Picton Murder Whodunnit

I can certainly see the appeal of the randomized murder-mystery. More than most genres of IF, once you know the solution to a mystery there’s not much to hold interest on subsequent replays – and even on the initial play-through, if the author’s telegraphed the true culprit too strongly it might be even less compelling. On the other hand, mysteries tend to be really engaging puzzles for players. If you could write a good mystery that could be randomized, so the player knows the game is playing fair and can be surprised the second, third, fourth, and fifth time they run through it – that’s a game you could enjoy for a long time.

While I think I’ve played maybe half a dozen iterations of the concept, though, they’ve always left me cold. Some of this is I think is down to the fact that I tend to be less interested in randomized narratives – I’ll gladly sink untold hours in a pure, zero-story roguelike, and I’ve 100%’ed every single Assassin’s Creed game save the most recent one. But even in those often-grindy games, if there’s a system for randomly generating quests, my brain just flips a switch and is completely uninterested, even if functionally the random ones are almost exactly the same as the bespoke quests I just put a hundred hours into playing through. But I think my diffidence at the sub-genre isn’t due to personal preference alone: it’s hard enough to write one mystery, much less a mystery that can be reshuffled multiple times and still be satisfying.

Picton Murder Whodunnit – yes, we’re finally getting around to it – is, of course, a randomized murder mystery. It’s built using the Strand system, which I wasn’t previously familiar with, but looks like it was designed to create updated versions of some of the old Magnetic Scrolls games. Anyway I like the engine well enough, offering the option of choice-based or parser-based interaction, though there isn’t the ability to play offline so far as I could tell and the online version was sometimes laggy.

The conceit here is about as traditional as you can get: you play a police inspector called to a country manor to investigate the suspicious death of a peer of the realm, and you’ve got to identify which of a quartet of suspects (the conniving widow, the vicious son, the grasping brother, or of course the supercilious butler) did the deed. The game discloses that the solution is randomized each time, so while e.g. the widow is always portrayed as conniving, only one time out of four does this tip over into a murderous motivation.

As you can tell from the characters straight out of central casting (the butler’s even named Jeeves), the milieu is spot on, and the writing is full of cheerfully over the top Britishisms that I quite enjoyed – the brother is described as wearing a “pompous cravat and tweedy, shoulder-patched green shooting jacket,” which definitely conjures the character. This is where the randomized nature of the game starts to pose problems, though, as every character is described in dark, unpleasant terms, even if that seems to make little sense. Here’s the ten-year-old son, Jimmy: “piercing, piggy blue eyes stare back at you fiercely. You get the impression he totally despises the police and you in particular. He’s probably guilty as Hell!” I mean, steady on there, matey, he’s not even a tween. But in order for the randomization to work, everyone has an equally-plausible motive, and everyone has a key to the gun cupboard (yes, even little Jimmy) which I thought felt artificial – at least give us one character who doesn’t seem to have a reason to off the Major!

The investigation is also less fun than I wanted it to be, though here the randomization is only partly to blame. There isn’t any physical evidence to examine, nor are there any clues to uncover or forensic details to analyze. Solving the mystery reduces to asking every character about their alibi, then doing another round to ask about everyone else’s alibi to see who’s the odd one out. This is made slightly more difficult by the fact that the murderer, of course, is happy to lie, and by the fact that I found some of the clues ambiguous, though possibly that’s down to me not fully understanding the manor’s layout rather than fuzzy writing. The alibis are also functionally the same in repeat playthroughs (like, maybe the brother will be writing a novel instead of smoking a pipe, but he’s always in the drawing room and always relies on Jimmy for corroboration), making the investigation feel repetitive even though the ultimate culprit may be different. This is especially the case because after the first run-through, which took maybe ten minutes, I managed subsequent replays in maybe two minutes apiece since there are so few things that need doing.

All of which is to say that while Picton Murder Mystery works fine and supports at least one or two fun go-rounds, the nut of the tightly-plotted but randomized mystery remains uncracked, and I’d personally trade it for a non-replayable but deeper investigation with the same setting and characters in a heartbeat.

Picton Murder Whodunnit is a short and sparse random-perpetrator murder mystery. I played through it only once, so I can't really comment on the mechanics of the random assignment of the murderer.

The setting of an upper-class mansion with its inhabitants (and a butler!) is something that inspires high expectations in me. There are many opportunities for drama and/or humour here, in the story/plot as well as in the characterization. Unfortunately, the game let me down on both counts.

Storywise, Picton Murder Whodunnit is just too sparse. There is no background on the murder or the victim. The family's affairs are given naught but the faintest of hints. Even within the constraints of the randomization, a vague account of the events leading up to the murder would have been very welcome.

An upper-class mystery-game of this kind stands or falls, in my opinion, by the pointed characterization of the people involved. There is a lot of room for either a deep psychological profile or a funny charicature of the well known rich family trope-characters.
In the game however, there is only the highly prejudiced opinion of the PC, which I found rather unprofessional for the expert-investigator the introduction claims he is.

Perhaps the randomization mechanism would blow my mind upon replay. I will not know because I will not replay. There just isn't enough there.

This game is written using the Strand Games engine, a nice and smooth interface that was developed to 'restore' the Magnetic Scrolls games from a few decades ago.

I like the way it looks and moves, it feels very smooth. The voice acting option (I think it was different text-to-speech readers, at least for some of them?) are a nice change of pace after how hard it is to get sound working in Inform.

The game itself isn't quite as alluring. It's a collection of fairly tropey characters in a fairly tropey setting (a major, a dilettante child, a butler, etc. in a manor). And the puzzle seems to consist in just asking who was where and figuring out which one person was lying.

The game is pretty easy. The murderer and the alibis are selected at random. (Spoiler - click to show)Only the murderer has no alibi, falsely claiming "I was in place X; character Y saw me there" but when you ask character Y, Y will say that they didn't see the murderer. You go outside, accuse one person, and the game ends.

I would have liked the puzzle to include a little more depth, something that gave me a sense of surprise (while still being solvable in hindsight).

I don't know why, but in Chrome 89 for macOS 11.2.3, the sound effects were horribly choppy. I switched to Safari 14 and it worked much better. It appears that opening this game (or any Strand game?) pins a CPU in Chrome.

The voices sounded computer generated, which was weird, but sometimes they really worked. I'd be curious to learn more about how the author made these sounds.