EJ and I meant to get this up earlier, but alas: life got in the way. This is also a real chonker of a post-mortem so thanks in advance to anyone who reads it! This was written by us collaboratively, with my thoughts at the end and EJ’s to follow in another post.
To talk about this game we first have to back up and discuss our previous work, Lady Thalia and the Seraskier Sapphires. (We never wrote a post-mortem on that one, so this is fair game, right?) Particularly, we have to discuss the circumstances in which that game came about.
EJ, who may have played too many visual novels and/or RPGs heavy on the conversation mechanics at an impressionable age, is most comfortable writing conversational/social puzzles (she even tried to include them in her parser games back in the day, although they were not very well-executed – Twine makes it much easier). After Social Lycanthropy Disorder, which was nothing but that, and with the more mechanically-minded Encorm as a proper coauthor, we decided we wanted to branch out.
That said, we didn’t want to go too far on an experimental first outing so we came up with the idea of a game that had two “phases” – one focusing on social puzzles, and one on more physical object-interaction style stuff. The idea to make it a heist game followed on from there, and thus Lady Thalia was born. Since 1) the game was destined for the lower-key Spring Thing rather than IFComp and 2) we didn’t really think many people would actually play it, we then proceeded to go ham and cram every single thing we like into the game. Period English setting? Check. Unnaturally witty prose? Check. Tropey lesbian enemies-to-lovers slow-burn romance? Check, check, check!
And then, people liked it! So we decided to write a sequel. (Well, okay, we had kinda always planned to, but if it had been a flop we would have cut our losses.) But first, we had to figure out: what the hell did we actually do, and how do we do it again???
After some deliberation we isolated the following points as integral to what made LTSS work besides the more obvious gameplay systems:
- No “game overs” – if you make too many mistakes, the PC is disappointed in herself and other characters are disappointed in her (and you miss out on an extra scene at the end), but the momentum of the game is never broken
- Writing-wise, an ethos of combining silly tropes and humor with rigorous historical research and attention to detail (the intersection of trashy and highbrow, as Encorm describes it)
- Designing the game entirely around what we personally would like and not tailoring things to any expected audience
The last bullet point is obviously hard to pull off when writing a sequel, and is frankly hard in general when you’re entering things into a competition. Also, while the game is not very serious or personal, writing something so self-indulgent gives criticism the potential to cut surprisingly deep: if people don’t like it, then it’s not just that we’re bad writers and/or bad game designers, but also that we have terrible taste! This is why the first game was made for the more low-key Spring Thing, and why the series continues to be entered in the Thing despite the positive feedback we’ve gotten suggesting that it might do OK in IFComp. That said, this strategy had already resulted in our best-received game ever so we were determined to stick with it for round 2.
Planning and Project Management
Step 1 in making this game was to get a better setup going for cooperating on the game. For LTSS, EJ did most of the game’s writing and coding in the Twine editor, while Encorm wrote, edited, and debugged via a Google doc (with the text then pasted into Twine) or by just stealing the keyboard. Version control and passing files back and forth was done via Google Drive. This worked well enough for LTSS, in which Encorm, as a newbie Twine author, was fully involved in the planning but taking a more secondary role in the execution. But when working on Starbreakers (which was closer to a 50/50 effort, though with Encorm mostly writing and EJ mostly coding) it became clear that this wasn’t tenable and we needed a better solution for our future collaborations.
A quick search led us to Em Lazer-Walker’s article on using Tweego and Github to make collaborative Twine projects, and after a lot of fumbling (we’re not actually good at computers) we managed to get it up and running. Regular Twine doesn’t play nice with version control, so this was an improvement in pretty much every way, and getting to write code in an editor with syntax highlighting and spell-check plugins is actually very nice!
Step 2 was trying to figure out project management, both because this project was more ambitious than LTSS and required some actual planning, and because left to our own devices the two of us have preferences for pace and structure and rigor of scheduling that are very different from one another. We have to admit that we did not entirely figure this out this time around; we had hoped to avoid crunch time and have time to get playtesters, but we didn’t quite manage it. But we think (we hope?) that we learned a lot from the process.
In particular, we badly whiffed our estimate of how large the game would be and how long that would reasonably take to write. The structure of the Lady Thalia series is such that a player will only see about half the total text on any given playthrough, so we have to do a relatively large amount of writing per amount of visible game. LTSS clocked in at around 30k words (not counting code), so we estimated that LTRR would land around 40k and promptly forgot to do anything planning-wise with that knowledge. (And then we got carried away and wrote even more – scope creep is a killer). The game as it stands is somewhere north of 45k words (Twine’s built in word counter doesn’t count link text, which we have a lot of) and as such we barely got it in under the wire. That said, this should give us a good baseline to use when planning out the schedule for future games.
We had a lot of things we wanted to do with LTRR, either as improvements from the previous game or just things we wanted to try. Not all of them made it into the game but we’re going to dissect what did, and how we feel about how they turned out.
There’s three ways we reward players for succeeding in Lady Thalia: additional content, shiny points, or providing an advantage on a later puzzle. In the first game we occasionally accomplished the latter by allowing players to bypass certain puzzles entirely, which is something we were never entirely happy with. It amounts to “rewarding” via taking away parts of the game, which isn’t really a reward at all as long as people are enjoying the gameplay!
This time around we tried to design the various heists such that they’re still interesting if a player has done flawlessly up to that point, at the cost of upping the difficulty somewhat if they come into any particular puzzle unprepared. This is one of many places where it would have been very useful to get playtester feedback, since at the time of release we had no idea if we’d balanced things appropriately. Based on the reviews it looks like we did, but we can’t count on being that lucky every time!
EJ has, again, played too many visual novels so of course this time around we wanted to include actual relationship mechanics between Thalia and her rival. We decided that their relationship should revolve around some personal growth for Thalia, and as such should involve her learning to treat Mel nicely (but not too nicely) as well as learning that Actions Have Consequences, Sometimes With People You Like. We originally planned to have Thalia’s nighttime activities be able to affect Mel’s investigation the next day (the confrontation with the policeman on Night 1 is a remnant of this) but due to the aforementioned scheduling issues this ended up too complicated and had to be cut for time. This left the relationship mechanics confined to the interludes and left our original setup (picking the “Correct” dialog option out of three choices) too opaque when standing on its own. As @DeusIrae pointed out in his very excellent review, under these conditions it would have made more sense to copy the finesse mechanics here and give an increasing number of points for various types of answers with a penalty for picking too many of the same type. Rest assured, it’s on the list for the post-comp release.
Overall, we wanted to make sure the sequel was able to stand on its own for new players but was still fresh and interesting for people who liked LTSS. Writing wise this was its own challenge, of course, but from a gameplay standpoint we were actually very worried about the heist puzzles. Repeating puzzle structures from the previous game would run the risk of fatiguing returning players by making everything too similar, but trying to be too different would possibly just end up confusing people, and/or losing the magic that made the first game work. We settled on a combination of new puzzle ideas mixed with old puzzle structures heavily disguised by new writing. This seemed to work out pretty well (nobody complained and people seemed to like the new puzzles, in particular the gauntlet runs at the end of Nights 2 and 4). If (ok, when) we make Lady Thalia 3 we’ll still have to take this into consideration, but we’ve got a better feel for it now.
Under The Hood
In addition to setting up Tweego, we wanted to make a few other changes to the way the game worked to make things easier on us. Some of them worked better than others - for example, the daytime conversations in this game do not branch nearly as much as in the previous, which again nobody noticed. (Hooray!) On the other hand, we tried to revamp the finesse system to make point tracking easier by awarding points only for achieving the best outcome of a situation (which would lower the available total on the first night from 9 to 4, for example). Unfortunately this leaves the player in the position where (due to our no-failure ethos) they can successfully complete the heist but still get a fat 0 from Gwen. This, to put it lightly, sucks for the player. So we changed it back at the last minute and put a band-aid over it by having a maxFinesse variable increment every time we can award possible points. There must be a better solution, but neither of us is very good at coding so this worked well enough.
Some Fun Facts, and Other Closing Thoughts from Encorm
- Nobody mentioned the rickroll I put in Night 1, which is surprising. It is subtle but also very dumb. EJ talked me out of a lot of dumber jokes than that one, thank goodness, so I’m glad I stayed on the right side of the line.
- L’Oignon Pourpre was a placeholder name that made it in because no one thought of anything better, but it’s not actually as weird as Le Rat Mort (the real bar it was based on). There’s quite a lot of real historical detail thrown into this game, as 1910s Paris had a thriving lesbian subculture and was in general a very interesting place to be. The three women in the bar are also based on real queer women of the time (Collette, Vita Sackville-West, and Josephine Baker) with some liberty taken with how old they would have been and when they were in France.
- Writing in Commonwealth English as an American is hard, but not as hard as stopping after you’ve written ~20k words over a month that way. I think I seriously weirded out some coworkers for a bit before I managed to rein that in.
- I have been absolutely charmed by the reception this series has gotten, and continues to get. It’s very much not the standard thing that does well in the IF Scene, but the people who love it REALLY love it. I’m personally very happy to have contributed to nine people’s favorite IF series.
EJ will follow on with her closing thoughts when she has time. Thanks for reading this very large post-mortem, and thanks for everyone (both players and authors) who participated in this year’s Spring Thing. I had a blast because of all of you, and I hope to be back next year!