Unpopular D&D Opinions

[kyoryu]

Then they’ll switch to Rapier and Shield, or be slightly behind.

“It is possible to build a bad character” is not a strong critique of a game. If it’s EASY, that’s an issue, but simply being possible isn’t bad.

Not at all.

But you should be able to build a moderately reasonable character with a "cursory" understanding of the rules, making choices that appear reasonable at the surface level. By "cursory" I mean things like "oh, yeah, Fighters use Str mostly, with some Con and a bit of Dex, so I should probably prioritize those".

No system (or at least very few) can stop you from making a deliberately bad character. That's not an important bar, I don't think.

[JNAProductions]

Not at all.

But you should be able to build a moderately reasonable character with a "cursory" understanding of the rules, making choices that appear reasonable at the surface level. By "cursory" I mean things like "oh, yeah, Fighters use Str mostly, with some Con and a bit of Dex, so I should probably prioritize those".

No system (or at least very few) can stop you from making a deliberately bad character. That's not an important bar, I don't think.

And... That character works fine in 5E.

If you see "Fighters need Strength, Dexterity, and Constitution," build a normal Human Fighter with the statline of 16 14 14 10 8 10, and put ASIs in round-robin fashion to Dex, Con, and Str... You'll be fine.

You'd have, at level 8, 18 Strength, 16 Dexterity, and 16 Constitution. You're a bit behind a more optimized Fighter, especially in your Wisdom save... But in terms of fighting, you'll do absolutely fine.

[kyoryu]

And... That character works fine in 5E.

If you see "Fighters need Strength, Dexterity, and Constitution," build a normal Human Fighter with the statline of 16 14 14 10 8 10, and put ASIs in round-robin fashion to Dex, Con, and Str... You'll be fine.

You'd have, at level 8, 18 Strength, 16 Dexterity, and 16 Constitution. You're a bit behind a more optimized Fighter, especially in your Wisdom save... But in terms of fighting, you'll do absolutely fine.

Agreed.

The next interesting thing is then, practically, what the optimization benefits are. I usually think of three benchmarks in terms of overall balance in character design:

1) New player, with a reasonable but cursory understanding of the rules, making reasonable-looking decisions
2) Experienced player with a more in-depth understanding of the rules, making more optimal but still logical decisions
3) Veteran player with an in-depth understanding of the rules, making highly optimal decisions that appear counter-intuitive

"How bad of a character can you make if you deliberately sabotage yourself" isn't in the list. Also, I presume that those are different numbers - the only question is how different they are.

[OldTrees1]

I am not sure how many time it is required to post the anydice link for this. https://anydice.com/articles/4d6-drop-lowest/

It also shows us that the average roll is roughly 16, 14, 13, 12, 10, 9. This is pretty close to the D&D 5e elite array (15, 14, 13, 12, 10, 8).

Yes, you are right that the Standard 5E array is slightly worse than the expectation of 4d6b3. However it is rather close.

Agreed.

The next interesting thing is then, practically, what the optimization benefits are. I usually think of three benchmarks in terms of overall balance in character design:

1) New player, with a reasonable but cursory understanding of the rules, making reasonable-looking decisions
2) Experienced player with a more in-depth understanding of the rules, making more optimal but still logical decisions
3) Veteran player with an in-depth understanding of the rules, making highly optimal decisions that appear counter-intuitive

"How bad of a character can you make if you deliberately sabotage yourself" isn't in the list. Also, I presume that those are different numbers - the only question is how different they are.

Also the benchmark(s) the game designers chose to calibrate the game at. For example a game calibrated to #1 is going to be different for new players than one calibrated at #2. This is independent of how different #1 and #2 are.

In 5E D&D for example, I could start with a 14, and keep it a 14. However I would hope the rules guide the new player to get from a 14 to 16 by 17th level. Anything above that is optional in 5E.

[Mordar]

Assume point buy or the pathetic "standard array" (which is quite frankly wrong based on math)?

I am not sure how many time it is required to post the anydice link for this. https://anydice.com/articles/4d6-drop-lowest/

Looking at the results, the average rolled array should be 16, 14, 13, 12, 10, 9.

The 15, 14, 13, 12, 10, 8 standard array is lower.

Is that slight penalty just intended to offset/disincentivize based on the safety the array offers? You don't end up with two or three single digit scores, or anything more than slightly below the average for a given stat.

In practice I doubt it has that effect, since people would scrap any rolled array that was less attractive because virtually no one enforces the idea that you have to play what you roll.

- M

[Anonymouswizard]

Have you tried 2e? Super simple character creation… with numerous options (from kits to Skills and Powers to a complete "build your own class" point buy) to "complicate" things enough to allow nearly any concept.

Yes, still not convinced it's better than point but. Seems complicated in all the wrong ways.

I'm also not against default character creation giving players pools of points to put into discussion categories. 15 skill points spent as I wish feels more meaningful than getting 16 and told I can only get the full benefit by selecting from this sub list.


A lot of systems could stand to more explicitly lay out what they think a decent starting character is. D&D has more problems just because it's more likely to be people's first game.

EDIT: also, a lot of groups either use a rolling method that produces significantly higher values, do some variety of 'if your roll is worse than the standard stay you can reroll/take the array, or both. I once had a character begin with FOUR TWENTIES because of a poorly designed rolling method (4d6b3 and 2d6+6 being my preferred if I have to roll, and generally creating reasonable arrays).

[Pex]

D&D needs to drop the automatic stat increases & feats, they need to be an option in the DMG & DM/world building sections of the splat books. They cover some nasty flaws in design about certain classes basically not changing over 20 levels while others get more and more tools to shape game play.

Unpopular, yes. Accurate, no. ASIs and feats are neither symptoms nor causes of the problem you describe. Their addition to the game since 3E provide for character customization which is a good thing. People quibble in their implementation across the editions, but the core concept is sound and a positive to the game.

[Anonymouswizard]

Honestly, whine the way they work in 3.X is now popular on this board, I like the way ASIs (and Feats) are tied to class in 5e. Yes, I don't like classes, but if we're going to assume they have to be there I like the idea that messing around with multiple classes stops your raw capability (as measured by stats) from getting too high. In theory the paid is that there's a lot more things you can do, but D&D isn't really set up to reward it.

What I'd do in D&D is separate Feats and ASIs and give about three of each to most classes. Makes those extra ability points more precious while mitigating the risk of a new player ending up too many points behind that +4 they should have at later levels. But then again if I had to have levels I'd likely turn D&Dv into Modern AGE, which gives a stat point and a rank in a Talent (read: level in a three feat chain) every level send ignores such silly things as 'class'.

[Telok]

Unpopular, yes. Accurate, no. ASIs and feats are neither symptoms nor causes of the problem you describe. Their addition to the game since 3E provide for character customization which is a good thing. People quibble in their implementation across the editions, but the core concept is sound and a positive to the game.

You're right, they aren't symptoms or causes. I said they are covers, patches, bandaids, plasters over the problem that about 1/3 of the classes are stuck with "get more fighty numbers" for the entire game. Which itself wouldn't be an issue if they had good class options beyond combat. Playing game without ASIs & feats just takes the bandage off so you can smell the pus.

Here's another one: Its bad design when good game advice boils down to "use the core dice mechanic as little as possible outside this one subsystem". 5e runs best when you use the dice as little as possible outside combat.

[Pex]

You're right, they aren't symptoms or causes. I said they are covers, patches, bandaids, plasters over the problem that about 1/3 of the classes are stuck with "get more fighty numbers" for the entire game. Which itself wouldn't be an issue if they had good class options beyond combat. Playing game without ASIs & feats just takes the bandage off so you can smell the pus.

Here's another one: Its bad design when good game advice boils down to "use the core dice mechanic as little as possible outside this one subsystem". 5e runs best when you use the dice as little as possible outside combat.

That I agree with. DMs should be able to make rulings. They should not have to make up their own game system to run part of the game. The game designers should have done that. That's why the game system is bought.

[Tanarii]

What I'd do in D&D is separate Feats and ASIs and give about three of each to most classes.

That's probably work well if you could turn an ASI into a feat-lite, and a Feat into an ASI lite. Something like an ASI can be turned into a 2/3 of a feat value, and a feat could be turned into a +1 plus 1/3 of a feat value. In other words, getting less than a full swap but giving you an option is you really didn't want the asi or feat.

[LibraryOgre]

Point buy is better than classes.

I tend to agree, though, I get weird with it.

As I mention, a lot, because I am a shameless self-promoter, I worked up a version of 2e AD&D that uses a modification on the Player's Option Point Buy classes to make the entire game point buy. I'm low key, sometimes, working on a means to even make your racial characteristics point buy (in that you buy everything, either from a DM-determined pool of points, or using your point total to determine XP table). Want to play an elf wizard? You pay more XP to have a lot of elfy abilities... or you could just say "I'm an elf wizard" and not take anything remotely elfy. Or take some elfy things you want, and leave the things you don't.

This does a few things on the 2e chassis (for good or for ill):

It lets you make the character you want
It put character creation at game start, then leaves it alone.

In 3e, and to a lesser extent 4e and 5e, character creation is an ongoing process, picking up new classes and subclasses, with feats adding abilities you never had before. In AD&D, if you picked a class, you were part of that class, barring some unusual circumstances. While you had a couple exceptions where you got more abilities (ranger and paladin casting, druids and monks and their abilities), for the most part, what you could do at level 1 was broadly what you could do at level 20... but you could do it better at level 20.

[Quertus]

Yes, still not convinced it's better than point but. Seems complicated in all the wrong ways.

What?

First, I wasn't saying it was better than point buy, I was saying that it *is* point buy (if a still partially silo'd point buy).

Second, it *can* be as simple as "write class on sheet, done". Not terribly complicated there.

Third, the rest certainly *can* be complicated, as… that's kinda… in the nature of point buy. But… I'm curious why you feel it's complicated in all the wrong ways, and what you think the *right* ways to be complicated look like.

I tend to agree, though, I get weird with it.

As I mention, a lot, because I am a shameless self-promoter, I worked up a version of 2e AD&D that uses a modification on the Player's Option Point Buy classes to make the entire game point buy. I'm low key, sometimes, working on a means to even make your racial characteristics point buy (in that you buy everything, either from a DM-determined pool of points, or using your point total to determine XP table). Want to play an elf wizard? You pay more XP to have a lot of elfy abilities... or you could just say "I'm an elf wizard" and not take anything remotely elfy. Or take some elfy things you want, and leave the things you don't.

This does a few things on the 2e chassis (for good or for ill):

It lets you make the character you want
It put character creation at game start, then leaves it alone.

In 3e, and to a lesser extent 4e and 5e, character creation is an ongoing process, picking up new classes and subclasses, with feats adding abilities you never had before. In AD&D, if you picked a class, you were part of that class, barring some unusual circumstances. While you had a couple exceptions where you got more abilities (ranger and paladin casting, druids and monks and their abilities), for the most part, what you could do at level 1 was broadly what you could do at level 20... but you could do it better at level 20.

So… is it like the (horribly limited) "point buy" in the DMG, but filled in with options from Skills and Powers? If so, I'll have to take a look.

[Anonymouswizard]

What?

First, I wasn't saying it was better than point buy, I was saying that it *is* point buy (if a still partially silo'd point buy).

Second, it *can* be as simple as "write class on sheet, done". Not terribly complicated there.

Third, the rest certainly *can* be complicated, as… that's kinda… in the nature of point buy. But… I'm curious why you feel it's complicated in all the wrong ways, and what you think the *right* ways to be complicated look like.

2e isn't point buy until Player's Options, and I don't know anybody who uses it. Then there's the issue about having to go through twenty million kits to find the one that fits your concept (and unit people who played it back in the day I don't know which book has which kit). And if I ignore all that I'm left with a class system that, while better than 3.X's and 5e's, is still incredibly lacklustre.

Honestly if you're going to have classes both 4e (with many specific roles) or BD&D (with it's broad archetypes) are better. Which reminds me, I need to pick up s physical copy of the Rules Cyclopaedia.

[Mordante]

ABSOLUTELY THIS

The Fighter is an NPC class that exists to give NPCs a feat and extra HP. Should NEVER be used by a PC.

I strongly disagree with this.

[oxybe]

Classes and Point Buy are solutions to 2 different problems IMO.

Classes are great if you want to go from level 1 to 20 and go through the whole zero to hero gamut.

Point buy is great if you really want to explore that specific level 7-10 sweet spot

Classes allow you to dole out progression in specific doses at a specific time. It also allows you to control when players should be gaining certain abilities or type of abilities and the challenges they should be facing. As a designer you can decide that at level 5 people start gaining limited flight and short range teleportation by simply not making those types of abilities available pre-level 5. Does that mean it restricts what kind of character you can or cannot make? sure, but not every concept should be viable in every game. if you REALLY want to play Blippo the flying Kobold at level 1, talk to your game master but if we're playing the (TOTALLY NOT D&D THEORETICAL GAME) Cellars & Centurions, flight is by design limited to level 5 and up.

and that's fine because level 5 is also where decent AoE starts kicking in as well multi-person support abilities. levels are always giving you something to look forward to, or at least they should.

OK, but what about point buy?

Point Buy i've found is... kinda bad at expanding the scope of what a character can do in the long term.

It's really good at exploring it's given band of strength or skill or whatever, but seems to fail hard at progressing through those bands in a reasonable fashion because it's too open and it's usually pretty easy to overspecialize or jack-of-all-trades yourself by spreading your points too thinly in one direction or another and can make scenario design... weird. but by assuming a base level of competence and giving out specific point-based packages with some differences in degrees of success but not scope, you can still keep a sense of progression that makes levelling up fun.

Does this mean that Blippo the flying Kobold will never be the true terror of the skies? yeah. probably, but that's because the game isn't meant to handle the power band of a hypersonic kobold zipping around and making strafing runs, but it is meant to be at a part where blippo can fly and do other things pretty darn well. he'll get more consistent with them over time, but it probably won't break the scope of the initial design.

[Xervous]

OK, but what about point buy?

Point Buy i've found is... kinda bad at expanding the scope of what a character can do in the long term.

Designing and testing a point buy system is giving me headaches when it comes to progressing abilities. The big question is how do you gate it? I have some strong incentives in place to diversify on skills and attribute scores yet my players seem utterly convinced that they need to race their way up to the highest ability ranks while ignoring every other part of the character.

“Okay people you’ve got 1800p characters now. Remember we all started at 600, now it’s time for a party stealth roll.”

And the assassin rolls more total successes than the other 4 party members combined after the bard rerolls his 0.

Shadowrun prevents deep plunges down one avenue with training requirements (but hey you can train multiple things at once). I could tune my numbers but I fear the better option might just be gating by point budget.

So yeah, if you want consistent tuning on a point buy system it needs to be constrained to a narrow range.

[Satinavian]

Splittermond gates higher abilities and skills by total used points. Works reasonably well.

[LibraryOgre]

So… is it like the (horribly limited) "point buy" in the DMG, but filled in with options from Skills and Powers? If so, I'll have to take a look.

Yep. In fact, if you're building your own XP table (instead of the DM giving you a pool of points and an XP table), you use your point total divided by 20 and compare it to the table in the DMG. I talk a bit more about it in this thread in Older Games. (Post 19 does a fair bit of the work)

2e isn't point buy until Player's Options, and I don't know anybody who uses it. Then there's the issue about having to go through twenty million kits to find the one that fits your concept (and unit people who played it back in the day I don't know which book has which kit). And if I ignore all that I'm left with a class system that, while better than 3.X's and 5e's, is still incredibly lacklustre.

A term I use to describe a lot of class systems is "chunky". If I want to play an archer in 2e, I have a few good choices, with fighter, ranger, thief, and bard being the best among them. But if I take one of those, I have to take the rest of it. My fighter-archer has the ability to wear really heavy armor, even if I don't see him that way. My bard-archer can cast spells and sing people happy, even if I just want to be able to shoot stuff. My ranger-archer has to be good and is inexplicably a supreme two-weapon fighter, because those are the size of the chunks.

Kits let you reshape the chunks a bit, but they're still chunks, and they may have their own weird bits... like a lot of archer-specific kits become incompetent at melee fighting.

Chunky game design is fast to create characters; to make a Castles and Crusades character, pick your class, pick your race, roll your stats, and pick one or two primes, and you're done. If the CK has papers with each individual class and race written on them, he can print them and hand them to you. Super easy. But your fighter is going to be like almost every other fighter, except for one or two small choices. Your thief will be like every other thief.

The more options you introduce, the finer control usually is, but that increases time, as well.

Honestly if you're going to have classes both 4e (with many specific roles) or BD&D (with it's broad archetypes) are better. Which reminds me, I need to pick up s physical copy of the Rules Cyclopaedia.

TBH, I love 4e's roles determining much of your powers, and source determining a lot of your aesthetic. While I found their powers system to be way too much (pages and pages of individual powers), that game had a ton of good ideas.

[NichG]

So yeah, if you want consistent tuning on a point buy system it needs to be constrained to a narrow range.

Seems correct.

Or maybe make it necessary to scavenge lots of advantages from different places if you really want to break into an extreme range.

E.g. you can get up to 3 dice from a skill, up to 3 from a stat, by getting a buffing magic you can add 1 die to anything, by improving your Luck you can reroll up to three dice in any roll, by improving your Destiny you can lower the target number for success on each die, if you get enough Wealth you can get gear or special locations which add flat successes, ...

So by the time you're getting 10 successes on a Stealth check when the normal range for NPCs in the world is 1-3, you're going to have to have picked up other stuff along the way because depth caps fast, even if what particular other stuff you pick to make it happen is still up to you.

Another way is to buy abilities rather than numbers. It costs such and such to become able to climb a sheer wall. It costs such and such to stay out of sight for 3 seconds once per minute, a different amount to share stealth with the party, a different amount to shape fire, etc. Focus on what you can do and not dice pools or mods.

[Xervous]

Seems correct.

Or maybe make it necessary to scavenge lots of advantages from different places if you really want to break into an extreme range.

E.g. you can get up to 3 dice from a skill, up to 3 from a stat, by getting a buffing magic you can add 1 die to anything, by improving your Luck you can reroll up to three dice in any roll, by improving your Destiny you can lower the target number for success on each die, if you get enough Wealth you can get gear or special locations which add flat successes, ...

So by the time you're getting 10 successes on a Stealth check when the normal range for NPCs in the world is 1-3, you're going to have to have picked up other stuff along the way because depth caps fast, even if what particular other stuff you pick to make it happen is still up to you.

Another way is to buy abilities rather than numbers. It costs such and such to become able to climb a sheer wall. It costs such and such to stay out of sight for 3 seconds once per minute, a different amount to share stealth with the party, a different amount to shape fire, etc. Focus on what you can do and not dice pools or mods.

Part of the allure I see in point buy is a streamlined delivery method for features and the opportunity for WYSIWYG options. You want big numbers on <thing>? Just increase <thing> and the one related <other thing>. If you’ve got abilities that do nothing beyond adding numeric bonuses you’ve just introduced hidden discounts that add extra steps and learning to character creation. In some cases this may be good, like allowing Green invested characters to have cheaper and higher progression on armor values because that’s their niche. IMO abilities in point buy should generally be about new ways of doing things, new things to do, or situational/tradeoff effects.

To address my concerns of players not wanting to invest anything in most skills I implemented something like a 401k match for skill ranks. Say you currently have access up to tier 7 abilities, each rank you have gets ‘matched’ up to half your tier access (rounded up). It doesn’t mandate that the high tier barbarian suddenly gets good at sneaking, but it allows him to get on the curve for a modest investment

[KorvinStarmast]

It also shows us that the average roll is roughly 16, 14, 13, 12, 10, 9.

Which is what Standard Array Should Be. I am very much on side with Pex regarding "Having an 18 stat at level one should not be prohibited!"

Yes, you are right that the Standard 5E array is slightly worse than the expectation of 4d6b3. However it is rather close.

It is three Point Buy points worse.
Watch what I do with Point Buy and 27 points to achieve standard array, by the way.
The 15, 14, 13, 12, 10, 8 ~ 8 {+9 +7 +5 + +4 +2 +0} is 27 points. Which means that Point Buy should be either 29 or 30 points, not 27. to approximate dice rolling.
But it isn't. Oh well, so it goes.

In 5E D&D for example, I could start with a 14, and keep it a 14. However I would hope the rules guide the new player to get from a 14 to 16 by 17th level. Anything above that is optional in 5E.

Phoenix pointed out that in Tier 4 the 18 is where the model fits for a prime requisite, but I've got an example to share: good friend of mine did a Mountain Dwarf Vengeance Paladin with a 16 Strength who never bumped strength, at all, from level 1 through 10. (He tended to pump Charisma and take feats). He did fine (helped that he eventually found a belt of hill giant strength, yes it did, and magic items are a part of the game, they just aren't a guaranteed part ...)

[JNAProductions]

Is point buy or the standard array claiming to be equal to dice rolling?

Because it’s better than 3d6, slightly worse than 4d6b3, and scads worse than 6d10b2.

[KorvinStarmast]

Is point buy or the standard array claiming to be equal to dice rolling?

They are both an alternate to the standard means, per the basic rules, of rolling 4d6b3 for 5e stat generation - 3d6 don't enter into it. This is D&D 5e, not Basic, not OD&D.

As to 6d10 b2: what game is that from? We are talking about D&D.

(I had a DM who had roll 2d6 +6, which gave a floor of 8 ... worked well enough)
(I recall from a point in time around release date that Mearls made some observation about the math for SA and rolling being the same, but I can't be arsed to find his inane quip. He ought to know better. Given his consistent habits of issuing careless utterances as one of the two faces of D&D 5e, JCraw being the other, I guess that's just a case of Mike being Mike).

[Morgaln]

A term I use to describe a lot of class systems is "chunky". If I want to play an archer in 2e, I have a few good choices, with fighter, ranger, thief, and bard being the best among them. But if I take one of those, I have to take the rest of it. My fighter-archer has the ability to wear really heavy armor, even if I don't see him that way. My bard-archer can cast spells and sing people happy, even if I just want to be able to shoot stuff. My ranger-archer has to be good and is inexplicably a supreme two-weapon fighter, because those are the size of the chunks.

Kits let you reshape the chunks a bit, but they're still chunks, and they may have their own weird bits... like a lot of archer-specific kits become incompetent at melee fighting.

Chunky game design is fast to create characters; to make a Castles and Crusades character, pick your class, pick your race, roll your stats, and pick one or two primes, and you're done. If the CK has papers with each individual class and race written on them, he can print them and hand them to you. Super easy. But your fighter is going to be like almost every other fighter, except for one or two small choices. Your thief will be like every other thief.

The more options you introduce, the finer control usually is, but that increases time, as well.

I fully agree A class, at least the way D&D does it, is ultimately a bunch of restrictions. It does sell those restrictions as class features but what it does come own to is that you can't do anything your class doesn't allow you to. You want to be able to hit weak spots for extra damage? Nope, you're not a rogue. You want to have special knowledge about deserts? Only Rangers can do that. You want your character to be a history buff? Sorry, not a class skill. It's like everyone can only learn things relating to their job and can't have any knowledge outside of that. As if no one ever learns stuff just because they're interested.
And in some cases it doesn't even make sense. A street kid who makes his living picking pockets for some reason knows how to hit vital organs expertly because she is a rogue, but the ranger who should know how to drop his quarry with a single arrow for some reason doesn't?

[Anonymouswizard]

A term I use to describe a lot of class systems is "chunky". If I want to play an archer in 2e, I have a few good choices, with fighter, ranger, thief, and bard being the best among them. But if I take one of those, I have to take the rest of it. My fighter-archer has the ability to wear really heavy armor, even if I don't see him that way. My bard-archer can cast spells and sing people happy, even if I just want to be able to shoot stuff. My ranger-archer has to be good and is inexplicably a supreme two-weapon fighter, because those are the size of the chunks.

Kits let you reshape the chunks a bit, but they're still chunks, and they may have their own weird bits... like a lot of archer-specific kits become incompetent at melee fighting.

Chunky game design is fast to create characters; to make a Castles and Crusades character, pick your class, pick your race, roll your stats, and pick one or two primes, and you're done. If the CK has papers with each individual class and race written on them, he can print them and hand them to you. Super easy. But your fighter is going to be like almost every other fighter, except for one or two small choices. Your thief will be like every other thief.

The more options you introduce, the finer control usually is, but that increases time, as well.

Honestly, if a game does that I'll like it better. While it's still not my preference I'm fine with BD&D having just seven classes, and even with the race as class model. You're picking an arrow, not a profession. (Plus I believe there's a couple of extra classes poking around, I've got an urge to play a Warrior-elf).

TBH, I love 4e's roles determining much of your powers, and source determining a lot of your aesthetic. While I found their powers system to be way too much (pages and pages of individual powers), that game had a ton of good ideas.

Oh, 4e had many good ideas. It's a shame 5e threw them all away.

[JNAProductions]

They are both an alternate to the standard means, per the basic rules, of rolling 4d6b3 for 5e stat generation - 3d6 don't enter into it. This is D&D 5e, not Basic, not OD&D.

As to 6d10 b2: what game is that from? We are talking about D&D.

(I had a DM who had roll 2d6 +6, which gave a floor of 8 ... worked well enough)
(I recall from a point in time around release date that Mearls made some observation about the math for SA and rolling being the same, but I can't be arsed to find his inane quip. He ought to know better. Given his consistent habits of issuing careless utterances as one of the two faces of D&D 5e, JCraw being the other, I guess that's just a case of Mike being Mike).

Does the PHB ever say that Point Buy is meant to be equal to rolling?

If not, then your expectation that it is is unfounded.

[Pex]

Which is what Standard Array Should Be. I am very much on side with Pex regarding "Having an 18 stat at level one should not be prohibited!"

The problem though is 16 14 13 12 10 9 still denies the 18 to humans and forces variant human to pick a half feat if he wants the 18. However, the problem is not the array but the game refusing to give humans a +2. D&D has a history of treating humans unfairly. It's not all bad with bonus feats and skills since 3E, but I'll never be happy until Point Buy allows an 18 at 1st level AND humans can have that 18 too by default, which is why I like Pathfinder Point Buy, even Pathfinder 2E's method.

[kyoryu]

Oh, 4e had many good ideas. It's a shame 5e threw them all away.

The best part of "roles" is that it gave you a good baseline expectation of what your character would do. And they weren't as cut-and-dried as people thought, as the source also usually impacted where you fit in the roles - martials tended to be more damagey, so a "martial defender" (aka fighter) would do pretty reasonable damage, especially for a defender. But if you picked up a defender, you knew that "outdamaging everyone" wasn't your real goal or purpose, and so if you did want to be a damage dealer, you should probably pick a striker instead.

[Anonymouswizard]

The best part of "roles" is that it gave you a good baseline expectation of what your character would do. And they weren't as cut-and-dried as people thought, as the source also usually impacted where you fit in the roles - martials tended to be more damagey, so a "martial defender" (aka fighter) would do pretty reasonable damage, especially for a defender. But if you picked up a defender, you knew that "outdamaging everyone" wasn't your real goal or purpose, and so if you did want to be a damage dealer, you should probably pick a striker instead.

Yeah, 4e was really good at describing what roles were supposed to do and then creating classes to those roles (with some exceptions). It caused some courses to use things they used to have, but it turned to give them a stronger identity in return. While I'd rather pay 13th Age than 4e one of the things I'm glad seems to have carried I is building classes to a role (if less explicitly).

If I'm going to have to select a cheap I want to know what that class is supposed to do. Both 4e and BD&D achieved this, but in completely different ways. Most of the issues with class systems seem to come along with the idea of having more than one.