D&D and the RNG?

Brainstorm ideas of possible additions to the game. Read this before posting!

Moderators: Forum Moderators, Developers

Forum rules
Before posting a new idea, you must read the following:
CarpeGuitarrem
Posts: 250
Joined: November 19th, 2007, 7:46 pm
Location: One among the Fence

D&D and the RNG?

Post by CarpeGuitarrem »

So now that I've got attention...I'm going to spring some rather interesting musings that I had whilst wondering why people have so many complaints about the RNG. See, what I started wondering is maybe, if the RNG truly is fair, there's something in the way that randomness is presented that gives a false impression to the player. Thing is, I have a hard time dealing with some vague idea of randomness generated by a computer, and so I fell back to a good ol' standby: dice. Same difference, really, and when you think about it, all of Wesnoth's combat can be modeled with d10's. And that's when I came to a funny conclusion...

Wesnoth Combat and D&D Combat??
May seem weird to some, but it turns out that these two have more similarities than you might suspect. Don't believe me? Let me summarize the content of the two.

D&D Combat
  • Roll a d20 and apply your attack modifier
  • Check the result against the target's defense number
  • If the result exceeds the target's defense number, roll a die for your damage and apply your damage modifier
  • The target takes the result as damage to its HP
Wesnoth Combat
  • Roll a d10
  • Check the result against the target's defense number
  • If the result exceeds the target's defense number, check the damage for the attack
  • The target takes the damage number as damage to its HP
(Note: Yes, Wesnoth works with percentages, but actually, if you divide each percentage by ten, it doesn't change anything, because the percentages for defense go in increments of 10)

The similarities are eerie. Now, what's the big differences between the two? Besides the number scaling (which is really irrelevant, a d20 and a d10 are equally random), the only differences are that D&D adds static modifiers to its random results, Wesnoth has no randomness in the damage dealt, and Wesnoth lumps a bunch of combats back-and-forth into a single attack. The question is, what does this mean?

Ding Dong, the Bell Curve is Dead?
Well, not exactly. Wesnoth quite obviously does have a bell curve of sorts when it comes to probability, because it has multiple instances of randomness. Here's the deal, though. The bell curve isn't where people expect it. And considering that a bell curve provides stability, I think this is where people are having a problem with the RNG. Not with the randomness itself, but the fact that the randomness doesn't figure into stability. Instead of stability with a random element, it becomes all a random element, of sorts. Here's why.

In a roleplaying system such as GURPS, a bell curve is very evidently present, because the random results are added together. When you roll 3d6, your odds of getting a 3 are much lower than your odds of getting a 10, because there's only one way to get a 3, but many ways to get a 10.

In Wesnoth, however, while there are multiple random results, the results are not added together. Let's say you have 4 attacks for a unit. Those 4 attacks are separate entities, and have no relation to one another. If you're hoping for a result of over 40%, and you get 70%, that doesn't affect the other random results, and so the surplus % is not needed.

So, while there may be a curve, the curve is not to be found in the actual attack value itself, but in the number of successes. Which means that the chance of success for an individual attack is utterly and totally random, with no chance of stability. That's why people blame the RNG so much.

"What's the big deal?" you may ask. "When you look at the damage totals, it averages out right." Here's the thing, though. When someone looks at the game, they don't always see that. They don't see the big picture, but rather each individual sword swing. Because they know that when you swing a sword, and have been trained to do so, you don't have an equal chance to do poorly (getting a 10% result against any defense) or to do well (getting a 100% result against any defense). You have a good chance at doing well (getting a 50-60% result against any defense), and a much smaller chance at doing poorly (getting a 10% result) or very well (getting a 100% result).

In D&D, I hear a lot of people talk about "lucky" and "unlucky" dice. In Wesnoth, I hear a lot of people talk about "luck" and "unluck" in the RNG. Coincidence? Considering that each individual attack in each game is based off of a flat curve, I highly doubt it. In both cases, people often lose sight of the big picture, and forget that the damage dealt always evens out in the end.

So...
Does this mean that we need to change anything? No clue. After all, RISK follows the exact same model (even more so than D&D, actually), and it's managed to get along just fine. The only disadvantage we have when it comes to RISK is that RISK dice are real and tangible. The RNG is not.

So yeah, there's some thoughts I just had. Throwing those out there. I have no clue what it means, rather, I'm interested to pursue that thought because I can.
Glory in Blood...Needs Programming Help!

If you have time, check out my ongoing serial story...
The Hidden: Secrets of the Future's Past

User avatar
JW
Posts: 5046
Joined: November 10th, 2005, 7:06 am
Location: Chicago-ish, Illinois

Re: D&D and the RNG?

Post by JW »

Well, the thing with DnD is that yes, theres a d20. It actually makes somewhat of a significance that there are 20 steps as opposed to 10. Let me show you:

In DnD each difficulty to hit step (+1 AC) increases your chance to miss by 5%. In Wesnoth each step increases your chance to miss by 10%. Because of this there are only so many variations of difficulty available (although there could be far far more in Wesnoth - they are just not utilized).

Also in DnD you have the capability to improve your chance to hit. Levelling gives +BAB, +stats which affect hit, +feats, or for mages +spells (magic missle 100% chance to hit [which you get at lvl1 anyway...]). In Wesnoth there is no such capability unless your unit gains marksman or magic, which only 2 units do: the elf archer and the elf shaman. Both units are in 1 of the 6 factions.

There are also issues where the BAB for fighters decreases with extra strikes, so they generally will always hit at least once, whereas in Wesnoth this is not the case. When you add magical weapons to the mix....get my drift? There are just an incredible number of ways to alter the CTH on the user side of the equation that just are not available to the user in Wesnoth. If the elf is in forest you'd better have a mage, or prepare to miss a lot.

Another issue is how quickly units can die. In DnD generally fights against a small group last several turns whereas in Wesnoth units die more quickly. It's been well documented the power of Adepts v Drakes as the most extreme example.

When you add all of these things together, a minimal swing of luck at an inopportune time can drastically alter a game. For the benefit of everyone I will clarify how someone could rely on luck reasonably and then be drastically hurt when it fails him. I will give 2 scenarios:

1) your CTK an important ZOC unit is high (90%+) and is necessary to continue an attack. Failure will result in a brutal counter, and not attacking at all allows opponent to regroup after taking a small economic advantage. Let's say you have 2 mages that need 2 hits between the two of them to kill a blocking ghoul. The mages miss 5 times.

2) reverse of (1): you set a high defense unit in an important spot protecting an important hex you don't want the enemy to have. Your unit is at full hp and the opponent has no mages and only 2 hexes on bad terrain to attack from. If they kill your unit they get the high def terrain and a flank around your ZOC line. 2 grunts hit 4 times on your thunderer on mountain, and you are now going to be overrun.

In such cases you've made the best possible moves that you could. Your reliance on luck was the most reasonable approach to take. There was no greater certainty to succeed in the game. Unfortunately a small run of bad luck has potentially cost you the game quickly. 6 failures in luck can be devastating quite quickly in Wesnoth, just as 6 successes, or fewer (think high damage units like thunderers).

Anyway, this is even true in DnD....so the last statements don't separate, only distinguish one of the (not so great IMO) features of using luck.

And anyway, in Wesnoth there is a great variance in CTH: 30%-70% generally. Since these are so far from each other they can easily be mentally expanded.

hiro hito
Posts: 201
Joined: November 23rd, 2006, 8:00 am

Re: D&D and the RNG?

Post by hiro hito »

JW wrote:Unfortunately a small run of bad luck has potentially cost you the game quickly.
Yes I notice that usually only 2 turns are enough to deal with a game.
"Of course His Majesty is a pacifist. When I told him that to initiate war was a mistake, he agreed.Thus, gradually, he began to lead toward war."-Emperor Shòwa (Enlightened Peace)'s chief cabinet secretary

CarpeGuitarrem
Posts: 250
Joined: November 19th, 2007, 7:46 pm
Location: One among the Fence

Re: D&D and the RNG?

Post by CarpeGuitarrem »

JW: the use of a d20 versus a d10 is minimally significant, statistically. Rolling a 1 or 2 on a d20 is the same, statistically, as rolling a 1 on a d10. If you multiplied all defenses in Wesnoth by 2, and made the percentage be out of 200%, nothing would drastically change.

I did also account for the improvements as a difference: they're covered where I discussed the addition of a modifier. Which does increase the certainty of success in D&D. But the fact of the matter is, the core mechanic still operates very similarly.

D&D
Roll a d20, add modifiers, and compare to the target number.

Wesnoth
Roll a d10, and compare to the target number.

Which means that the probability rules behave the same.
Glory in Blood...Needs Programming Help!

If you have time, check out my ongoing serial story...
The Hidden: Secrets of the Future's Past

Lord_Aether

Re: D&D and the RNG?

Post by Lord_Aether »

I believe JW's point was that the player has more (at least perceived) control over the chances to hit in D&D; whereas in Wesnoth, it depends on the unit and terrain.

User avatar
JW
Posts: 5046
Joined: November 10th, 2005, 7:06 am
Location: Chicago-ish, Illinois

Re: D&D and the RNG?

Post by JW »

CarpeGuitarrem wrote:JW: the use of a d20 versus a d10 is minimally significant, statistically. Rolling a 1 or 2 on a d20 is the same, statistically, as rolling a 1 on a d10. If you multiplied all defenses in Wesnoth by 2, and made the percentage be out of 200%, nothing would drastically change.
No. 5% (1/20) is not the same as 10% (1/10). And measuring things in terms of 200% is also wrong.

I think you should understand statistics a little better before posting about them.....

***edited to be kinder
Last edited by JW on December 23rd, 2008, 7:01 am, edited 1 time in total.

User avatar
JW
Posts: 5046
Joined: November 10th, 2005, 7:06 am
Location: Chicago-ish, Illinois

Re: D&D and the RNG?

Post by JW »

Lord_Aether wrote:I believe JW's point was that the player has more (at least perceived) control over the chances to hit in D&D; whereas in Wesnoth, it depends on the unit and terrain.
Well, they have more actual control in that they can tune their characters for CTH. Also as they level they increase CTH. So, in fact as well as perception, there is in fact more control over CTH on the player's (used synonimously with attacker's) end.

CarpeGuitarrem
Posts: 250
Joined: November 19th, 2007, 7:46 pm
Location: One among the Fence

Re: D&D and the RNG?

Post by CarpeGuitarrem »

JW wrote:
CarpeGuitarrem wrote:JW: the use of a d20 versus a d10 is minimally significant, statistically. Rolling a 1 or 2 on a d20 is the same, statistically, as rolling a 1 on a d10. If you multiplied all defenses in Wesnoth by 2, and made the percentage be out of 200%, nothing would drastically change.
No. 5% (1/20) is not the same as 10% (1/10). And measuring things in terms of 200% is also wrong.
I never said that 5% and 10% were the same. What I said was that altering the number of sides on the die doesn't alter the fact that the mechanic is the same. I also specifically stated that the odds of rolling a 1 or a 2 on a d20 (which is 2/20, or 10%) was the same as the odds of rolling a 1 on a d20 (1/10, or 10%).

The 200% thing was just an example, although I do want to point out that the fact that Wesnoth uses percentages has no bearing on the randomness factor. It doesn't have to cap out at 100, because nothing adds up to that top percentage. Ergo, it's an arbitrary number.

I agree that D&D offers more control. As I've stated, that's a peripheral detail if you're looking at the core mechanics.
Glory in Blood...Needs Programming Help!

If you have time, check out my ongoing serial story...
The Hidden: Secrets of the Future's Past

User avatar
JW
Posts: 5046
Joined: November 10th, 2005, 7:06 am
Location: Chicago-ish, Illinois

Re: D&D and the RNG?

Post by JW »

CarpeGuitarrem wrote:I never said that 5% and 10% were the same. What I said was that altering the number of sides on the die doesn't alter the fact that the mechanic is the same. I also specifically stated that the odds of rolling a 1 or a 2 on a d20 (which is 2/20, or 10%) was the same as the odds of rolling a 1 on a d20 (1/10, or 10%).
That much I did misread.
The 200% thing was just an example, although I do want to point out that the fact that Wesnoth uses percentages has no bearing on the randomness factor. It doesn't have to cap out at 100, because nothing adds up to that top percentage. Ergo, it's an arbitrary number.
No, it is not.
I agree that D&D offers more control. As I've stated, that's a peripheral detail if you're looking at the core mechanics.
I don't understand how player control of luck is a "peripheral detail" if your original point is about the psychology of luck.

Another point I would like to make at this time is that DnD is more than just rolling the dice to determine the outcomes of battles. There is actual roleplaying involved, decisions made without luck involved at all, and at any time the DM can alter any outcome to what he prefers. The argument that luck is a "core mechanic" of DnD is therefore debatable - though that's not really important to the actual discussion about "luck and fairness."

CarpeGuitarrem
Posts: 250
Joined: November 19th, 2007, 7:46 pm
Location: One among the Fence

Re: D&D and the RNG?

Post by CarpeGuitarrem »

JW wrote:
The 200% thing was just an example, although I do want to point out that the fact that Wesnoth uses percentages has no bearing on the randomness factor. It doesn't have to cap out at 100, because nothing adds up to that top percentage. Ergo, it's an arbitrary number.
No, it is not.
How is it not an arbitrary number? It wouldn't matter if you removed the "%" symbol and called defenses "1", "2", "3", etc.
I agree that D&D offers more control. As I've stated, that's a peripheral detail if you're looking at the core mechanics.
I don't understand how player control of luck is a "peripheral detail" if your original point is about the psychology of luck.
I'm not looking at an exhaustive definition of everything that goes into why people have problems with the luck. Also, how do you know that player control over luck is what's going on here? It's just a hypothesis that the amount of control players have over luck is related to how "fair" they view the luck system to be. I'm looking at one facet of it.
Another point I would like to make at this time is that DnD is more than just rolling the dice to determine the outcomes of battles. There is actual roleplaying involved, decisions made without luck involved at all, and at any time the DM can alter any outcome to what he prefers. The argument that luck is a "core mechanic" of DnD is therefore debatable - though that's not really important to the actual discussion about "luck and fairness."
When I say "core mechanic", I'm using the terminology that Wizards of the Coast uses to describe D&D's d20 mechanic. Without it, it would be just roleplaying, and not d20/D&D.
Glory in Blood...Needs Programming Help!

If you have time, check out my ongoing serial story...
The Hidden: Secrets of the Future's Past

User avatar
JW
Posts: 5046
Joined: November 10th, 2005, 7:06 am
Location: Chicago-ish, Illinois

Re: D&D and the RNG?

Post by JW »

CarpeGuitarrem wrote:
JW wrote: No, it is not.
How is it not an arbitrary number? It wouldn't matter if you removed the "%" symbol and called defenses "1", "2", "3", etc.
Then they would have no relevance to the actual percentage chance to hit. Arbitrary? Perhaps, in the sense that elvish archer could have 65% defefnse in forest, or 89%, or 37.26%. The developers chose 70%, perhaps arbitrarily, but the number itself is not arbitrary as it has actual meaning.
I'm not looking at an exhaustive definition of everything that goes into why people have problems with the luck. Also, how do you know that player control over luck is what's going on here? It's just a hypothesis that the amount of control players have over luck is related to how "fair" they view the luck system to be. I'm looking at one facet of it.
I'm hardly giving an "exhaustive definition." I also don't even see how the "eerie similarities" to how luck is used in DnD and Wesnoth are even "looking at one facet" of how fair luck is. Especially when I've been trying to give you reasons why the "eerie similarities" are actually not so similar. If you want to take my counterpoints as an "exhaustive definition" then that's your interpretation. I'm simply pointing out flaws in your comparison.

User avatar
Thrawn
Moderator Emeritus
Posts: 2047
Joined: June 2nd, 2005, 11:37 am
Location: bridge of SSD Chimera

Re: D&D and the RNG?

Post by Thrawn »

Just like to say that I agree with JW on most of his points btw. Presentation is a bit lacking, but content is correct.

DnD and wesnoth isn't the best comparison
...please remember that "IT'S" ALWAYS MEANS "IT IS" and "ITS" IS WHAT YOU USE TO INDICATE POSSESSION BY "IT".--scott

this goes for they're/their/there as well

roadkill
Posts: 92
Joined: November 3rd, 2008, 11:48 am

Re: D&D and the RNG?

Post by roadkill »

another point, (talking about 3.5 since that is what im experienced in).

After you spec a character in Dnd (say a low-level fighter), You can min-max it so that character is VERY hard to hit.

It'll take most monsters a natural 20 to hit you (if the dm follows the excounter rules, to the letter). It make the game almost deterministic, you can be pretty certain your character will not die.

If you are playing with average players and the DM gives them easy encounters to keep them alive (no-one likes dieing, unless your a hardcore "roll"-player). Its mind-numbingly easy.

However on very very lucky rolls by the DM in dnd, especially at low levels... you can find yourself being insta-killed. Which people whine about, and the DM usually has to fudge.]#]

Wesnoth is a completly different kettle of fish though. I have't studied the RNG, but I'm pretty confident that its good enough, because even low-quality rngs can live upto the correct statistics. And unless you know what numbers are comming up (cheating) the wenoth one is FIIINE.

You talk about bell-curves.

The wesnoth bell curve for damage per attack with 4 attacks is fairly flat. quite a high chance to hit with no attacks, quite a high chance to hit with all 4. (assuming a 50% chance to hit all or nothing is a 6.5% chance).

With a multiple dice bell/curve you may make the bell taller and narrower. Which makes it easier to predict. Cue all the old topics about making wesnoth deterministic...

So what about stopping people complaining about the RNG? well, the new bell curve would make it exceptionally bad luck when something goes wrong... and there will be less chance for the luck to swing back in your favour.

So cue more people blaming the RNG (& quitting after a bad turn, because its more unlikely, even quitting after less significant things!) instead of LRN2PLY which these folks should be doing!

So I can't see how doing anything will help. And I think the wenothian bell curve is fine how it is (flat for one attack! and quite wide for subsequant attacks!)

CarpeGuitarrem
Posts: 250
Joined: November 19th, 2007, 7:46 pm
Location: One among the Fence

Re: D&D and the RNG?

Post by CarpeGuitarrem »

JW wrote:
CarpeGuitarrem wrote:T
How is it not an arbitrary number? It wouldn't matter if you removed the "%" symbol and called defenses "1", "2", "3", etc.
Then they would have no relevance to the actual percentage chance to hit. Arbitrary? Perhaps, in the sense that elvish archer could have 65% defefnse in forest, or 89%, or 37.26%. The developers chose 70%, perhaps arbitrarily, but the number itself is not arbitrary as it has actual meaning.
Having a percentage isn't everything, for those who aren't programmers and numbercrunchers. Most games I've played don't feature percentages in any prominent position. See, you may be looking at it as a programmer, but that's not the only valid viewpoint. Honestly, a typical gamer could care less. They know that higher numbers are better than lower, and there it stands. If they want more information, they can use the damage calculation button.
I'm not looking at an exhaustive definition of everything that goes into why people have problems with the luck. Also, how do you know that player control over luck is what's going on here? It's just a hypothesis that the amount of control players have over luck is related to how "fair" they view the luck system to be. I'm looking at one facet of it.
I'm hardly giving an "exhaustive definition." I also don't even see how the "eerie similarities" to how luck is used in DnD and Wesnoth are even "looking at one facet" of how fair luck is. Especially when I've been trying to give you reasons why the "eerie similarities" are actually not so similar. If you want to take my counterpoints as an "exhaustive definition" then that's your interpretation. I'm simply pointing out flaws in your comparison.
It was never a perfect analogy. I was trying to distill out a core similarity to prove a point about the bell curve, or lack thereof.

roadkill, there's a difference between deterministic and having a good chance to hit. If Wesnoth were to become a deterministic game in its current sense, that would involve the ability to max out unit stats, as you mention. In a game with a bell curve, though, it isn't deterministic, neither is it totally chaotic. Yes, it is more deterministic, but is that a bad thing? I'm not saying Wesnoth should be a deterministic game, and I'm not saying that it's bad as it is. I'm saying that players may take more positively if it has a small degree of determinism.

Also, it's not possible to have something go terribly wrong, sticking you in a "luck rut", just because you have a bell curve. In fact, the worst possible scenario (when you have a bell curve) happens even less often than if you have a flat curve. In a bell curve situation, the odds of getting a 1 (out of 10) might be 1/20, and the odds of getting a 2 might be 1/10. In a flat curve situation, though, the odds of getting a 1 would be 1/10, same as the odds of getting a 2, or a 5, or a 10. So actually, you get the terrible outcome (the bad turn) more often in the current Wesnoth system. Which means that you get less complaints of bad luck because of one bad turn, if you switch to a bell curve of some sort.
Glory in Blood...Needs Programming Help!

If you have time, check out my ongoing serial story...
The Hidden: Secrets of the Future's Past

User avatar
JW
Posts: 5046
Joined: November 10th, 2005, 7:06 am
Location: Chicago-ish, Illinois

Re: D&D and the RNG?

Post by JW »

CarpeGuitarrem wrote:Also, it's not possible to have something go terribly wrong, sticking you in a "luck rut", just because you have a bell curve. In fact, the worst possible scenario (when you have a bell curve) happens even less often than if you have a flat curve. In a bell curve situation, the odds of getting a 1 (out of 10) might be 1/20, and the odds of getting a 2 might be 1/10. In a flat curve situation, though, the odds of getting a 1 would be 1/10, same as the odds of getting a 2, or a 5, or a 10. So actually, you get the terrible outcome (the bad turn) more often in the current Wesnoth system. Which means that you get less complaints of bad luck because of one bad turn, if you switch to a bell curve of some sort.
So what exactly are you talking about when you say you want Wesnoth to switchto a "bell curve of some sort"? You wrote in your OP that:
CarpeGuitarrem wrote:Ding Dong, the Bell Curve is Dead?
Well, not exactly. Wesnoth quite obviously does have a bell curve of sorts when it comes to probability, because it has multiple instances of randomness. Here's the deal, though. The bell curve isn't where people expect it. And considering that a bell curve provides stability, I think this is where people are having a problem with the RNG. Not with the randomness itself, but the fact that the randomness doesn't figure into stability. Instead of stability with a random element, it becomes all a random element, of sorts. Here's why.
...
In Wesnoth, however, while there are multiple random results, the results are not added together. Let's say you have 4 attacks for a unit. Those 4 attacks are separate entities, and have no relation to one another. If you're hoping for a result of over 40%, and you get 70%, that doesn't affect the other random results, and so the surplus % is not needed.

So, while there may be a curve, the curve is not to be found in the actual attack value itself, but in the number of successes. Which means that the chance of success for an individual attack is utterly and totally random, with no chance of stability. That's why people blame the RNG so much.

"What's the big deal?" you may ask. "When you look at the damage totals, it averages out right." Here's the thing, though. When someone looks at the game, they don't always see that. They don't see the big picture, but rather each individual sword swing.
...so you want each individual swing to have a bell curve? How would that be done? I won't make any statements based off assumptions and will wait for your response instead.

Post Reply