The New World design topic #3: Dialogue Checks

[Vault Dweller]

When designing AoD dialogue system, our goal was simple: your character’s skills must determine conversations' outcomes (i.e. success or failure). The dialogue checks were equally simple: if your skill is high enough, you pass the check, otherwise you fail. It created 3 problems:

1. You never had to consider what the NPC would respond best to. Any tagged line would result in instant success if you have the skill, meaning that your dialogue option was not an attempt (as it should be) but guaranteed success, which made considering the options redundant.

2. Since all dialogues had multiple checks to simulate realistic conversations, it didn’t matter how many checks you passed and how well you were doing until that last check that resulted in failure (i.e. early success didn’t contribute to anything and thus didn’t matter).

3. The rigid nature of the system forced us to lower the checks to make the hybrids (i.e. jacks of all trades with lower skills) viable, which in turn made playing talkers an easy mode.

We did have a couple of interesting dialogues. When you talk to Lorenza, she asks you some questions to understand your motivations better (before she makes her decision), and your answers modify the checks later on, making them easier or harder.

In The New World we’d like to engage the player, make him/her consider the options instead of clicking on the line with the tag matching your highest skill, yet still keep the system skill-driven. It’s not an easy task as this problem doesn’t have a perfect solution, so I’m asking you to consider both systems (see below) and vote for the one where the pros outweigh the cons.

Option 1 (AoD):

NPC text

1. [skill 1] Dialogue option A
2. [skill 2] Dialogue option B
a.[success] NPC favorable response
b.[failure] NPC negative response

Option 2 (the proposed system):

The biggest conceptual change is that the tagged lines would now represent an attempt without any guarantees of success. It’s up to the player to read people based on the available info and consider what would work best. You can have two different streetwise lines, for example, one would result in a positive reaction, the other in a negative.

That brings us to the second biggest change. Most lines would no longer lead to success or failures but result in positive and negative reactions, represented numerically. Your skill level would act as modifiers, magnifying positive reactions and reducing the effect of blunders. The final check would tally up the reactions, which will determine whether you’ve succeeded or failed.

Let’s say your Persuasion is 3. You’re offered 3 arguments. The NPC will respond very favorably (+2) to argument #1, favorably to argument #2 (+1) and very negatively to argument #3 (-2). Your skill will modify these reactions to 4, 2, and 0. Let’s say the final check’s value would be 10, so assuming the conversation has 3 nodes with tagged lines, you’ll need to score at least 2 very favorable reactions and 1 favorable (or 3 very favorable ones) to pass the check. In longer dialogues you’d be able to fail a few times and still recover.

This system will maintain the importance of skills and encourage further investment but it will shift the focus to figuring out which lines would work best. Obviously, it might increase meta-gaming but that’s your choice and thus not our concern. Every time the player is offered to make a choice with different outcomes, 8 out of 10 people would want to know the outcomes in advance and the exact way to get to the outcome they want.

NPC text:

1. [skill 1] Dialogue option A
a. NPC very favorable reaction: +2
2. [skill 1] Dialogue option B
a. NPC negative reaction: -1
3. [skill 2] Dialogue option C
a. NPC favorable reaction: +1

Anyway, let us know what you think and if you have any concerns.

PS. In unrelated news Steam introduced developer pages, which are handy for developers and publishers with 200 games and not so handy for developers with 2 games. Still, we'd appreciate if you 'follow' us in case there's more to this whole thing.

https://store.steampowered.com/developer/irontowerstudio/

 

[Binky]

The new system. Definitely.

I'd also suggest more in the way of negative consequences for being a lying, treacherous weasel. People don't take very kindly to being lied to, betrayed, and manipulated. Thus, you fuck over too many people (or just one wrong person) and they come after you. Hard.

 

[Tigranes]

#2 is obviously far superior, but the question is whether you can actually pull it off. It's exponentially more complicated to plan out than even AOD dialogues, which were already pretty complicated.

I remember playing AOD for the first time and commenting that Lorenza is one of the best dialogues for this reason. I would be pretty happy if, even within the confines of #1, we were able to have Lorenza-like dialogues with key NPCs - where, at least, you know you're good at streetwise but you have 3 different streetwise options and you have to have read & understood the NPC's motivations to figure out which one.

Obviously dumbfucks will reloadfest until they get the "best outcome" like some hentai game, but who cares about that

 

[Vault Dweller]

#2 is obviously far superior, but the question is whether you can actually pull it off. It's exponentially more complicated to plan out than even AOD dialogues, which were already pretty complicated.

It's definitely doable.

I remember playing AOD for the first time and commenting that Lorenza is one of the best dialogues for this reason. I would be pretty happy if, even within the confines of #1, we were able to have Lorenza-like dialogues with key NPCs...

You will. This much I can guarantee you.

Obviously dumbfucks will reloadfest until they get the "best outcome" like some hentai game, but who cares about that

My sentiment exactly.

I did discuss it with another developer and his concern was:

"The player has to interpret what those arguments actually are, evaluate which is meant to be the stronger one, and then offer that back up. If you make it too obvious which is stronger and which is weaker, you have largely defeated your own goal. But if you make it too obscure, there is a significant risk that the player will believe he has picked the more persuasive option. He believes he has expressed, "Persuade him effectively." You interpret his choice as, "Persuade him ineffectively." And the player gets pissed off."

 

[Tigranes]

Well, that would be TNW's version of the popular AOD complaint that "I thought in X situation Y response is an obvious one, but the available options did not support it." It is an inherent limitation of premade dialogues. Breaking it down to a "gather points" system would mitigate this, however, since you're not failing outright but losing a point or two.

This might be an area where a degree of beta testing could happen for key pieces of dialogue. One of the things that is most difficult for writers to anticipate is what pieces of information are picked up and not picked up by readers/players/etc. - so at the least you can test whether players actually come out of the initial Antidas conversation with a decent grasp of what makes the guy tick, etc.

eAnother challenge seems to be how to communicate the 'points progress' mid-dialogue in a natural way. "Antidas nods along with your argument, though he does not seem fully persuaded yet" is simple but crude, but integrating that info to the actual flow of dialogue risks making the dialogue itself more mechanical, because of this need to show the player where they are at.

 

[lukaszek]

deterministic system > RNG

 

[Vault Dweller]

Your skill level increases positive reaction and reduce negative reaction, which is pretty much how it works in real life. Give 10 sales reps the same script (i.e. the same lines to say) and the results will be very different because the skillsets are different. What you say matters, of course, but how you say it matters more.

 

[Black Angel]

I think it's possible to pull the new system off. Aside from AoD's own achievement with Lorenza, Arcanum did this with Kerghan. In dialogue with Kerghan, the choice nodes are the exact same, but (iirc) you need full rank of Persuasion (and according to the wiki, probably 20 CHA). Even then, you have to navigate through Kerghan's dialogue options carefully; you need to choose 'the right option' every time, and only then he will agree to stand down. Otherwise, he would get pissed off, stops you from saying anything else and force you to choose a side.

The problem with Kerghan, however, is that it wasn't exactly clear if I chose the right option, because that kind of engagement only appear at that very moment, and I can't remember if the dialogue with any other NPCs before that happens the same way.

So, yeah, I think TNW's proposed system of positive/negative reaction system might fix Arcanum's Kerghan problem. The question is, would there be some degree of outcome according to how many points you achieved at the end of a conversation in TNW?

 

[HeatEXTEND]

The proposed system sounds exactly like my utopia of a dialog system, I'd never dream someone would actually dare to consider implementing something like it. ITS geared to barrel over the battlements once again huh? Godspeed.

Spoiler

 

[Daedalos]

New system by far. The new system sounds kickass and also fair and fun

 

[MF]

Cool. Sounds a little like my game's disposition & sway system, except it's still tree-based, right?

Will the reaction modifiers carry over in later conversations with the same character? That's what I'm doing, but it's very tricky to write for and I wouldn't recommend that if you have more than 20 NPCs.

 

[Jaedar]

New system is better, but I would suggest a trait/talent or something that coupled with your skills give you mechanical feedback about your choices (like CASIE aug in the new deus ex's). Or maybe just a skill that's entirely about reading people. At max level you might even let it be as clear as that talent in the first fallout which colors all dialogue blue/yellow/red.

Forcing the player to think a bit about which dialogue to pick is great, but just like how you can completely steamroll combat with a properly built combat char with little strategy, I think it should be possible to steamroll dialogues with a properly specialized character.

 

[hivemind]

New system is better, but I would suggest a trait/talent or something that coupled with your skills give you mechanical feedback about your choices (like CASIE aug in the new deus ex's). Or maybe just a skill that's entirely about reading people. At max level you might even let it be as clear as that talent in the first fallout which colors all dialogue blue/yellow/red.

Forcing the player to think a bit about which dialogue to pick is great, but just like how you can completely steamroll combat with a properly built combat char with little strategy, I think it should be possible to steamroll dialogues with a properly specialized character.

"no"

 

[Zanzoken]

Good on you for recognizing the weaknesses in the AoD approach. Not that I didn't enjoy the merchant or loremaster runs, but there was no challenge to it.

A good RPG system needs to incorporate both player and character skill.

 

[agentorange]

If there's going to be more likelihood to "fail" a dialogue situation, then I think there should also be more variety in the outcome of these conversations. That is, situations where a negative outcome can have a somewhat beneficial effect or even situations where being too silver-tongued can be a bad thing (like the Dean Domino conversation in Dead Money). There were a few occasions in AoD where this could happen, like the final mission in Teron where failing to pass checks to could lead to alternate paths, but I also didn't mind when conversations lead to an instant death or locked me out of content because the mechanical structure of the dialogues prepared me for that possibility.

However if a game has a more realistic, or natural system like what is proposed in that second dialogue system, I would expect more realistic outcomes rather than the binary fail or pass. One of the only examples I can think of for a game that did this was Alpha Protocol, where severely pissing off someone you are talking to could have beneficial effects, as they would be more careless due to their overwhelming hatred of you, and vice versa where someone liking you too much could interfere with their sense of duty. Otherwise despite implementing this more natural system it ends up boiling down to the exact same thing as the more simple system, just in a more round about way. Like "That brings us to the second biggest change. Most lines would no longer lead to success or failures but result in positive and negative reactions, represented numerically," to me seems like the exact same thing in both cases: a positive change is a success and a negative change is a failure. Or when you say "it didn’t matter how many checks you passed and how well you were doing until that last check that resulted in failure (i.e. early success didn’t contribute to anything and thus didn’t matter)," how is going through a dialogue tree in this new system and ending up with +6 positive rating but then getting a -2 in the last dialogue check and thus failing to get a net positive outcome any different from that previous system?

 

[Erebus]

The new system looks very promising. Even though CRPG dialogues have grown lengthier and more elaborate over the years, they're seldom used to challenge the player in interesting ways (either the dialogue contains no challenge at all, or the challenge is very easy, or it's entirely tied to the PC's skills). Forcing the player to really think about the arguments to use in a dialogue (much like he would think about the tactics to use in a fight) is an excellent concept, and having to use several arguments makes the whole thing more flexible.

There is a risk that, sometimes, the player might genuinely not understand why the option he chose was the bad one, and feel frustrated because of that. In some cases, a subtle, subsequent hint might be useful to explain the reason why the argument wasn't effective.



For added complexity (because why not ?), I'd suggest that, in some cases, the coherence (or lack of coherence) of the arguments used by the PC could play a role.

For instance, let's say the PC is trying to convince a young lord to provide some help against the powerful and villainous Lord Evil. After gathering some information, the player has deduced that the young lord is eager to acquire glory and worried that he might not live up to his ancient family's name. So, when the young lord asks why he should help the PC, the PC answers :

"Just think, my lord, how glorious it would be for you to have the courage to oppose Lord Evil ! Even your worthy ancestors have never faced such a dangerous and terrifying enemy !"

The young lord's reaction shows that he's tempted, and already dreaming about being a hero. But he's worried about the dangers of opposing the powerful Lord Evil. Let's imagine the PC then says :

"Don't worry, my lord. I know for sure that Lord Evil's reputation is exaggerated. His actual strength is much inferior to what it seems to be."

The argument in itself isn't necessarily bad, but it directly contradicts your first argument, in which you mentioned that Lord Evil was super-dangerous and that it would be heroic to oppose him.

Instead, if the PC says :

"All the other lords want to be rid of Lord Evil, for they know that he might one day send his army to invade their lands, burn their castles and kick their dogs. But none of them has the courage to be the first one to stand up. If you oppose Lord Evil, they'll quickly rally to you and send their forces to support you."

That second argument is strengthened by the first argument, instead of contradicting it.


Obviously, I'm not suggesting that coherence should be an issue in all dialogues (it would be way too much work, and unnecessary in many cases). But it could play a role in a few major dialogues.

 

[Absinthe]

I think there should also be some skillchecks which result in failure regardless of whether or not you meet the threshold simply because it was the wrong kind of solution to the problem. Break players out of that "if all you have is a hammer, everything looks like a nail" mindset a little. And I think it would be worthwhile to reward players for picking disposition penalizing choices and succeeding in the end because you obtained greater concessions/whatever in doing so, even if they don't necessarily like you, or maybe they will end up liking you for the disposition penalizing options because you didn't sugarcoat things and told it straight. You don't want to encourage players to just automatically pick the path of least resistance at all times. You want them to think about their options.

 

[vazha]

Is failure / success immediately evident or is it vague and takes time for the player to discover whether he made the right decision or not?

 

[Vault Dweller]

Immediately evident. It will be very clear when a person doesn't buy your arguments or bullshit. Nothing over the top, normal reactions.

 

[Stormcrowfleet]

As someone with almost 100 hours in AoD and almost all achievement while doing mostly talky or hybrid run, system #2 is more appealing.

That being said, it is only in so far as you guys put the time in the checks having meaningful impact. What I mean by this is:

Situation A
Success is triggered on 10 success rate. But because you have 10 in Speech, you can choose 3 time the worst argument and pass. What's the use of the worst argument ?

Situation B
Success is triggered on 10 success rate. But because you have only 1 in Speech, you can choose 3 time the best and still fail. What's the use of the best argument ?

IDK if it would be possible, but I'd still add some "you need at least X in the skill to see this option", and I'd still add some "because you have 10 in this skill you can see this super useful answer to go directly to the point and even maybe a unique discussion where you insta win" (I'm thinking here something along the lines of convincing Delar to send the bandits kill the mine Aurelian).

Do I make any sense ?

 

[Polydeuces]

The new system sounds great! However, would there be multiple possible outcomes? Say if you have max Intimidation and low Persuasion, your dialogue checks will change to reflect that you bullied them into accepting your strategy instead of sweet talking them?

 

[ShadowSpectre]

I like the proposed new system. What would also be good is a cumulative outcome if you talk to the same NPC or a contextually related one later (say you help them with one thing, then you offer to help again, it should be easier to achieve a more favorable outcome/be more forgiving after the first time and vice-versa if you got a negative outcome the first time around).

 

[Grampy_Bone]

One thing I liked about Kingdom Come Deliverance was if you failed one check you could try one of the other tacks. So if being nice didn't work you could switch to pulling rank, intimidating, bribery, etc. It worked okay.

But yeah the proposed system sounds good.

 

[Vault Dweller]

As someone with almost 100 hours in AoD and almost all achievement while doing mostly talky or hybrid run, system #2 is more appealing.

That being said, it is only in so far as you guys put the time in the checks having meaningful impact. What I mean by this is:

Situation A
Success is triggered on 10 success rate. But because you have 10 in Speech, you can choose 3 time the worst argument and pass. What's the use of the worst argument ?

Situation B
Success is triggered on 10 success rate. But because you have only 1 in Speech, you can choose 3 time the best and still fail. What's the use of the best argument ?

If you're asking if we're planning to balance the checks properly, the answer is yes (as these are balance not design issues).

IDK if it would be possible, but I'd still add some "you need at least X in the skill to see this option"...

While the game will have some hidden lines, we prefer to limit it to high stat lines (things that a low Per character simply won't notice and thus remain unaware of) and info acquired in-game. When it comes to lines, it's not that hard to figure out what to say. It's much harder to say it the right way, even if the line is something as simple as "compared to what?" (when your client says that your product is too expensive).

The new system sounds great! However, would there be multiple possible outcomes? Say if you have max Intimidation and low Persuasion, your dialogue checks will change to reflect that you bullied them into accepting your strategy instead of sweet talking them?

When appropriate.

 

[Whisper]

Your skill level increases positive reaction and reduce negative reaction, which is pretty much how it works in real life. Give 10 sales reps the same script (i.e. the same lines to say) and the results will be very different because the skillsets are different. What you say matters, of course, but how you say it matters more.

You sure of it? But they say same thing. So logically it shouldnt matter much how they say.