『g@me.』

"No, I... it's just, I got confused..."

"And this is news?"

"Huh?"

"Just come out with both guns blazing... like you always do"

"Phoenix Wright: Ace Attorney"

Because I won't make special posts for news or other news: English translations of Edogawa Rampo's The Strange Tale of Panorama Island (a crime story with a distinct fantastic element to it),  Case of the Murder at D-Slope (the first original Japanse locked room story) and The Fiend with 20 Faces (and Ashibe Taku's Murder in The Red Chamber) are set to be released by good old Kurodahan. Who are situated in Fukuoka. Yeah. And there is a library beneath the attic nowadays. And I still post at Criminal Element. But enough idle talk.

One of the weirdest essays I wrote for university was one on problems concerning reading comprehension within detective fiction. It was just for a Japanese language course and everyone was free to choose his own theme and while most people rewrote other essays/researches in Japanese, I for some reason chose to write a new essay on the topic. At first, I was actually planning to write something about how people read detective fiction and the implementation of that in videogames, but I really couldn't see me pull it off in Japanese. Or any language at all actually.

But I've wanting to write something on deduction systems in videogames for a long time, so why not now? Note, this is mainly on how deductions of the player are integrated into videogames, so I'll not go into topics like pixel-hunting and such. I'll only go into games that actually do expect some deducing by the player, even if it's just a little, so a game like Tantei Famicom Club is not discussed. And I hardly play PC games, so I won't discuss them. And finally, this is precisely not at all what I had in mind at first for this post, but I've been postponing it for months now, so I just came up with something that at least touches the subject. 'Cause that's how I roll.

Enough researches on reading comprehension in detective fiction indicate that readers of a (the same) detective story, often come up with a wide variety of deductions (see for example Kojima 1996, 1999 in the Attic). These differ in both width and depth; some ideas might have several layers of thought behind them, other just one. Some might be very original, some rather predictable.  And the same should happen with detective games, as long as we have (some kind of) puzzle plot. The problem is that while a book is pretty much a one-way trip that is pretty much automatic, games often give the player a certain range of freedom to move around. The same with detective games. Whereas the story in a book will always develop from point A to B to C, it's sometimes possible in games to move from point C to A to B. In fact, many gamers will find a game that's always lineair quite boring.

At certain points in the story (be it a book or a game), a check is made on the readers' deductions. In a novel it's often (though not exclusively) the ending, where the detective reveals his solution. The right solution. Compare it to your solution and you know whether you were succesful or not. In games however, this often results in a paradox. Game developer Takumi Shuu explains it in an article as follows:

"Mystery fiction and games.... at first sight, the two seem compatible, but they are actually very contradictive when you look at it from the position of a game creator. The theme of a detective is to 'unravel the mystery'. But on the other hand, you enjoy it the most when 'you're surprised when the mystery is unraveled at the end' ....  and this is actually the complete opposite. What you expect from a game is naturally 'to become a great detective and solve the mystery'. But if so, you neccesarily lose the enjoyment of being surprised at the end. A great detective can't explain his own deductions and get surprised by it... Just like that a magician can't get surprised by his own magic tricks. How can we overcome this contradiction? The key to mystery games lies precisely there...." (Takumi, 2010)

And indeed, most games don't cope with this paradox too well. Almost all games I know work with a model wherein the detective-character in the game lays out a deduction on his own and only allows input from the player at certain, select moments in the deduction. It's actually little more than a test of your deductions. As it would be quite impossible (@ the moment) to create a system that can react to every deduction possible made by man, they go the other extreme: they only check at certain points of the deduction to see if you have been paying attention. The problem with this model is that a lot of the deductions are done for you, you can only show your own ideas at certain select times, but the automatic part of the deduction often spoils a lot. Because the questions for the player have to be 'set up', the introducing sentences of the deduction often hint at a certain answer in advance. It's easier to look for the answer if you have a definite question. Say you think it was an alibi trick, but then you notice that all the questions are about strings and needles and bolts. It will probably influence your deduction.

In the Tantei Jinguuji Saburou (Nintendo Famicom/DS; Sony PSX/PS2/PSP) games, this is really reduced to the bare minimum, with many questions npothing more than mere tests of memory (questions like "who am I supposed to visit now?" are not rare).  Detective Conan - Rondo of the Blue Jewel (Nintendo DS) is also like this, with the character starting a deduction on his own and only checking with the player once in a while asking you to select the murderer, the murder weapon or something like that. A vaguer variant exists in Hayarigami (Sony PS2/PSP), a horror-adventure game that actually does allow the player some creative freedom. At certain points in the game, the player is asked to answer questions about the case. The player can answer with rational, realistic answers, or more supernatural answers. The story actually changes depending on the players' choices, resulting in either a rational explanation or a supernatural explanation for the stories. Here the questions are not so much testing (though they do test certain point that _have_ be filled in correctly), as much as giving the player the freedom to play around with the story.

Trick DS (Nintendo DS) works a bit different. The game basically also tests your deductions by comparing your ideas to crucial points in the automatic deduction, but the catch here is that you must select your answers beforehand. The player is forced to create in-game deductions. By combining cards (that represents things like possible weapons, crime scenes and other relevant information), deductions are made. The player has to pick out the deductions he thinks are right in advance before he can proceed into the automatic deduction sequence. The catch is that it's possible to make wrong deductions (thus possible to enter the deduction sequence with the wrong 'ammunition').

The system in Trick X Logic (Sony PSP) is probably the most like how people deduce while reading a real book. Here the player is also asked several (checking) questions, but just like with Trick DS, the player has to come up with possible answers on his own. This is done by selecting keywords from the text ("He can't read" and "He was seen reading a book"), in order to generate mysteries ("Why was the man reading a book if he can't read?"). This mysteries can be combined with other keywords to solve them, thus creating insights ("It was an imposter" or "He actually can read"). This can be done with contradicting pieces of text, but also with pieces of text confirming each other (creating 'facts'). Either way, by combining these texts, certain insights are made, which are used to answer the test questions.

In fact, the Trick X Logic system is IMHO the most natural, as this is exactly how people read detective fiction (at least people with some experience with puzzle plots). But even then it has to cope with that problem: you have the (leading) questions in advance, so you can work your way back to the answer. And that is something these games all have in common.

So what did Takumi Shuu (from the article) do? He turned it around in the Gyakuten Saiban / Ace Attorney (Nintendo GBA/DS/Wiiware) games. In those games, the player-character (and sometimes also the player) often has no idea what's going on. The player-character is just forced into situations wherein he is forced to strike back even though he has no deductions, no idea at all. He just looks for contradictions within a character's testimony and attacks him on those points, confident that by revealing one lie at the time, he will eventually arrive at the truth. Say a witness says he was watching TV while you have proof that there was an outage. You have a contradiction and you call your witness out on that, but you have no idea where this is going. In the beginning, the player-character (and the player) is just poking in the dark, hoping to find a weak spot.

And I really like this system for a game. The initial poking around is much more interactive than with other detective games, allowing the player to actually bumble around. The story often unfolds almost real-time, with the player-character getting new information every time he uncovers a lie and uses that to (immediately) alter his deductions. It's not like an exam like in the older deduction model. Which makes this system a lot more fun as a game.

Not to say that the old model is obsolete. The system used in Trick X Logic really holds a lot of potential, and if we would like... stick it into a multiple interactive storylines - structure like in 428 ~ Fuusa sareta Shibuya de (Wii/PS3/PSP), who knows what could happen? Deductions made by the player in one storyline that instantly change the development of the other (ongoing) storylines...

I really should finish some games one of these days. Like Glass Rose (PS2). Which is actually a Japanese adventure about a serial murder case in a 1920s mansion. I should like that game a lot more than I actually do....