A two-dimensional reputation system

[TrashMan]

I like the OP's suggestion.

 

 

Having just a like/dislike system is not very deep. Just because I dislike you, doesn't mean I'll attack you on sight. A single variable is not really enough

[JFSOCC]

They could do numerics for "Known for" which get turned on based on which quests you completed and how. So if you're always discrete, you wouldn't be known for it, except to those who know these sorts of things

 

Then you have certain NPC's who have dialogue trigger based on a small list of things you've done which they could know.

 

so, a Crime lord NPC would check

Thieves guild reputation 45

Thieves plot quest done loudly 0

Thieves plot quest done discretely 1

Police plot quest done loudly 0

Police plot quest done discretely 1

Major game event one done 1

Major game event two done 0

 

etc.

So he could comment on your discretion, and wouldn't know you've also worked for those coppers, known you've done something of significance in the game world. and that you've helped them out for 45 points worth.

 

Then you could have certain dialogue options trigger based on these prerequisites.

Maybe I'm overthinking it.

[Ffordesoon]

@Frenetic Pony:

 

Oh, it was a really enjoyable game, no doubt. But you can't deny that it's kind of a mess in a lot of ways - most particularly the non-sandbox-y bits, although those did have a few neat ideas that weren't implemented all that well.

 

 

For example, the idea that you would lose experience when you didn't do what Lucien wanted in the Spire was a great idea, but you lost so little experience that it didn't actually matter.

 

Or the part where Reaver (what a wasted character, by the way - especially when it was Stephen f**king Fry doing the voice!) tricks you into having your youth taken away, but it doesn't affect your attractiveness very much at all if you're wearing the right stuff.

 

Or the part where Lucien kills your wife and kids offscreen, but they were essentially ciphers who served no useful purpose, so you (or at least I) weren't very invested in them anyway.

 

Or the limp final choice, where one of the choices was simply made invalid by another design decision (the "wealth" one - it was trivially easy to get all the cash you might ever need by just buying up property, so there was no great need for more dough). The other two weren't great either, but they did at least represent choices that might be vaguely meaningful to the player.

 

 

Basically, the issue with Fable II is that it's unfocused. Really tremendously enjoyable, but unfocused. You know?

[Lephys]

They could do numerics for "Known for" which get turned on based on which quests you completed and how. So if you're always discrete, you wouldn't be known for it, except to those who know these sorts of things

 

Then you have certain NPC's who have dialogue trigger based on a small list of things you've done which they could know.

 

so, a Crime lord NPC would check

Thieves guild reputation 45

Thieves plot quest done loudly 0

Thieves plot quest done discretely 1

Police plot quest done loudly 0

Police plot quest done discretely 1

Major game event one done 1

Major game event two done 0

 

etc.

So he could comment on your discretion, and wouldn't know you've also worked for those coppers, known you've done something of significance in the game world. and that you've helped them out for 45 points worth.

 

Then you could have certain dialogue options trigger based on these prerequisites.

Maybe I'm overthinking it.

 

I definitely think that the "known for" aspect should be represented in the reputation system, in whatever precise manner. 8P

 

Of course, that would still be a separate factor as it relates to your "How this person or this group regards you at the moment." For example... The Thieves' Guild might regard you with respect if THEY know of things you've acquired without anyone else knowing about them, but might react with disappointment or disinterest if everyone knows about the things you've done ("everyone" being "common people who aren't experts in stealth and therefore shouldn't be aware of your stealthy acts"). OR, The Guild might respect you for being known (by the right people) for stealthily accomplishing things, while a specific member (or just some other individual) might only have respect or admiration for you if NO ONE knows how you acquire the things you acquire or handle the situations you handle because they're so stealthily handled. Or, another individual might even admire you for boasting about your skills/accomplishments, even though the Thieves' Guild might frown upon this (but still appreciate your skill).

 

Things just tend to be more... robust, when there's not just a single "good points -- bad points" scale for everyone's regard for you, even when a variety of different things affect that scale in different ways. You only get 3 different ways to affect it: A positive change, no change, or a negative change. It's just very limiting, especially when dealing with something as extensive as human(oid) reaction/regard.

 

It's the same as how Water or Air spells do different things from fire spells, not absolutely better or worse things than fire spells. Other factor values are needed to determine the inherent value, so there isn't really an up or a down, but, instead, a variety of directions.

Edited April 19, 2013 by Lephys

[TrashMan]

What kind of scales are we actually looking at here? Universal scales or faction/individual ones?

What are the things you'd expect people would keepin mind when dealing with you?

 

Variables tied to specific NPCs:

1. Respect/like - certain actions that the NPC's knows you did may affect ther level of respect (in Leyphises example, sneaking or doing a certain quest may affect this differenlty for different NPCS)

2. Fear - NPC may think you are a wimp, might be unsure, or fear you

 

Global/facton-specific variables:

3. Reputation? - this is basicly rumors. What people say you did.

[Adhin]

Yeah I think to do the thing any real justice you'd need it both on an individual basis and a faction basis and anyone who's need some random guard or unnamed of a faction would need both. A, this faction doesn't like you, or doesn't care about you but this guy in it is still going to help you out because of things you've done for him in the past kinda thing. That's usually the kinda they they only ever do via the story and that person is just kinda 'out of' the faction loop till there part in the scripted story stuffs over with.

 

But no matter how many points or categories they could throw you in depending on any mixture of stuff it'll have to be represented in dialog of you just have A LOT more data that all boils down to the same 3 things, hated, neutral or liked. Feel like a broken record at this point, but people keep just bringing up more data points which games, including new Vegas already had. They just failed to represent that outside of a little title next to the faction name.

 

****s gadda be represented in a more granular manner or its pointless.

[Lephys]

But no matter how many points or categories they could throw you in depending on any mixture of stuff it'll have to be represented in dialog of you just have A LOT more data that all boils down to the same 3 things, hated, neutral or liked. Feel like a broken record at this point, but people keep just bringing up more data points which games, including new Vegas already had. They just failed to represent that outside of a little title next to the faction name.

 

****s gadda be represented in a more granular manner or its pointless.

Well, the point (or at least mine... I mean, what I'm thinking on the matter) is that you have not just more numerous data points on the same +/- scale, but actually multiple scales. The tricky part is hammering those out the right way; picking exactly what they'll be.

 

Really, these things are represented in plenty of games. Just... very simplistically. You've got your like/hate scale, and then you've got all these other things handled not by a scale, but by individual skill/dialogue checks and/or quest triggers.

 

For example, you helped that orphanage? *Trigger flip* NPC Bob now trusts/respects you with his firstborn child. There is no scale for "Okay, you seem to be looking out for those in need, but how do I know you're not just trying to get people to like you? I mean, how well do I really know you?" that's affected by different amounts or types of deeds, there. Then, you've got the "Do this thing for me, and we'll be best friends for life" quest. Again, a binary switch.

 

You don't allow for "Okay, I'm afraid of what you might be capable of, because I heard this rumor about you, but, as far as I'm concerned, you're still good people, because of what else I know about you."

 

Another thing to consider is this: What if certain factors aren't always created by your direct relationship with a given NPC, but instead are basically statuses (statii?) of NPCs? As in, a person can be highly frustrated with a situation (and frustration can even be compounded by your dialogue choices/actions) without them actually developing a permanent dislike for you based on that. But, a super frustrated person isn't going to react to things like a calm person would. A frustrated person might say "I DON'T HAVE TIME FOR THAT!", whilst a calm person might say "Okay, I'll tell you what I know real quick," or even agree to deals that they otherwise wouldn't have, etc.

 

Similarly, someone can be frightened or panicked without being directly frightened of you, in general. And, things you do or say might cause them to become quite frightened or panicked without permanently attributing fright and panic to your presence, but they'll still react differently.

 

*shrug*

 

It still feels very difficult to really come up with some kind of rough draft of the skeleton of a whole, working system like this. But, I think it's worth exploring.

 

Basically, there need to be more varied effects on the status and reactivity of NPCs (and groups of NPCs) -- for the purposes of creating more dynamic situations and outcomes -- rather than simply a greater number of effects on the same narrow range of effects (like/hate/neutral) on NPC demeanor.

 

What we've got is kind of like a weapon system in which you sharpen/improve your sword enough and it becomes an axe. Maybe you sharpen it more and it becomes a polearm. Each weapon has completely separate sets of properties, so it shouldn't be limited to a certain quality range on a single scale. An axe should have its own scale, a sword its own, and a polearm its own.

[Amberion]

How about this? You and your party have certain values for reputation. These values are myriad, and can cover everything from 'tendency toward selfless acts' to 'reputation for wanton slaughter of civilians' and all sorts of different 'things' that you can do in the game that would give you a higher or lower reputation in these things.

 

Different factions have different likes and dislikes for different reputations. If you gain a reputation for killing thieves, for instance, the thieves guild would probably have you on kill on sight. If you get a reputation as a crack thief, the thieves guild might want to have a little 'chat' about working on their turf. If you gain a reputation for selfless acts, the local church might ask you to help them with a charitable donation, or someone might ask for help, only to lure you into a trap.

 

As an example: Killing a thief would net you +1 thief killer reputation. Killing a guard would net you +1 anti-authoritarian reputation. The higher those values go, the more they affect the reputations that have those values in importance. Some people won't care one way or another whether you kill lots of guards, or kill thieves. Some will like both, some will hate both. some will like one or the other, and so on.

[Lephys]

^ I think it might need more beyond that (like the individual NPC reactive statuses, for example), mayyyyybe... But it's definitely a good system facet, even on its own.

[JFSOCC]

that might work if everyone checked against several different reputation markers. So say, eacg faction checks against 5 out of 20 markers, and you need 3 to be positive and no more than 1 negative to get certain dialogue. At this point you have a lot of freedom in creating reactivity.

[Lephys]

^ Yeah, and you could probably come up with some reasonable mathematical threshold for determining when someone's personal experience with your character(s) overrides their faction's experience/view of your character(s).

 

Simplistically, something like "If Fear is 70, and Faction's Unwillingness To Cooperate With You is 50, NPC A's Fear causes him to obey you for the time being."

 

So, basically, an NPC without a faction would have only their personal factor values to check. Whereas, an NPC with a faction would simply have an extra check on faction modifiers (which could be different for different NPCs -- maybe a new recruit in the Followers of Ciethe is less powerfully swayed by their creed than an elder member) to measure the difference between the two and see if it hits the threshold or not for that specific effect.

[Randomthom]

I like some of the ideas being spouted here & far be it from me to crush them but they may need to be tempered with some elements of realism.

 

Consider;

9 potential different reactions to the player/group means writing 9 different conversation options for many conversations. This also increases the likelihood of quest bugs etc. which eats up more development time.

 

One thing I do think is important and should be included is the concept of fame/notoriety/visibility. Why do you gain a +/- to a reputation score if nobody saw you do what you did to earn it?

Killing someone in the centre of town for all to see (righteous or not) will definitely earn you some respect/fear/badass points. Killing them in a back street where nobody saw you do it shouldn't earn you anything (unless you take their head off & wear it as a trophy!)

 

I'd love to see the game have different options at certain points depending on your fame/infamy/obscurity. It might be that you can't do a certain quest because you'll be recognised (or at least you'll need to do an extra quest first to secure a good disguise). Equally another quest might rely on your fame to get you into a high-to-do party to dig up dirt on a questionable noble.

 

It would be fun to play a character who kills people indiscriminately but secretively who also atones by anonymous giving to a temple of a good God, all-the-time remaining a relative unknown to most people.

[Kjaamor]

Consider;

9 potential different reactions to the player/group means writing 9 different conversation options for many conversations. This also increases the likelihood of quest bugs etc. which eats up more development time.

 

Indeed, and this is my main concern with all these technically good ideas. They're fantastic in the world of insta-idea-to-code but aren't terribly realistic.

 

Reputation systems are only ever as good as their impact upon gameplay. You can have the most superb system in the world, but if you don't implement it then it really doesn't matter.

 

My understanding (perhaps 'assumption' would be a better term) of P:E is that it will use 'simple' +/-per faction reputation. Even in using this, extreme reputation differences will probably affect a handful of quests and minor reputation differences are unlikely to affect anything. As much as I love complicated systems, I can't see how increasing the axes of reputation will affect the gameplay to a degree worthy of its implementation.

 

I know the good/evil axis is somewhat childish but it is very easy to implement. In terms of practicalities, New Vegas (as an example of heightened methods mentioned here) only used good/evil for gameplay, and the other differences purely affected the arbitrary titles given on your character sheet. Which is fine, but that's really what you have to expect.

[Amentep]

I like some of the ideas being spouted here & far be it from me to crush them but they may need to be tempered with some elements of realism.

 

Consider;

9 potential different reactions to the player/group means writing 9 different conversation options for many conversations. This also increases the likelihood of quest bugs etc. which eats up more development time.

 

This depends, I think.  Is having "9" values much different from checking 7 values - Quest, faction, reputation, Int, Wis, Chr, Speech?  Realistically you're always going to use a "general impression" and then NPC specific reactions based on other elements in the narrative.  I suppose really the question is how robust is the general impression?

 

Lets assume that the 9 listed values from the first post are actually considered clustered reactionary states of NPCs

 

And lets cluster them as such

Sympathy / Indifference / Disgust = for charity

Affection / Neutrality / Hatred = for fellowship

Admiration / Respect / Fear = for following

 

 

Now each individual could have a personal and faction based reaction so

 

X_n = personal charity

A = faction charity

Y_n= personal fellowship

B= faction fellowship

Z_n = personal following

C = faction following

 

and where n = specific NPC

 

Also, characters may strongly value some personal reactions over others (a woodsman may never follow another man but may value fellowship with others for example) so lets tag α, β, δ as how an npc values charity, fellowship and following

 

So then reaction becomes

Charity = αX_n + A

Fellowship = βY_n + B

Following = δZ_n + C

 

Now some quests may create some things may create a situation where a character will never like - or always like the PC and we'll give that a value of ♣_n and K will represent the total "quest reaction" where K_n = sum of modifying value specific quests might give a PC.

 

So we get general impression := ♣_n((αX_n + A) + (βY_n + B) + (δZ_n + C) + K_n) where -100 =

 

Each Faction Reaction can be A + B + C for each faction

 

Town Reaction can be  Σ^T  ∞ _T1 X_n + Y_n Z_n + F where T is the townspeople at that local and F is the reaction of the dominant faction of the town.

 

Overall Reputation can be plotted as the the sum of faction and town reactions in the world.

 

Or something. I mostly just wanted to use sub- and superscript and pseudo-mathematical notation this morning.

[Lephys]

Consider;

9 potential different reactions to the player/group means writing 9 different conversation options for many conversations.

Not necessarily. In the OP, you've only actually got 4 scales. 3 horizontal scales, plus an overlapping vertical one. You're not going to have a response that makes an NPC like you, one that demands respect, one that produces fear, one that produces hate, one that produces... etc.

 

The point of these factors is that they can be altered and reached by what you do. Not to give you the dialogue tools necessary to directly craft the exact effect you want in every single person you talk to.

 

Some people would be immune to one or more scales. And, even with the ones who weren't, it wouldn't make sense for you to have the means to affect them on every possible level, all at your fingertips. You can't go around saving orphanages for 70% of the game, then get to a person and suddenly be ferociously intimidating. You've got no practice, experience, or reputation to back you up.

 

There would be VERY few dialogues with such a plethora of options.

[Randomthom]

Granted, it might always be 9 but my point is still true, that it means a lot of extra work when writing & coding the dialogue.

[Lephys]

^ Point acknowledged. Don't forget, however, that, for a lot of this stuff, you've already got the dialogue options and check systems set up in oodles of previous games. They simply fail to represent more than a single scale of final effects.

 

So, in a lot of instances, instead of having 4 dialogue options grant varying amounts of negative points, and 4 grant varying amounts of positive points (to the same scale), you can simply work in the extra scales/demeanors and have each of the negative choices affect a different scale, and each of the positive choices affect a different scale.

 

This isn't so much a purely new giant block of content to add to the game. It's simply a new way of handling lots of already-existent cRPG content. Granted, I won't argue that it probably wouldn't be extra work as compared to NOT-doing it. But, almost any complex system/addition is, and it wouldn't really be as much as it seems (a default game doesn't have 1-2 choices per dialogue, and now we're trying to bring that up to 9 with our 9 demeanor factors, to put it simply).

[Ffordesoon]

Also, full voiceover work severely limits a developer's ability to deliver on choices. Text doesn't.

[Kjaamor]

Also, full voiceover work severely limits a developer's ability to deliver on choices. Text doesn't.

 

Voiceover work offers more limits than text, but that is not to say that text itself does not have practical limits. The writing and the coding around the writing take time, and like so many other features it's a case of assessing whether the end product justifies the time put in.

 

I don't think anyone is arguing that the implementation of further reputation/alignment systems is a bad mechanic, it's just a case of weighing up the effort it would take and whether or not time would be better spent on that than, say, level design, combat mechanics or character development.

 

In terms of expansive, checked dialogue, Fallout 2 is still the king of all the games I've played, and it's the reason why I love it so much. Yet the IE fantasy games (PS:T excepted) got away with next to no checks and it didn't stop them from being some of the best RPGs ever made.

[Solstis]

I just have to say after reading the original post and a few comments, the OP's idea is very similar to one I've considered. I'm sure it won't occur to everyone but a system similar to the one he has attempted to describe is undoubtedly a stepping stone to the future of RPGs and will be included in a groundbreaking game. I really hope it's this one. Such a system is the basis for a more in depth character experience and a more realistic game world. This really has my support and should be on the list of potential additions.

[Arsene Lupin]

I kind of think the systems to emulate are (or, rather, will be) Divinity: Original Sin, where there are a whole host of variables. Pessimism and Faith, Affection and Affinity, etc. They've got variables for -everything-.

[Hormalakh]

Hi, good topic. I really like the idea especially having two different "aspects" quantified (emotional vs. practical).

 

Three things I'd like to add.

 

1- Each individual NPC should have their own "calculated rep" for you. They would derive their calculated reputation based on several variables: your actions against/for that individual or conversation choices made, the reputation of the faction of which they're part of, any extenuating circumstances modifiers, specific character skills/bonuses hat affect reputation. You can have each of these variables have different weights for each individual NPC. Some people (the sheep) always go with the mob mentality (0.7*faction reputation+ 0.3*everything_else) whereas others care more about what you do to them individually (actions/speech*0.6+faction*0.1+etc etc). 

 

Then of course, the faction reputation can change based on a calculation done on certain influential people in that particular community. Perhaps through conversation we can figure out who these individuals are who modify reputations ("Oh he's the king, we love our king and listen to him") as well as figure out who has an individual mind vs hivemind ("oh jack? He's always striking out on his own.")

 

2- If we were to draw a graph of reputation on the y-axis and actions done by player on the x-axis, the reputation line should NOT be linear but rather "S-shaped" in that there are two asymptotic lines at "100" and "0". Thus as you are liked/disliked more and more by factions/individuals, it requires more and more significant actions to make them "LOVE" or "HATE" you. It should be fairly easy to be neutral within a faction and doing one good and one bad might cross each other out, but you need to make a concerted effort to have a whole faction love you. Usually the repuation systems with a linear line become sort of silly because you keep repeating the same conversation line to them and after 5 times of repeating the same thing or giving them some money, they suddenly love you. which brings me to my third point...

 

3- Do not allow repetitive actions to alter the reputations so much. This was done terribly poorly in Arcanum for eample, where if I complimented an individual three times, suddenly they loved me. I always thought that was weird.

[Tamerlane]

It should be fairly easy to be neutral within a faction and doing one good and one bad might cross each other out, but you need to make a concerted effort to have a whole faction love you.

This right here is an important thing. It is a thing that I hate, and a big part of the reason I brought up New Vegas on the first page.

 

The idea that opposed actions action can bring you back to neutrality is a weakness in a lot of reputation/morality systems. Saving an orphanage then burning a box of puppies does not make you "neutral", it makes you a giant **** who has done some good. Helping faction A then hurting faction A doesn't make them ambivalent towards you, it makes them wonder what the **** your deal is.

 

New Vegas got that bit right. You save an NCR train from getting blown up then murder one of their patrols, and they'll consider you a wildcard, not neutral. Even Mass Effect, for all that its paragon/renegade stuff was horrible - spoiler alert: pretty damn horrible - completely separated paragon and renegade points and only allowed values to increase (though there was a lot of dumb hidden stuff in ME2 because, again, the paragon/renegade system was pretty damn horrible).

[Hormalakh]

 

It should be fairly easy to be neutral within a faction and doing one good and one bad might cross each other out, but you need to make a concerted effort to have a whole faction love you.

This right here is an important thing. It is a thing that I hate, and a big part of the reason I brought up New Vegas on the first page.

 

The idea that opposed actions action can bring you back to neutrality is a weakness in a lot of reputation/morality systems. Saving an orphanage then burning a box of puppies does not make you "neutral", it makes you a giant **** who has done some good. Helping faction A then hurting faction A doesn't make them ambivalent towards you, it makes them wonder what the **** your deal is.

 

New Vegas got that bit right. You save an NCR train from getting blown up then murder one of their patrols, and they'll consider you a wildcard, not neutral. Even Mass Effect, for all that its paragon/renegade stuff was horrible - spoiler alert: pretty damn horrible - completely separated paragon and renegade points and only allowed values to increase (though there was a lot of dumb hidden stuff in ME2 because, again, the paragon/renegade system was pretty damn horrible).

 

 

yep. you're right - i agree, that's a better way to do it. the point still stands though. it should be a nonlinear curve towards good or evil. 

 

perhaps you start neutral, but then doing good and doing bad can shift you between hated -> wildcard -> loved: all the different spectrums or whatever else.

[AGX-17]

Also, full voiceover work severely limits a developer's ability to deliver on choices. Text doesn't.

Strictly speaking, it doesn't. It's budget that matters in that case. If they had enough money, they could voice every line. What has been mentioned by Obsidian is the issue of going back and revising content, which again is a budgetary, (as well as logistical,) issue related specifically to the cost of hiring actors and the potential issue of availability if they wanted to recod new lines.

Edited May 6, 2013 by AGX-17