(Dialogue) skill checks: which is better any why?

[JarlFrank]

Skill checks, especially in dialogues (persuasion, intimidation, seduction, but maybe also things like doctor), are a nice and very well-liked feature on the Codex. There are two (or three) ways to implement them, though.

The first way is a simple check whether your skill is high enough or not. If your persuasion skill is 50, you will succeed. If it's 49, you will not. Some games also have this with lockpicking: there's no skill check roll, there's just a check on how high your skill is and then you either can't even attempt to pick or you automatically succeed (Bloodlines did this).

That's a rather... well... not-so-good way of doing it, I think. Why the fuck will you fail the speech check with 49 skill points, which is just about as much as 50, while you will succeed with 50 all the time? What is this i dont even

Then there's the percentage type of implementation. Depending on your skill and the difficulty of the skill check, there's a certain chance of success. If you have a low skill and the check is difficult, you might have only 1% chance of success, while with a high skill and low difficulty it will go up as far as 99%. This way, there isn't as much of a difference between 49 and 50, since each skill point affects your chance for success.

This is considerably better than the previous way, but I prefer the D&D way, which is kinda in-between. Say, the skill check has a difficulty of 50, so you throw a D20 to determine success. you need to have at least 30 speech to be successful, and then only if you throw a 20. With 50 speech, you will always succeed. But with 49 speed, you will also succeed almost always - except if you throw a 1.

I think that's the best system. It requires the player to have a good skill, but it isn't as rigid as the "hard border" system.

So, which is better and why? DISCUSS!!

 

[SimpleComplexity]

I prefer the D&D way, which is kinda in-between. Say, the skill check has a difficulty of 50, so you throw a D20 to determine success. you need to have at least 30 speech to be successful, and then only if you throw a 20. With 50 speech, you will always succeed. But with 49 speed, you will also succeed almost always - except if you throw a 1.

I think that's the best system. It requires the player to have a good skill, but it isn't as rigid as the "hard border" system.

 

[curry]

Throwing dice is good only if you can't re-roll right after failing. Otherwise it's just stupid waste of time.

 

[torpid]

Softer checks can lead to save-scumming though, so the skill check should close off if it fails. The player should have 2-3 chances at most, and then the lock jams, the computer shuts down or the person you're speaking to tells you to beat it.

 

[KalosKagathos]

The first way is a simple check whether your skill is high enough or not. If your persuasion skill is 50, you will succeed. If it's 49, you will not. Some games also have this with lockpicking: there's no skill check roll, there's just a check on how high your skill is and then you either can't even attempt to pick or you automatically succeed (Bloodlines did this).

That's a rather... well... not-so-good way of doing it, I think. Why the fuck will you fail the speech check with 49 skill points, which is just about as much as 50, while you will succeed with 50 all the time? What is this i dont even

The problem in your example is the spread of skill values. It shouldn't be more than 1-10 (lol at Fallout 2 with its 1-300 scale). That way each point is an investment, especially if XP is limited, and hard checks are much easier to swallow, as the difference between neighbouring skill values is 10% of the scale, not 1%. Bloodlines was on the right track, but of course Troika had to fuck it all up with Blood Buff and Research.

In general, I think hard checks are better because soft checks inevitably introduce save game abuse.

 

[SCO]

Easily implemented if there is no "goodbye" option or more generally a cycle in the graph of the dialog whenever there is a check.


Though people like to talk again to npc's to refresh their memories...

Oh you mean save scum? Despicable undermen.

 

[torpid]

The first way is a simple check whether your skill is high enough or not. If your persuasion skill is 50, you will succeed. If it's 49, you will not. Some games also have this with lockpicking: there's no skill check roll, there's just a check on how high your skill is and then you either can't even attempt to pick or you automatically succeed (Bloodlines did this).

That's a rather... well... not-so-good way of doing it, I think. Why the fuck will you fail the speech check with 49 skill points, which is just about as much as 50, while you will succeed with 50 all the time? What is this i dont even

The problem in your example is the spread of skill values. It shouldn't be more than 1-10 (lol at Fallout 2 with its 1-300 scale). That way each point is an investment, especially if XP is limited, and hard checks are much easier to swallow, as the difference between neighbouring skill values is 10% of the scale, not 1%. Bloodlines was on the right track, but of course Troika had to fuck it all up with Blood Buff and Research.

In general, I think hard checks are better because soft checks inevitably introduce save game abuse.

Yeah I agree with this. I just reread my reply and realized that I derped: preventing multiple tries only works if the player doesn't save-scum. That's why the ideal is a point buy system with a limited amount of levels per skill, and hard checks.

 

[Excommunicator]

I hate most random rolls and I hate strict arbitrary checks.

So, neither.

 

[Leimrey]

I prefer the first method (where your skill must be high enough or you will fail all the time) as it eliminates the purpose of save/load scumming and provides a better incentive for the player to raise those non-combat skills that are used in the dialogue checks.

 

[Mastermind]

Skill checks, especially in dialogues (persuasion, intimidation, seduction, but maybe also things like doctor), are a nice and very well-liked feature on the Codex. There are two (or three) ways to implement them, though.

The first way is a simple check whether your skill is high enough or not. If your persuasion skill is 50, you will succeed. If it's 49, you will not. Some games also have this with lockpicking: there's no skill check roll, there's just a check on how high your skill is and then you either can't even attempt to pick or you automatically succeed (Bloodlines did this).

That's a rather... well... not-so-good way of doing it, I think. Why the fuck will you fail the speech check with 49 skill points, which is just about as much as 50, while you will succeed with 50 all the time? What is this i dont even

Because almost enough is never good enough. I generally prefer this system. The rest of your post is essentially an endorsement of save scumming.

Anyway, there are other options.

Your skill level determines whether you can attempt to perform a certain action. So if your skill level is 50, you can attempt to lockpick locks of level 50 and below, or persuade characters with their own speech skill at level 50 or below. The details could be filled in with active or passive abilities (like perks in fallout). I'm actually thinking of using a similar system in a game I'm designing. I haven't gotten the speech skills part down yet (it will need the most work as speech is generally handled in the shittiest way imaginable in pretty much every game I can think of), but I can give you a general idea of how it will work:

Each skills (say we have speechcraft in this case) has a value from 0 (unknown) to 100 (grandmaster). With the first level of the skill you gain that skill's base ability. In the case of speechcraft, it's the ability to persuade someone using logical arguments (Do X because X is good to do). As you gain levels, you can learn new abilities related to your skills. So you could emotionally manipulate them (do X because your dead child would want you to), flirt with them (hey, you're cute, do X for me would you babe?), appeal to survival instincts (do x or your house will get murdered, your dog will get raped and your wife will get burned to the ground), appeal to greed (do x and you'll be richer than you can imagine). Different NPCs can be more susceptible to different abilities. A nicedwarf scientist would be more susceptible to logical arguments than to idle threats, so presumably someone with a speechcraft of 30 using logical argumentation could persuade him while someone with a speechcraft of 50 using idle threats might not. This means a 49 speechcraft character could potentially win out whereas a 50 speechcraft character would not. At the same time there is no randomness (though there is some uncertainty if your character cannot read another character's personality properly) to abuse.

 

[Shemar]

Like someone else already said the problem lies in the scale of the skill, not the way the check is implemented. My favorite system only has 5 levels of each skill, each level costing exponentially more, so there is no 'almost but not enough'. And depending on the type of action, a skill check die roll may or may not be required. If the skill check is something the character will be able to try again and again, like picking a lock without any time pressure, or something without much of a random element (climbing a wall or breaking down a door) then there is no point is doing a skill check. If a skill is a one off attempt, like a dialogue persuation or intimidation, then a skill check is more appropriate. Regardless of whether a check is used or not, skill checks should be a lot more reasonable than how DnD does them. A d20 is way too wide a random element for my taste. Of course a good DM will set up the DCs so that the chance of success will be something realistic, but a human DM has the advantage of knowing exactly what the player's capabilities are when designing his adventure. A computer does not.

 

[Jaesun]

This is considerably better than the previous way, but I prefer the D&D way, which is kinda in-between. Say, the skill check has a difficulty of 50, so you throw a D20 to determine success. you need to have at least 30 speech to be successful, and then only if you throw a 20. With 50 speech, you will always succeed. But with 49 speed, you will also succeed almost always - except if you throw a 1.

I like this better, AND as long as there is no feedback to the player on succeeding or failure. For example:

CHARNAME: Oh please Mr Guard, guarding the royal treasure room! I shall bribe thee with 100 gold!
(engine does roll)
Guard: I cannot! I shall fulfill my duties as a guard. Begone filthy peasant!

not like:
Guard [FAIL] I cannot! I shall fulfill my duties as a guard. Begone filthy peasant!

 

[Darth Roxor]

All values should be hidden. For checks, rolls and outcomes. That's the best way to prevent save-scumming.

If I see [SPEECH: 90%] and I fail, I know that reloading is almost bound to make me succeed. If I only see [SPEECH], it will make me more hesitant to reload because I might be wasting my time if I have a 5% chance to succeed.

And if there's no [SPEECH] bracket at all, I probably won't even know that I just failed a check, so why the fuck would I even think about reloading.

So yeah, hide as much as possible for the greater good.

 

[JarlFrank]

Yeah, I agree with the hiding. Speech checks should look like any other dialogue option. Giving it them a special color like Bloodlines did is okay, too, but more than that - like putting the [Speech: 80%] like Fallout 3 did - is a bad idea. Not only does it encourage save scumming, it also makes players decide what to say by looking at the values. "Oh, I only have a 20% chance of success so I won't even try. If the game hadn't told me I'd have chosen this dialogue option. Thank you, game!"

@Excommunicator:
How would you do dialogue skills, then? I'm curious if you got a better solution.

 

[Unradscorpion]

Wait, so what are the pros of the soft limits?

 

[Mastermind]

Wait, so what are the pros of the soft limits?

It requires the game to hide your own rolls from you. I always love it when RPGS do that.

 

[deuxhero]

This is considerably better than the previous way, but I prefer the D&D way, which is kinda in-between. Say, the skill check has a difficulty of 50, so you throw a D20 to determine success. you need to have at least 30 speech to be successful, and then only if you throw a 20. With 50 speech, you will always succeed. But with 49 speed, you will also succeed almost always - except if you throw a 1.

49+1=50

 

[Darth Roxor]

This is considerably better than the previous way, but I prefer the D&D way, which is kinda in-between. Say, the skill check has a difficulty of 50, so you throw a D20 to determine success. you need to have at least 30 speech to be successful, and then only if you throw a 20. With 50 speech, you will always succeed. But with 49 speed, you will also succeed almost always - except if you throw a 1.

49+1=50

I think a more important remark is why would 49 speed win you a speech check.

 

[Antihero]

This is considerably better than the previous way, but I prefer the D&D way, which is kinda in-between. Say, the skill check has a difficulty of 50, so you throw a D20 to determine success. you need to have at least 30 speech to be successful, and then only if you throw a 20. With 50 speech, you will always succeed. But with 49 speed, you will also succeed almost always - except if you throw a 1.

49+1=50

Not sure if you're serious, but something against critical failures?

 

[JarlFrank]

Er... damn. Calculus was never my strength.

But yeah, critical failure can be introduced for checks like that. Or maybe a 0 roll.

 

[deuxhero]

This is considerably better than the previous way, but I prefer the D&D way, which is kinda in-between. Say, the skill check has a difficulty of 50, so you throw a D20 to determine success. you need to have at least 30 speech to be successful, and then only if you throw a 20. With 50 speech, you will always succeed. But with 49 speed, you will also succeed almost always - except if you throw a 1.

49+1=50

Not sure if you're serious, but something against critical failures?

Critical failures/success is on attack rolls and the odd strange technique (and possibly ability rolls?) in the d20 system, skill rolls are straight checks of if skill level+modifiers+roll>=difficulty.

 

[zeitgeist]

I think a more important remark is why would 49 speed win you a speech check.

You'd be a fast-talker?

 

[deuxhero]

Also, Bloodline's system of a low "points" (0-5) system benefited avoided the "50 points needed at 49 points", as 4 points in presuade was vastly better at it than having 3 points.

 

[Jaesun]

Giving it them a special color like Bloodlines did is okay, too, but more than that

I disagree. Hide the skill check completely. If a player is raising their speech skill, it should be organic and completely invisible to the player (as in special colors or [SPEECH])

 

[Mastermind]

Giving it them a special color like Bloodlines did is okay, too, but more than that

I disagree. Hide the skill check completely. If a player is raising their speech skill, it should be organic and completely invisible to the player (as in special colors or [SPEECH])

This sounds almost as dumb as Bioware's "hide what your character is actually going to say" system. Hiding their own rolls from the player is newfag derp. If I'm playing an RPG, I want detailed output of my actions easily accessible (a concept that seems to be long dead). Your system would at best work for an adventure game with light RPG elements. But solving the problem of save scumming like this is like hiding damage rolls to solve Fallout's combat save scumming. It's the equivalent of removing a pimple with a chainsaw.