Toby's Nose postmortem

Since the Author’s Note at the game’s end already explains the origin behind Toby’s Nose, I’m going to use this postmortem to talk more about the technical design.

The feedback I received wasn’t quite what I expected. Almost everyone mentioned the technical side, expressing surprise at how the game doesn’t alert you when various clues match together, and expressing more surprise that the game can be solved even without gathering all the evidence.

I say I didn’t expect to hear this response, and yet I thought a great deal about both these points when I made the game. I just never imagined they would draw so much attention.

This game has a single puzzle: to identify the murderer correctly. It isn’t like other puzzles in games that require objects to be moved around in relation to each other – for keys to be fetched for doors, for actions to be carried out in a specific order, for ingredients to be combined in recipes. It’s a question of knowledge. You start out not knowing the murderer and, by considering various things in the story, you should finish the game by realizing who’s guilty. The puzzle is something you have to work out in your own mind, because it’s only by working it out in your mind that you will gain the necessary understanding rather than having the game hand it to you.

Alternative designs existed. I could have had Toby make a statement when all the relevant clues were gathered. But then the game would have been announcing an impending solution, relieving the player from having to think for themselves. The puzzle would change from being mental to being mechanical, no different than gathering keys. And this wouldn’t help players who hadn’t found all the necessary clues anyway, since such players wouldn’t receive the announcement.

Which brings me to what constitutes a “necessary clue” in the first place. Everything in the game might be a clue, although most things are not; but on the other hand, entire situations sometimes function as clues. The most obvious example would be the brothel. Toby can smell perhaps twenty or thirty scents in this location, and although they don’t all factor directly into the murder, they do factor into establishing the circumstances behind the murder. In order for the player to understand that, the player has to sniff around and get an idea about what’s happening in this particular area. There would be no way for the game to ring a bell and say “Now you know what’s going on here!” because every player will come to that realization at a different speed.

That said, there are a few scents in the game that are extremely incriminating individually, and whenever Toby smells these, they automatically go into the inventory. This was a mechanic that I almost didn’t implement, because it means the player is relieved from having to deduce that these clues are important since the game announces it via the inventory. This went against my original philosophy, but I implemented it anyway after considering just how many scents there are. I figured the player would need help remembering everything.

Some players still wondered why the game didn’t announce when a link had been made between different clues. To explain my design choices here, I’ll use the blood as an example. It’s one of the first clues the player can smell, and the description you get is: “This blood is not fresh. It was spilled at least five days ago.”

Now, if you were to go around smelling different blood stains, you might assume they were from different people. But every blood stain you can smell in this game gives the same text, which indicates that you’re smelling the same blood every time. That’s the link connecting each blood-stained item, and the same logic applies consistently with every other clue. Discovering identical scents in numerous contexts is how you piece the puzzle together.

In addition to that, the game does at one point bluntly announce where the blood came from when the player smells a certain important object. So between these two methods of linking clues together, I’m satisfied that the game does enough without outright answering the mystery for the player.

The other point I wanted to address is how the game can be solved at the beginning with a random guess. I did consider gating the solution to prevent such a thing from happening, but I decided against it for a few reasons.

At the most basic level, it was a technical problem I didn’t have time to mess around with before the Spring Thing deadline. But that aside, I also asked myself: at what moment would the solution’s “gate” be opened? It would have to be when a certain number of clues are gathered, which brings me right back to my original idea behind making the game. I didn’t want an announcement via mechanics that “now you have the answer.” Not to mention that different players might find enough information to solve the mystery by gathering different clues. You don’t need to gather them all to realize who’s guilty. And anyway, on a narrative level, Toby wouldn’t necessarily know when he has all the clues. How could he know?

Well, I could’ve gated the solution until some arbitrary number of clues were gathered – say 50% – but what’s the point? Those 50% might be enough to deduce the murderer, or they might not, depending on the player’s own comprehension. Such a mechanic also might send the player the false message that “now you have the answer” even if the player doesn’t have the biggest clues yet. Therefore it had to be all or nothing, and I went with nothing because it supports the philosophy behind the game, and it’s also logical within the story. Holmes wouldn’t know a random bark from an informed bark. He’s depending on you, the player, to know what you’re doing. And his presumption about your role is an important theme in the story, which gets called out directly at the end.

After those considerations, my other option was to gate the solution until some arbitrary number of turns had passed in order to force the player to explore the game. But I didn’t do this. I wrote text where Holmes explicitly tells the player to take their time, that there’s no rush, and I added a warning after the first bark to once again explicitly tell the player to consider all the clues first. With those features in place, I decided that if a player wanted to brute-force the solution without exploring, I’d let them, because that meant this sort of game wasn’t for them.

Last but not least, I wanted the player to be able to guess incorrectly, without necessarily finding all the clues. Solving the game should require a leap of faith where, after deciding for yourself that you think you’ve gathered enough evidence, you jump. The risk is that you might fall down, but that’s okay. I wrote full-fledged endings for every incorrect guess to reward players even for making a mistake, and to hopefully give them the incentive to keep trying.

Overall, the downside to the game’s design is that a player without enough clues has to review the story again and look for more clues. But since other games force players to comb back through different rooms looking for missing items, or to repeat even more tedious actions as failure follows failure (Varicella and Spider and Web lost many points with me), I feel comfortable standing by the challenge level in Toby’s Nose. An easy “undo” will put you back into the story if you guess incorrectly. Everything is sniffable from one spot, and all the big clues are just a few sniffs away, so that you can always return to the drawing-room’s “home base” and try smelling something else. Not to mention that reviewing clues and looking for what you might’ve missed is what detectives do!

I wrote this game in two months on the heels of ParserComp, and lots of the feedback I got for Down, the Serpent and the Sun went into Toby’s Nose. I’m very grateful for all the ParserComp reviewers. Their advice aligned me more toward what games I want to create and what games I enjoy playing.

Down, the Serpent and the Sun had puzzles that existed just to be puzzles. That was a mistake. I’m not very invested in puzzles, even though I do enjoy them when they’ve got meat attached (my love for Hadean Lands is too great to fit inside a parenthetical). With Toby’s Nose, I wanted to remove every mechanical puzzle, streamline every verb so that there’s no guessing involved, make the goal and the game’s controls obvious at the outset, and allow the player to just explore and read and think. But I didn’t want to make a totally “puzzleless” game either, where the player merely floats forward; I did want the player to think, to engage the text, to turn it over and poke at it.

How much people enjoy Toby’s Nose will probably depend on how much investigating they’re willing to do. But this time I’m confident that I wrote something I’d like to play – especially since there’s lots to examine through smell! I’m always examining things in parser games.

Thanks again to everyone who’s given me feedback over the past months, to Aaron for hosting the Spring Thing festival, and to the other authors!

After playing this game, I dug out my long-hidden Sherlock Holmes book and decided to read it again. It has been 3 years since I touched that book. This shows how great Toby’s Nose is to revive my old love for the fictional detective.

Basically, this idea for this game is unique and is something fresh compared to the other generic stuff that we get on normal Twine or parser games. Reminds me of The Space Under the Window, where we find something significant in the text and type it out on the parser, hoping that is works.

The things that could have been better in this game is the signal that you are actually progressing. It’s like I gathered about 10 or 15 clues. Is it enough? Do I need to explore more to find the bad guy? Each clue led to every person in the room too, so it was not too easy to guess who did what. (Should the shovel be blamed on the gravedigger?) In the end, I gave up and did a trial-and-error spree on all the suspects. It’s funny that the one guilty happened to be the last one I chosen.

Overall, this is still one brilliant game which deserves first in Spring Thing. I hope to see more great games from you, Chandler!

Thank you! I’m glad to hear that this game inspired you to dig out your old Holmes books. I only wish that Toby himself were featured in more of Doyle’s stories!

At first the clues may seem to point to a few suspects, but once they’ve all been pieced together there’s only one suspect who fits. I should’ve mentioned in my postmortem that this game now has a hint menu complete with a walkthrough, though, for anyone who gets stuck. I always wanted to have that included in the game, but for some reason it kept crashing originally when I was programming the table of hints. Midway through the festival I was able to figure out how to reprogram the table without crashing the game (even though I don’t know what I did differently) and I updated the download file, so the version available now should include those new hints. But some people may still have the old version from when Spring Thing first started, and if you do, then please discard that file and download it again! The updated version has some disambiguation errors fixed too.