Replay and Multiple Endings for Mystery Games

[sake-bento]

Thank you guys all for the great feedback. If I might trouble you for a little more, it would be my honor.

A lot people have mentioned that there's not a huge replay value, which is a completely valid point, as there's really only one true ending. Are you guys okay with "good" endings where you don't solve the case? As in, would you mind an ending where you don't find the murderer, but instead wind up making sweet love to Chance and running away with her (just an example, don't get too excited)? Do you mind having a murder mystery that has more than one "true" end? Is it okay if you don't even solve the case in some of those "true" ends?

[Jake]

Leaving aside that I didn't find Chance too interesting in the first place, all of those things would be perfectly fine with me... :3

[Linky]

A lot people have mentioned that there's not a huge replay value, which is a completely valid point, as there's really only one true ending. Are you guys okay with "good" endings where you don't solve the case? As in, would you mind an ending where you don't find the murderer, but instead wind up making sweet love to Chance and running away with her (just an example, don't get too excited)? Do you mind having a murder mystery that has more than one "true" end? Is it okay if you don't even solve the case in some of those "true" ends?

Other good endings that don't solve the case can't possibly do anything but good! But to amplify they value-addingness, there should be a myriad of them - in order to really increase the replay value. This gives players a strong incentive to delve back into the story after playing it through the first timeand re-explore it for all the other possible endings (but this might require the endings being very different from one another...)
Yeah it might not be viable in creating a lot of other endings, but the more the better/deeper the story.

But, there should be only 1 'true' ending. It wraps up the story really nicely that way and is consistent as well.

I think it should also be explicity stated to players playing the game ahead of time that there are a number of possible different endings. This way, players will fully understand so.

[lepapillonrouge]

Maybe make two or three good endings, but only make one ending canon throughout the entire piece.

Dang. XD Someone said that.


...I just want to see more Mikolaj and Kizaki...if that's actually possible. (Manly man and megane -shot-)

[jack_norton]

A lot people have mentioned that there's not a huge replay value, which is a completely valid point, as there's really only one true ending. Are you guys okay with "good" endings where you don't solve the case? As in, would you mind an ending where you don't find the murderer, but instead wind up making sweet love to Chance and running away with her (just an example, don't get too excited)? Do you mind having a murder mystery that has more than one "true" end? Is it okay if you don't even solve the case in some of those "true" ends?

I don't remember the name (is an old game), but there was a detective game in which there were like 4-5 suspects, and at beginning of the game one would be randomly picked to be the "true" murderer. Of course the game was built such in a way (was an adventure) so that the leads/clues could change accordingly: it was fun because even if you knew more or less the "scheme" of the murder, you had to pay attention anyway since could have been another person.

Now, how to implement this with a VN... uhm I think would be rather impossible (or very hard).

About your question I like more endings, even those in which you don't catch the murderer, or maybe you do other stuff, like... sweet love with Chance? not a chance! I want Miss Bergstrom! from the feedback I got from Vera game, they liked the "good endings" in which you don't kill the werewolf, so I think I'm not the only one that wuld appreciate more endings

[Jake]

I don't remember the name (is an old game), but there was a detective game in which there were like 4-5 suspects, and at beginning of the game one would be randomly picked to be the "true" murderer. Of course the game was built such in a way (was an adventure) so that the leads/clues could change accordingly: it was fun because even if you knew more or less the "scheme" of the murder, you had to pay attention anyway since could have been another person.

Now, how to implement this with a VN... uhm I think would be rather impossible (or very hard).

I've come back to it again and again on this forum, but seriously: if you can, check out Westwood's Blade Runner, circa 1997; it's far from impossible:

http://en.wikipedia.org/wiki/Blade_Runn ... ideo_game)

Mechanically, it wasn't so distinguishable from a VN; most of the game revolved around a set of decisions as to where to go, who to talk to and what to say, with a few hidden-object-esque minigames and very occasional interludes of point-and-click-to-shoot gun action (which could probably also be implemented in something like Ren'Py pretty easily). The plot started off with the same scene every time, but could potentially have a completely different ending from one playthrough to the next (whoever wrote the plot summary on Wikipedia apparently only played it once...), with different characters randomly being on different sides, including the protagonist.

Now, obviously if you're going to tell two completely different stories in the same VN it's probably going to be almost as much work as just writing two separate linear VNs, but that's not something that's new for VN development anyway. I suspect that the work probably lies in structuring the stories such that any given major character can be an innocent or a criminal and not bend the path of the story too much, meaning that the player has to notice the small clues to pick up on who's actually guilty even if they've seen that story arc before. I guess some amount of random scene variation just to avoid the clues being the only bits that change also would be a good idea.

[Topagae]

It reminds me of a less lame Fahrenheit

I wonder if it'll run on a modern machine.

[jack_norton]

Ahhhhhhh yes!! haha I loved that game (well, Blade Runner is also one of my favourite movies of all times). Yes, I remember it came on 5 CD (a record for those times) and was the first game to use Voxel rendering for character sprites. I've played it soo many times to try to see all the different endings. That's a very good example of what I mean.

Though I reiterate what I said, that would be extremely hard to make it well. Already writing 2 endings for Vera game without major mistakes was hard! (or maybe it's just me that has difficulties writing detective stories)
<sorry for derailing the thread>

[Topagae]

When writing huge branches and intertwining storylines with lots of choice You REALLY need a chart.

Here's one of the many free flow-chart mapping programs out there. I use this one.

http://freemind.sourceforge.net/wiki/in ... /Main_Page

A fairly complex game like Mass Effect had 3 chalkboards of Flow charts for this kinda crap. It's seems to be the only way really. That or not even the 100 million dollar budget guys have found a better way yet >_>

[sake-bento]

I don't remember the name (is an old game), but there was a detective game in which there were like 4-5 suspects, and at beginning of the game one would be randomly picked to be the "true" murderer. Of course the game was built such in a way (was an adventure) so that the leads/clues could change accordingly: it was fun because even if you knew more or less the "scheme" of the murder, you had to pay attention anyway since could have been another person.

The plot started off with the same scene every time, but could potentially have a completely different ending from one playthrough to the next (whoever wrote the plot summary on Wikipedia apparently only played it once...), with different characters randomly being on different sides, including the protagonist.

Now, obviously if you're going to tell two completely different stories in the same VN it's probably going to be almost as much work as just writing two separate linear VNs, but that's not something that's new for VN development anyway. I suspect that the work probably lies in structuring the stories such that any given major character can be an innocent or a criminal and not bend the path of the story too much, meaning that the player has to notice the small clues to pick up on who's actually guilty even if they've seen that story arc before. I guess some amount of random scene variation just to avoid the clues being the only bits that change also would be a good idea.

I read a book that did that once. It was a popup book, and you would spin this wheel to change all the physical clues so that the culprit would be a different person. It was wild. I think the Clue movie did something similar, too.

Writing a game with a changing solution would actually be a fantastically fun challenge. I was tempted to do something like that with RE:A++ in which the true identity changed in different playthroughs. It didn't really seem practical in the realm of the storyline, though. A mystery game that's a bunch of different mysteries with a similar beginning sounds great, though.

[jack_norton]

Writing a game with a changing solution would actually be a fantastically fun challenge. I was tempted to do something like that with RE:A++ in which the true identity changed in different playthroughs. It didn't really seem practical in the realm of the storyline, though. A mystery game that's a bunch of different mysteries with a similar beginning sounds great, though.

We have different concept of "fun"
Jokes apart, surely is possible, but in practice you would have to split the story in several smaller narrative blocks that gets "activated" only on certain conditions. And I think that keeping the flow of the story natural wouldn't be easy (also you would have to check for consistency).
With bionic heart actually I did something like that:
- beginning is the same, then there are sub-paths that leads to the same "nodes", but tells the story from different angles
- then, depending on a important choice there were 2 main blocks: stay with tanya or helen
- of those two, they would split into several more directions. some "paths" would lead to same ending, despite they were originated from a different conditions

It was fun, but also really hard to write, and not sure was worth it in terms of gameplay (some people don't like to replay the game again to see different endings so you have to take into account that too).
Anyway whatever you choose to do I'm sure the result will be cool as always

[Topagae]

@Jack

Every branch requires you to double the amount of content you need to write, and check the consistency of that branch against other branches. Much like anything that branches there's also redundant stuff that needs to be consistent.

The complexity grows exponentially and becomes incredibly difficult. To solve this most stories like you mentioned branch, but reconvene in a set # of places to deal with all the huge amounts of complexity.

Example: You have three branches, it gets kinda muddled, but they all end up at the same place.

[jack_norton]

Example: You have three branches, it gets kinda muddled, but they all end up at the same place.

Yes that's what I did with bionic heart, some branches ended in a "endings" (game ending) while others reconvene in a specific place/node.
< sorry again to sake_bento for the thread derail lol >

[sake-bento]

Haha, we should probably split this off, because I really do like this discussion.

I imagine a mystery with multiple culprits would work more like Spirited Heart than Bionic Heart, though. There will be some common scenes, and at certain scenes (like the romance encounters in SH), the dialogue would be different based on the clues that need to be dropped.

[jack_norton]

Maybe some mod can cut those last posts into a new thread
A system like SH would probably work better indeed, but would still be a pain: in SH the various character you can romance never met each other (except for a short scene of Yuza and Janimee), so you don't have to check for consistency for that.
In a detective story usually the relationships between each character are essential for the story and for determining who is the murder and the motivations (y hates z, y loves x), so it would still be lot of work, unless the characters never met each other in a scene (maybe they're in custody or some other reason but should be plausible).