This is literally just copy-pasted from a post on my blog, but I felt as though it was worth getting the words on here directly. I hope that’s okay!
At 50,000 words in length, this was my second largest ever work of interactive fiction, and easily the largest parser work specifically. Given how much effort went into it, I’m incredibly pleased with the result. As nice as it would have been to crack the IFComp top ten, I kind of appreciate that I can keep that goal for next year. I’m not exactly running out of worlds to conquer here, but reaching the top ten feels like something that I can realistically shoot for with an ambitious enough game. Aiming for top five or number one essentially becomes a bit of a lottery, with the highest rated games needing to average a score somewhere north of an eight out of ten. At that point success seems less a matter of wowing judges (since the critical consensus is already “this is just about perfect”) and more a matter of hoping that a handful of low outlier votes don’t scupper your chances.
This evidently wasn’t a problem for Who Whacked Jimmy Piñata? this year, given that it had not just the lowest standard deviation of any of my games ever, but the lowest standard deviation of anything submitted in 2025. Remember when I intentionally divided opinion to win the Golden Banana of Discord? Here I have unintentionally done the opposite. There was talk of actually establishing this as an award last year – I put forward The Iron Potato of Accord as a name – and now I’m sort of disappointed it never happened. There are good reasons for avoiding such an award, though: my placing of 12th here is actually pretty rare, as historically the most common reason for a game having a low standard deviation is “Everyone agreed it was rubbish” and handing out an award for that is just rubbing salt in the wound.
In terms of how Who Whacked Jimmy Piñata? managed to achieve both a high placement and a low standard deviation, I think there are two big reasons. One is that last year’s Quest for the Teacup of Minor Sentimental Value placed highly enough that I was able to claim a code review by Zed Lopez as a prize, and he offered a huge amount of help. I wasn’t actually that familiar with the Inform programming language when I started the game, but I feel as though I’ve got a much better handle on it thanks to him. I also devoted quite a bit more effort to getting feedback from playtesters this time around, and they provided me with lengthy transcripts of what they’d typed in and got back: there’s a lot of dialogue in the final version of the game, and quite a chunk of it came from seeing people typing in “ASK SO-AND-SO ABOUT [SENSIBLE TOPIC]” and getting a generic “There is no reply” in response. Where remotely possible, I tried to ensure that anything a player had attempted during testing would produce a satisfying response if a player attempted it during the judging period.
The other reason I think Who Whacked Jimmy Piñata? got the result it did is that it built quite significantly upon the earlier Bubble Gumshoe games: Who Shot Gum E. Bear? and Who Iced Mayor McFreeze?. The critical response to Who Shot Gum E. Bear? was a surprise because I thought I had a fun idea for an unconventional murder mystery, with the Sugar City setting being a throwaway means of making it work: it needed a handful of quirky suspects to talk to, and Don Toblerone et al. fit the bill. Essentially, I thought people would enjoy the mystery and be indifferent to the setting and characters. In the end, people hated the mystery (expressing a preference for more traditional puzzles) but loved the set dressing. Who Iced Mayor McFreeze? was a response to that – giving Bubble Gumshoe some conventional puzzles to solve – but taking place in a more closed-off area of Sugar City that didn’t require me to write dialogue for interviewable characters (which would have been impractical in the time available).
Who Whacked Jimmy Piñata? combines elements of both games – including the map, which allows Bubble Gumshoe to travel from the docks of Who Iced Mayor McFreeze? to the street and businesses from Who Shot Gum E. Bear? via an abandoned building. For the sake of keeping a consistent setting, I tried to reuse as many elements as possible from the earlier games: traces of both previous murders can still be found, and the puzzles tend to revolve around objects and locations that had already been established. Who Shot Gum E. Bear? included a number of red herrings – the newspaper vending box, the VIP room in the nightclub, Don Toblerone’s candy cigarettes – that could be interacted with to an extent but didn’t actually form part of a puzzle or yield any information regarding the mystery. At least a couple of reviewers who took issue with this with the earlier game suggested it seemed as though I’d had their comments to hand while working on this one, and though I did have a print-out of the feedback for Mayor McFreeze next to me while working, a much bigger factor was simply that I was trying to add more puzzles while also reusing existing details. It’s not so much that I was thinking “Reviewer X specifically wanted to be able to do something with the candy cigarettes…” as that I was thinking “I need a key item for this puzzle” and also I’d already established that Don Toblerone’s desk has a pack of ciggies on it just waiting to be taken.
I can’t exactly say “This is the game that Bubble Gumshoe was always meant to be” because it’s actually drifted quite a long way from my original intent, but it has been nice to see the series go from something that had people saying “I was surprised to see a sequel to this” to something that people have expressed an interest in seeing more of. There will be more, though not necessarily in 2026 specifically because I’ve already started working on something else. IFComp was really what got me into parser-based interactive fiction – I find it funny to hear that there was once a panic over whether Twine would see choice games fully replace parser ones, because I never would have learned Inform without Twine – and I’m keen to see how much more polish I can add to Bubble Gumshoe. Though the response this year has been fantastic, it certainly leaves room for improvement: the game didn’t get a single ten rating, and I think the quality of the implementation is still probably the main area to work on. I learned a lot while making it and there’s plenty I would do differently if starting over today (but only if starting over – it would be a gargantuan amount of effort to rework it at this point). Currently I feel as though the jokes and bizarre setting are probably making up for a certain amount of friction, so if I can make the game play more smoothly I think that’ll be a significant boost overall.
