding on display – unlike some games that are wacky for the sake of being wacky, here the setting is there to play host to jokes, rather than just being a joke. Speaking of implementation, things are solid on that front too. There’s a lot of depth in Delivered!’s medium-sized map, with tons of scenery, some complex set-pieces, characters with a ton of non-obvious conversation responses, and prose that communicates the weirdness of the world without getting overly prolix. There is an occasionally bit of guess-the-verb annoyance – I quickly figured out how to (Spoiler - click to show)drill holes in the wall to spy on the neighboring hotel room, but getting the syntax right took some trial and error – but nothing too bad.

So this is overall a fun, unique game that I really enjoyed. I wouldn’t be me if I didn’t add a few critiques, though. First is that the puzzles, while generally reasonable in terms of their difficulty, are occasionally unmotivated. For example, after suffering a certain reverse that meant that the most logical source for more fuel wasn’t going to pan out, I was stymied for what to do next – turns out the solution was to open a certain package, with no indication so far as I could tell that it would have anything to do with resolving my dilemma. And I had a similar experience two or three more times, when I took an action because it was the only thing available to do, rather than because it seemed likely to advance the protagonist’s goals.

As for the other point: remember that cake? Delivered! has a lot of off-kilter aspects, but both the overall structure of the plot (look for MacGuffin, suffer reversals until finally obtaining MacGuffin) and the details of the puzzles (there’s a disguise one, and a couple swap-stuff ones, and some keys to find) actually wind up being pretty normcore. This isn’t necessarily a bad thing – heck, I’m wearing khakis right now! But given the lurid, gutterpunk vibe Delivered! so capably projects, I think I wound up feeling a little disappointed there wasn’t more gonzo on offer. In fact, for the author’s first full-bore piece of IF, the game is admirably controlled. From a design perspective, that was probably the right call, but I can’t help but hope that a notional third Hinterlands game will really cut loose.

Folks, I am too excited to bury the lead on this one – after trying and failing with seven or eight of them, I’ve finally won a Larry Horsfield game! True, it took an excessive amount of save-scumming and UNDO abuse, and it’s clear that the difficulty on this one is pitched way more towards beginners than is typical for his work (modulo one punishing design decision that’s thankfully pretty easy to work around). But I am still going to take my victory lap while I can.

Bug Hunt has one of the oldest video game premises there is – you’re a space marine, there are aliens, go shoot them and win. Much like Xenon-Xevious Resurgence*, it’s part of a larger series of games, though the grounded sense of place I noted in that one didn’t come through as strongly for me here; it really does feel like generic military sci-fi. There are a couple twists in the setup, though, viz: a) instead of playing one space marine, you actually swap between members of a squad, each with a slight difference in skillset or role, and b) rather than terrifying acid-blooded xenomorphs, per the cover art you’re hunting down overgrown but still-cute tardigrades.

After a brief bit of context-setting, the game quickly establishes its structure: the team splits up to explore each corner of a besieged colony, and you need to guide them in turn as they find, and hopefully best, an alien. The vignettes are all quite brief – the longest might take fifteen minutes or so – and all involve classic puzzles, but with a little bit of variety; one involves getting an elevator to work so you can explore an abandoned building, another finding an alien who’s hiding among others in a zoo. And again, none are too challenging on their own – sure, I got jumped a couple times, but some judicious UNDOing was usually enough for me to turn the tables on the beasties. The one wrinkle adding to the difficulty is that there’s a tight 80-turn timer, and some of the larger scenarios can easily eat up 50 or 60 of those to fully explore. Good thing the party separates, so the timer resets every time you swap characters!

Except, er, no, it doesn’t. I assume there’s some technical reason in ADRIFT behind this implementation decision – it’s similarly kind of annoying that you can only save and load the game as the initial protagonist – but as a result, the timer just keeps on clicking linearly as you hop from character to character, as though when the commander said “let’s split up!” and left, everyone shuffled around aimlessly waiting until he radioed back to say he’d killed a bug, at which point one more person left and the whole process repeated itself. Given this constraint, the 80 turn limit goes from tight to ludicrous – I’d imagine a reasonably-efficient playthrough that explored the full play area and checked out every bit of scenery, while solving the puzzles expeditiously, would still easily reach 250 turns or so (that abandoned office building is big – I confess I checked the walkthrough to avoid having to check out a dozenish empty, nondescript locations).

Fortunately it wasn’t too hard for me to savescum my way around the issue: every time I wasted an alien, I restored an earlier save, typed in the optimized path, then saved again. As a result the timer did little more than add a pleasant frisson of challenge, making my victory all the sweeter. Horsfield has written better games, I think – as mentioned, while this one is technically solid and has fair, well-clued puzzles, it doesn’t have as much of the immersive detail I’ve enjoyed in other, harder games – but Bug Hunt is still recognizably of a piece with that larger oeuvre, and so I feel quite satisfied in finally taking the W.

* I know, I know, this isn’t actually its name, but without looking, do you remember what the game’s called?

The RNG has been getting its laughs in – after putting the two games starting with X together, it also gave me the duo of Adventuron games back to back. There the commonalities end, though; for one thing, the protagonist is a Russian soldier fighting “the 19th century Eastern war” (is this Crimea?), but more importantly, while we imagine the protagonist of The Last Mountain as an intensely moral figure, here we’re playing someone with distinctly darker ethics. The opening crawl tells us that after volunteering to fight, Sidorf’s “ideals were quickly replaced by a survival instinct… he wants to do one last thing perfectly. Whatever the cost.”

Going into the game, I had several ideas for what that one last thing might be, and what cost might have to be paid. Points to the author: all of those ideas were completely wrong. This is a story I haven’t seen before, at least in exactly these contours. But partially I think that’s because while Sidorf’s motivations present a compelling enigma, once they begin to resolve things slightly fall apart. To dig into why, I’m going to need to spoil the plot, so fair warning that you might want to finish the game before reading the rest of this review (it’s short and worth playing, in my view).

To provide some padding before getting to the spoilery bits, let’s talk briefly about the mechanics and the prose. We’ve got here a linear series of set pieces, which makes sense because Sidorf is a grunt following orders – he shouldn’t have free rein to wander. The game uses a relatively stripped-down command set, and runs into some of the syntax foibles out-of-the-box Adventuron is prey to, but because it’s quite direct about prompting you about what action you’re supposed to take next, I generally didn’t have too much trouble, with just a few notable exceptions (Spoiler - click to show)(I knew exactly what I wanted to do in the late-game sequence where you need to sneak around the back of a tent and blow up some explosives, but it definitely took some wrestling with the parser to get that across).

The prose takes a similarly blunt approach. BLF boasts a translation credit, so it was clearly written in another language first, but the English is solid enough; the game’s prone to simple, declarative sentences that are closely grounded I the first-person narration, which is an effective way of communicating Sidorf’s voice. There’s the occasional off note – him calling his fellow Russian soldiers “comrades” feels like an anachronistic Soviet-era touch – but overall it fits the game quite well, with the relatively straightforward language not getting in the way of establishing the ambiguities of the protagonist’s desires and goals.

In fact, this combination of straightforward prose and aggressive prompting of the player is doubly important because Sidorf’s motivations turn out to be quite idiosyncratic – if the player were given more freedom or a muddier picture of the situation, the game could easily have turned into a frustrating experience, since they’d almost certainly wind up chasing the wrong goals. Sidorf doesn’t dream of performing an act of heroics, or of surviving to go back home no matter what: no, he’s resigned himself to death in battle, but wants to make sure the last letter home that’s found on his body is the best-written, most compelling letter anyone has ever seen. He’s also fixated on a very specific way of accomplishing that goal: purloining bits and pieces of the letters his fellow soldiers write. Understandably, none of them are especially likely to share their missives back to their sweethearts or tearful farewells to their children with someone who, as it turns out, is quite the socially-awkward weirdo; fortunately for Sidorf if not for the others, he’s willing to go to any extreme to get them to cough up the goods.

The game thus has a regular rhythm to its half-hour runtime: meet a new soldier or soldiers, then follow order for a while until you have a chance to kill them and take their stuff, until you have all the raw material you need to write your masterpiece, bringing the game to a close with a brief narration of Sidorf’s inevitable death. In its favor, this resolution is compellingly demented; against this, though, I simultaneously found it both annoyingly obscure and a little too pat.

On the obscure side of things, besides that one sentence in the intro talking about the death of Sidorf’s ideals and the triumph of the survival instinct, we don’t get any sense of how Sidorf hit on his ideas – they’re just taken as givens (and, one feels obliged to point out, they don’t seem to have anything much to do with survival). Beyond his monomaniacal acquisitive zeal, Sidorf doesn’t have much characterization, and indeed, the climax feels frustratingly anticlimactic. The contents of the final letter are never so much as hinted at, nor do we get any clue about who Sidorf’s family are, what they might think of what he’s going to tell them, or why he’s so concerned with making such an impression on them (and again, we don’t even know what war this is!).

Sure, to a certain extent this is beside the point; psychologizing a character who’s clearly meant to be an allegorical figure risks crushing an intellectual argument with banality (the game would hardly be more compelling if we found out, say, that Sidorf is desperate to impress a father who used to beat him). And there’s no indication that there’s something about the contingent facts around this particular historical conflict that brought on his mania. But to my mind fiction works best when it manages to ground its ideas in personality; to stick with the game’s milieu, Tolstory is surely working with abstractions in War and Peace, but the novel has survived because Pierre, Natasha, and the others feel like specific, idiosyncratic characters with depths that go beyond their mere function as elements in an argument about history. We don’t get anything like that here, and so Sidorf dies as he lives: a cipher.

As for the other way of looking at the game’s themes: I mean, writers are vampires, film at 11. To its credit, BLF stages this idea in a novel way, but as far as I was able to engage with the game, the novelty felt only skin-deep, and actually winds up undercutting the effectiveness of the argument. Like, even if we consider the intensely negative case of an amoral author who takes the stories or emotional trauma of their loved ones and turns them into a crass commercial product, we’d still say “compared to Sidorf, this isn’t so bad!” The things authors do to get ideas or inspiration from others don’t look very much like the stuff Sidorf does so he can steal letters from the corpses of his friends, so while the parallels may work intellectually, they feel schematic rather than visceral.

I always like to see parser games that are going for a literary effect, and BLF certainly looks good on that score; the plot, characters, writing, puzzles, and gameplay are very clearly arranged to advance a very specific set of themes. But for all the grubbiness of Sidorf’s experience, I found the perspective offered wound up being too high-level; there’s not enough blood in the veins to make the various dilemmas and atrocities here truly sing.

Here’s one of my pet theories of IF that I’m not sure I’ve written down before: there should be more parser games about sports. This isn’t due to any native affinity for them – more power to those who are into sports, I could care less about any professional teams, I only did a real sport for two semesters for all my high school and college years, and I’m the kind of schmuck who thinks it’s funny to respond “Interpol investigations” when the check-in question at a work meeting is “what’s your favorite Olympic event?”

No, it’s because of that old writing adage that action reveals character. We can get told that a character is clever or cowardly or chokes under pressure or what not, but until that gets on screen in some way – meaning, in a game, that they take some action that demonstrates the trait – it’s all theoretical. The trouble is, the sorts of character traits that can be revealed by the business of a typical parser game are fairly limited by the medium-dry-goods world model that tends to dominate: “resourceful” and “kleptomaniac” can only take you so far. Then consider that for a linear puzzle game, beyond the difficulty of coming up with and implementing multiple solutions to puzzles, it can also be a challenge for authors to invent reasons why different approaches might actually matter in narrative terms.

Sports offer a fresh way of engaging with these problems: beyond the fact that they create a rules-based framework that supports novel kinds of gameplay, their victory or scoring conditions also offer built-in consequences for a player’s choices, meaning that discrete, relatively-easy-to-implement physical actions can be freighted with narrative and/or thematic weight (This, by the way, is why my dark-horse pick for TV shows that totally should have gotten an RPG is Friday Night Lights). To be clear, I’m not saying that Madden 2023 would clean up at IFComp or anything – but that I do think there’s a lot of potential in parser games that use sports rather than conventional puzzles as their main gameplay elements.

Anyway, I wish that a) I’d written this theory down before playing The Last Mountain, and b) that I could count it as vindication of said theory, when the truth is that it could just be that a talented author like Dee Cooke can make any of their ideas look genius.

Yes, you might have lost track of the fact that this is technically a review somewhere in the previous four paragraphs of maundering, but I swear, these thoughts are relevant to understanding why this Adventuron game works so well, and feels (at least to me) so unique. The setup certainly isn’t one you’ve heard before: the player midway through a long-distance foot race with their running partner, Susan, who’s uncharacteristically flagging early as you tackle the last mountain before the finish line. You’ve got a water bottle, a flashlight, the race directions (there’s an orienteering component), and some walking poles, and with those you need to overcome a series of obstacles – getting tired, losing the trail, facing one last steep descent. Some of them are decision-points, some are inventory puzzles, and none on their own is that innovative – but again, the fact that they’re all happening in a race rationalizes the barriers, and adds a compelling urgency to solve them quickly.

Susan is the other part of the equation. The game deftly sketches your relationship with her – she’s somewhere between a friend and a mentor who helped bring you into this racing hobby – and presents her uncharacteristic fatigue as a central dilemma of the game. Again and again, you’re faced with the option (and Susan’s explicit prompting) to leave her behind so you can get a good finishing time. I’m guessing that most players won’t be tempted to ditch her, but still, the fact that the choice is there lends added weight to the individual puzzles.

The prose thus has to accomplish a lot of different things: create a sense of place, of course, while making sure to foreground Susan’s presence and give the player everything they need to engage with the game. It’s thankfully up to the task, and accomplishes all this with economy and without getting showy, too. Here’s a bit of mid-game scene-setting I especially liked:

"As the trees become denser, you realise how dark this forest can start to feel when the daylight isn’t so bright. You’ve never been here so late before. It makes it really difficult to identify the right path, even with Susan’s keen sense of direction.

"The forest has become really dense here. The smell of dry branches and the hooting of birds surround you, making you feel a little claustrophobic."

My one kick against the game is that I experienced a few guess-the-verb struggles, partially born of my own lack of experience with this kind of running, but also partially because the parser didn’t feel like it was meeting me halfway as I flailed about trying to figure out how to use the walking poles (yes, for those of you who have played the sailing sequence in Sting, I am aware of the irony of me of all people whining about this). Beyond that, I suppose one might complain that the player will guess what’s up with Susan well before the ending – but I don’t think that’s actually a fault with the game; the tension between the player’s suspicion that it’s serious, and the protagonist’s urge to do well in the race, is another piece of the engine that helps make it work so well. And for all that the reveal wasn’t a surprise, I still felt that it had emotional heft when it landed.

All of this is to say that I found The Last Mountain very good on its own merits, as well as instructive for the directions I think it suggests for future works, which is exactly the sort of thing one wants to come across in ParserComp!

Judging from the title, you’d be forgiven for thinking this game would be a forgettable throwback, puzzler where you wander around an ersatz fantasy environment, solving simple puzzles and looking for valuables to hoover up for no reason other than that they’re there, all the while enjoying/enduring various wacky scenarios and overenthusiastic jokes. And Steal 10 Treasures isn’t not that, certainly; yup, there’s a castle; no, there’s rationale for you to be raiding it; yes, the first puzzle involves refusing a poisoned ice cream; no, it doesn’t get any less silly from there. Still, it’s anything but generic, and merits inclusion in the freestyle category by dint of its interface: you type commands just like a regular parser game, sure, but it only recognizes actions that are a single letter long.

To give an example of how this works, instead of writing out a full command, you just type, say, A which if you happen to be in the dungeon would get interpreted contextually to ATTACK CLAM (I told you about the wacky scenarios). Or S might get you SMELL CLAM PLEASE; meanwhile, since there’s nothing to move around down there, pushing P just pops up PUSH ANYTHING, which unsurprisingly accomplish much when you hit enter. Navigation, meanwhile, is handled via the arrow keys.

This is a limited parser game, in other words – something I’ve had on my mind of late 5 – but a peculiar sort of one. Outside of navigation and out-of-game commands, there are about a dozen actions on offer, running the usual parser-puzzler gamut (including LICK, as I understand is becoming the style), and while the help screen doesn’t tell you all of them, since it only takes a minute to try out all the keys on the keyboard to learn the “secret” commands, the player generally knows exactly what their options are.

That’s the theory, at least – in practice, I often found myself at a bit of a loss for what to type. There are too many possible actions to be easily held in the head at once, and because many of the commands start with the same letter, the keyboard mapping sometimes felt about as intuitive as that of an early Ultima game (Ztats, anyone?) If P is push, then Y must be pull – so that means B is yell? C for climb is intuitive enough, as is T for turn, but then you’ve got V for converse. And sometimes the game seems willfully perverse: G isn’t mapped to anything, but rather than using that, you need to type a period to get an item. The result of all of this is that when I entered a new room and was confronted with a new situation, my first instinct was to just start hammering out QWERTY and continuing from there until I found an option that looked good.

I ran into the lawnmowering problem, in other words, where the player turns off their brain and tries to make progress by mechanically trying every choice until they hit on one that works. As I discuss in my Rosebush article, there are various strategies limited-parser games can use to make this approach less appealing – it’s a little gauche to keep flogging it, but I feel like you, specifically, would really enjoy it – like timing puzzles, actions that are contingent on the presence or absence of different NPCs, or concealed second-order actions, but Steal 10 Treasures doesn’t employ any of them.

This is a real kick against it, but I’m compelled to note that in practice, even as one part of me was cataloguing the ways the design didn’t quite work, another part was just enjoying the ride. Sure, silly treasure-hunts are played out at this late date, but the reason they’ve stuck around so long is that they can be a lot of fun. And the game’s gags and puzzles are solid enough to carry it pretty far – it’s just big enough to avoid being trivial without being so sprawling that it gets annoying, does a good job of clueing its puzzles and alternating big, multi-step ones with short, easy ones (I especially liked the decidedly non-standard way you deal with the dragon), and the jokes adeptly ride the line between wacky-silly and wacky-ridiculous.

As a result, the single-letter gimmick didn’t wind up being as much of a downside as I thought it’d be; it might have even wound up being a plus, making it easier for me to kick back and enjoy the ride. Not every game needs to be Hadean Lands: if all you’re after is beer and pretzels, isn’t it nicer to just lift a finger to signal for another round, rather than having to spell out your order every time?

Andrew Schultz has by this point created quite the collection of IF chess puzzlers – this is I think the fifth one I’ve played and reviewed? There are commonalities between all of them, of course: most notably, they’re all impeccably presented, with multiple helpful ways of displaying the board, accessibility options and hints to allow for maximum ease of play, and a light but engaging patina of story adding some narrative sugar to what could otherwise be dry exercises in logic. Impressively, while XOU is no exception, it doesn’t feel like a retread – unlike the earlier games, which hinged on proper piece placement or clever pawn-promotion tricks, or started out with the player on the back foot, here we’ve got a classic but tricky endgame scenario: the player’s got to achieve checkmate with only their two bishops.

It’s a well-chosen setup because it allows for a fun narrative layer, hinging on the difficulties of getting the bishop-who-only-goes-on-white-squares and the bishop-who-only-goes-on-black-squares to put their differences aside and work together. It also makes for a deceptively challenging puzzle. With the opponent having only a king to their name, you’re obviously in no danger, but it’s surprisingly easy for them to slip through your offense and force a stalemate – or even, since you can only defend your bishops with your king, knock out one of your pieces (this is still just a stalemate, of course, but it’s a much more humiliating way to go down).

As I’ve mentioned in my previous reviews of this series of games, I’m no chess maven but I’ve generally found a way to muddle through. That’s technically the case here too, though I’ll have to cop to rather more muddle than usual. It didn’t take me too long to crack the first phase of the puzzle, and it was fun to scissor my bishops past each other until they pinned the enemy king against the edge of the board. The process of herding the king into the corner for mate, though, was a much harder nut to crack, involving a forward-and-back pas de trois that I only groped my way towards through a whole lot of trial and error – somehow keeping track of all those diagonals was very taxing on my poor brain. When looking back at the solution in retrospect, it’s lovely and elegant, but it sure didn’t feel that way at the time.

That’s probably more an indication that I’ve found my level as a chess dilettante than a real critique of the game, though – and I’m guessing that for those with more familiarity with the game of kings, this stepped-up difficulty could well be a selling point. My favorite of these games remains You Won’t Get Her Back, which I think nailed a sweet spot in terms of difficulty while also having the cleverest marriage of gameplay and narrative, but XOU is a worthy addition to the collection too – if the concept seems at all appealing, you really can’t go wrong.

This is one of three ParserComp entries by the author, which is the kind of work ethic that I feel like I can’t directly comment on without being consumed by jealousy. Each is an old-school ADRIFT puzzler, of various flavors – here, we’ve got another installment in the author’s long-running Alaric Blackmoon series of fantasy games. While there’s some continuity with earlier entries, with references to previous adventure sprinkled throughout the opening, XXR (you’d better believe I’m not typing that title again) seems to work quite well as a standalone, with a straightforward and engaging premise: there are rumors of monsters on the periphery, so you and your buddy the king need to cross a desert to check things out.

I’ve played some of those earlier games, but never to completion, and I was hoping this would be the one to break the streak – but alas, it was not to be. All of them have a fine-grained style that require the player to spell out exactly what they’re doing, step by step, rather than bottom line their actions. On the positive side, this contributes to a pleasant sense of immersion; I enjoyed the low-key opening section, where you need to barter for camels and equipment for your desert trip. Sure it’s a little fiddly to have to buy the transportation but then visit separate vendors to get saddlebags and tackle, then purchase clothes appropriate to the desert heat, but it helps sell the reality of the world, and establish that the player characters – you can swap between Alaric and the king whenever you like – are going to be fish out of water (er) on their trip.

On the downside, though, this granularity combined with some of the foibles of the ADRIFT parser to make the puzzles even harder than I think they’re intended to be. The first major section of gameplay involves exploring a ruined city where you’ve taken shelter from a sandstorm, which ultimately requires using an abandoned metallurgical workshop to duplicate a key. While it wasn’t too tough to figure out what I was supposed to do in general terms, each step involved wrestling with the parser. A key item can be found in the debris lying around the place, but SEARCH doesn’t reveal it – instead, you need to CLEAN WORKSHOP (I feel like cleaning abandoned workshops is right up there with cleaning a rental car in the implausibility sweepstakes). Similarly, getting water into the quenching trough is a bit of a struggle:

"> FILL TROUGH WITH WATERSKIN
You cannot fill anything with the water skin.
> FILL TROUGH WITH WATER
You pour some water into the trough from your water skin."

So I was able to make some progress, and found some intriguing secrets in the city, but eventually my progress petered out; the game does include context-specific hints, but through some combination of the system seeming to get confused with a different puzzle and/or me being too thick to figure out what I was missing, it couldn’t get me on track. This is a shame since I did enjoy aspects of the world, but between wrestling with the parser and the punishing puzzles – as well as an annoying quirk of ADRIFT that meant that I couldn’t reload the game when I died while playing as the king – I wasn’t too sad to wash my hands of it. Besides, even just in this Comp I’ll have two more chances to finally get through one of Horsfield’s games!

For the past year or so, the IF community has seen numerous conversations about the ethics, efficacy, and prospects for using AI tools like Large Language Models or image-generation software to create IF. Various arguments have been advanced over epically long threads – often throwing off as much heat as light, it must be admitted – but perhaps our time would have been better spent waiting quietly, because sometimes an ounce of practice is worth a pound of theory. The Fortuna is perhaps a maximalist take on what using AI in IF can look like: every element of it, from the graphics illustrating each location and character, to the descriptions that flesh out its cruise-ship milieu, to the freeform conversation system that’s central to progression, is built around AI. And every single one is awful.

Looks, cards on the table: I come to this debate pretty skeptical of the AI pitch. One of the major reasons I engage with art is because it offers an opportunity to connect with other human minds, to expand my understanding and my perspectives, to experience something idiosyncratic and specific to the person who made it – LLMs and other AI approaches, with their views from nowhere, get in the way of that. I also find most procedural-generation pretty boring; I get that in theory it can open up new intellectual possibilities and reveal surprising connections, but in practice I tend to experience it as one bowl of oatmeal after another. So I may have some biases (I also am not really big on cruise ships, now that I think about it – never been on one, so my major associations with cruises are the destruction of Venice’s lagoon, David Foster Wallace’s A Supposedly Fun Thing I’ll Never Do Again, and the time my brother-in-law went on one and came back with swine flu). But even if I’d gone into The Fortuna a dyed-in-the-wool techno-optimist, what I found there would give me second thoughts.

Start with the art. Due to the use of AI, the game is lavishly illustrated by the standards of amateur IF. But instead of providing a nice aid to the imagination, allowing the player to more fully immerse themself in the world, instead we get, well, this:

Look upon my works, ye mighty and despair

Allegedly this is a set of greeters for the cruise ship, not a band of Cenobites on holiday from hell. Allegedly. But it’s not an outlier; all of the images are like this to a greater or lesser extent.

It’s hard to discern exactly what’s gone wrong here – obviously the image-generation tool got confused somehow. But aren’t there other tools that work better that the author could have used? And shouldn’t he have exercised a minimum of quality control and noticed how disturbing this picture is, how out of step it is with the happy vacation vibe the prose is working to convey, and killed it?

Speaking of the writing, it suffers from the same uncanny valley issues as the visuals; the well-known tendency of LLMs to hallucinate is on full display, and the prose manages to be vague and repetitive. Like, here’s the description of a uniform in a gym locker:

"The uniform in the gym locker room is a set of clothing worn by individuals for a specific purpose or profession. It is likely made of durable and breathable material to withstand physical activities. The uniform may include a shirt, shorts or pants, and possibly a jacket or other accessories depending on the specific sport or activity taking place in the gym. It may also have the logo or name of the cruise ship or gym embroidered on it. The uniform is typically neatly folded or hung in one of the lockers, ready to be worn by individuals engaging in physical activities in the gym."

This is a hundred words that don’t communicate anything beyond “a uniform in a gym locker”, and of course there are an awful lot of “may”s here – in fact, the uniform may not actually exist, given that I couldn’t try to take it.

The writing issues combine with the game’s conversation-focused design to create still further problems. In several places, progression requires you to build a rapport with various of the characters you meet, which is largely accomplished by asking them about themselves. This is accomplished via a freeform chat system, which is impressively open-ended and does allow for a wide variety of input to be understood (one character writes for the New York Times, and when I asked her about Jamelle Bouie, she correctly identified him as a journalist). Still, this element too runs into some familiar problems. The left hand may not know what the left is up to – as with one early character, who’s described as toying with their wedding ring in a bit of environmental detail, but who steadfastly insists they were never married when you ask them about it. And when you do learn more about them, well:

"> ask officer “how are you”

"I’m doing well, thank you for asking. I believe in hard work and dedication, and I strive to make the world a safer place as a port security officer. Ensuring that everyone can feel safe is important to me, and I hope to be a role model for young people. Growing up in a rough neighborhood and seeing the violence and crime firsthand motivated me to make a difference. After high school, I joined the military and served in Iraq for four years. I then pursued a degree in criminal justice and landed a job in port security. I want to show others, especially young people, that they can achieve anything they set their minds to."

This is generic and robotic, of course, but it’s maybe worth noting that this character is Black, and I dare say there’s something pretty stereotypical about the backstory being presented. The one other Black character has a similarly tormented backstory involving struggles with addiction. Meanwhile, when I asked the Times reporter about her job, after saying it’s a dream come true, she continued with “[a]s a female reporter, I am proud to work for a newspaper that values diversity and promotes women’s rights.”

To just state what I’m getting at: the characters who aren’t white men are largely stereotypes, defined by their race and gender. This is something a human author can unwittingly wind up doing, of course – but it’s also a real danger of LLMs, trained as they are on the products of a racist, misogynistic society. At least there’s a note of comedy from how the model extrapolates without any understanding of the world: there’s a devout Italian woman who’s super into C.S. Lewis, because the AI’s mashed up different stereotypes without knowing the difference between Catholics and Protestants.

Now, it’s possible that these characters get deeper over the course of the game, and that it improves in some of its other problematic aspects too, because I have to confess that I didn’t get very far into the plot. After boarding the ship, entering the VIP section, and being introduced to the cast of characters, there was an inciting incident that pointed to a mystery to be solved that presumably kicks off the narrative proper – but one of the first steps required solving a riddle to open a safe. The riddle seemed like one of those old chestnuts everyone knows, albeit with awkward, AI-y syntax (Spoiler - click to show)(”I am owned by the poor and the rich don’t need me” – it’s nothing, right?) except the obvious answer didn’t work, nor did a bunch of non-obvious ones I tried. Is this because I was typing things in wrong, or being especially thick, or did the AI generation flub the riddle? I don’t know, but regardless, there my journey came to an end, and I must admit to feeling some relief.

I am being uncharacteristically harsh on The Fortuna, and I should say it’s not because I think the game is coming from a bad place by any means. In fact, the best piece of writing in the game is the introduction, where the author speaks movingly about being inspired by IF as a child, and wanting to use the new possibilities afforded by AI tools to create a game that would similarly spark joy in its players. That’s a laudable goal, and one that pretty much all first-time authors fall short of (I know I certainly did!)

But I do think there’s something very badly broken about the approach here. While there may be a case for the IF community to embrace AI, it turns out that using it for every aspect of a project and dialing things up to 11 does not make for a convincing demonstration. The Fortuna’s various failures all, I think, have one root cause – which is that time after time, when an AI tool threw up something that didn’t work or didn’t make sense, the author didn’t take action to cut, modify, or customize in service of the larger vision. As the amateur IF scene is currently constituted, it’s an auteurist medium; one can certainly levy critiques about that situation, but as it turns out, taking the author out of the equation is a very bad idea.

Here’s one of the iron laws of interactive fiction: you are in good hands with Garry Francis. There may be a few other contemporary authors working at a similar clip, but even fewer, I think, hit the same consistently high level of quality, serving up adventures that might be old-school in their premises but boast airtight implementation, clear and engaging prose, and solidly-designed puzzles. Yet even judged against these high standards, Search for the Lost Ark is a standout, delivering a game that’s polished, funny, and satisfying.

From the title you might think that we’re in for an Indiana Jones style globetrotting adventure, but the actual setup, delightfully, is both more grounded and wackier: the Lost Ark was found long ago and had been hanging out in Chartres’ cathedral for several centuries, until being moved first to a village church and then – out of an admirable but perhaps overzealous protective instinct, Chartres being west of Paris – during World War I it was hidden in the nearby woods to keep it safe from German marauders. Now, as a priest-in-training who grew up in the area, you’ve been ordered back home to find the thing after the clergyman who hid it shuffles off to join the choir invisible. The only problem is, you’ve no idea where to start, and there’s something off about your immediate superior, the rather-wan, just-arrived-from-Eastern-Europe “Father” “Alucard.”

So yeah no points for guessing the plot twist, but this isn’t the sort of game that’s relying on the plot for engagement – and it knows you know that, meaning Alucard will engage in a bit of knowing vamping if you care to toss some pointed dialogue queries his way but you’re not going to short-circuit any puzzles by dint of your genre-awareness. Similarly, the writing does a perfect job of conveying a sense of place and highlighting the details you’ll need to focus on to complete your quest, all in a terse yet informative style that’s a model of effective prose. This is no mean feat, especially since it also has to communicate information that you, as a priest-to-be, would know about subjects like the structure of the Bible or the details of church architecture, but that you, as a player of IF, might not. Many of these tidbits are actually relevant to solving the game, but some are just lovely little factoids:

"> x hammer

"The sledge hammer has a heavy iron head and a long wooden handle. This combination makes it good for breaking big rocks into little rocks. Did you know that ‘sledge’ is derived from the old English ‘slægan’, which means ‘to strike violently’. No? Well, now you do."

I didn’t, so yes, now I do!

As for those puzzles, none are especially hard, but gosh, are they satisfying. There are a series of clearly-signposted obstacles, each with intuitive solutions, and if this were all there was to the game, it might risk feeling a bit too old-school, in a USE X ON Y sort of way, but there’s also a final metapuzzle: as you progress, you’ll come across a series of biblical verses, which you need to combine to reach your ultimate goal. I won’t spoil the details, but it takes just the right amount of thought, and while it rewards a bit of outside knowledge, the game characteristically provides everything you need even if the difference between Corinthians and Thessalonians is all Greek to you.

Writing this review with the benefit of hindsight, I’m not at all surprised that Search for the Lost Ark shared the gold medal in the classic category – this is a near-perfect execution of the traditional form, thoughtful and engaging and not overstaying its welcome. The only problem is, now the next time I play a Garry Francis game, my standards will be even higher!