The difference between a stat of 18 and a 20 is 5% or a single modifier point. In most cases, it's realistic to eventually boost a character's primary and secondary stats to three 18s or 20/20/14, possibly including a feat or two along the way.

Which do you all think is the more favorable stat array for which builds?

EDIT: Someone below mentioned an array of 20/18/16. This is quite fine, too. Between the standard array and point buy, one could easily start with 16/16/14 or 16/16/13, use 4 ASIs on stats and the 5th on a feat to get 20/20/14, 20/18/16 or 18/18/18.

It's probably not optimal, but a pure fighter or either 19:1 or 16:4 fighter multiclass can get 20/20/20 starting with 16/16/14 and only taking stat ASIs. This is pretty nice with a dip into barbarian or monk for Unarmored Defense.

I would personally say that 3 18s would be better because…

1) Maxxed stats don’t come until late-game and at that stage saves mean WAY more than a 5% hit chance, especially because only the biggest baddies have crazy AC and usually you have to go through a lot more low AC targets before you get there

2) Ideally, if they are in the most-used Saves of Wis, Dex, and Con.

3) For MAD classes, definitely 18/18/18.  I’d rather have a Paladin’s aura give 2 more Save bonus to my allies than ensure the Paladin hits a small percentage more.

But objectively?  I don’t think anyone can do enough analysis to accurately determine what would happen late game because every campaign is different.  Let’s just say no one will say definitively which is best. 

What metric are you using for "favorable"?

In a solo adventure, it is helpful to be well rounded, but in a party, it is important for everyone to be good at their own thing. 5% may not seem like much, but being redundant with another PC can be worse than nothing.

Quote from Memnosyne >>

What metric are you using for "favorable"?

In a solo adventure, it is helpful to be well rounded, but in a party, it is important for everyone to be good at their own thing. 5% may not seem like much, but being redundant with another PC can be worse than nothing.

 When I say "favorable," I treat it as an experimental ideal. Lots of white-room scenarios I've seen on the forums lean towards the two 20s setup for damage:hit/save% ratios, but a paladin with that setup sacrifices something important, whether it's staying power, team support or the probability to land hits; same goes for monks and rangers. Bladesinging wizards who spike Dex and Int go down pretty fast when whittled down with saves, despite having a very high AC.

One of the nice things about 5e is that pretty much any party makeup can work, but it's not uncommon in the games I play in to see a barbarian, paladin and plate fighter all meet up at the proverbial tavern.

I feel the best case for the two 20s goes to medium armor casters. Leaving the 14 in Dex allows for a solid AC and allows the rest of the ASIs to go into the casting stat, Con and possibly a feat. Clerics and druids are well-kitted here, as this encourages higher values in the big three saves.

For me it would depend a lot on the character.

Generally I would want 20 in the primary stat

Most Rogues only really need Dex although con and (to a lesser extent) wisdom is nice. 20 Dex would be a must though probably go 20/18/16 rather than 20/20/14.

Simlarly most pure spell casters you want a 20 in the casting ability with a decent  Con and Dex, again I would prefer 20/18/16 but would prefer 20/20/14 to 18/18/18

20/20/14 would be optimal for a clear two ability build, fighter is the most obvious one but would also be the case for medium armor barb

I would only really consider 18/18/18 for the MAD classes, Monks, Paladins, unarmored Barbs, melee clerics, and rangers but even then I would often prefer 20/18/16 or sacrifice feats to more ASI, the monk I took to level 20 ended up 20 Dex, 20 Wis 16 Con.

I'm on board with the 20/20/14 route for my Monk.  Having a +5 to hit, a better DC on all my abilities that call for saves, max unarmored defense number and they come into play every combat, every roll, every action.  Adding 1 to my chances for every roll I make pretty much is a big thing.  Especially when my option is to further boost Con to be a little beefier, and as a Monk, IF I get hit, at high end it is going to hurt....a LOT and adding a handful of hit points aren't going to make a big difference.  As others have said, I think it also depends a fair bit on class, because there likely are several who would net better overall benefit from the 18 split.  My experience with my Monk has been that every point has made a difference.  Maybe my dice gods are more fickle and have us MEET AC on a hit or save often, so we can see there, the 1 point made the difference.

Quote from Falwith >>

I'm on board with the 20/20/14 route for my Monk.  Having a +5 to hit, a better DC on all my abilities that call for saves, max unarmored defense number and they come into play every combat, every roll, every action.  Adding 1 to my chances for every roll I make pretty much is a big thing.  Especially when my option is to further boost Con to be a little beefier, and as a Monk, IF I get hit, at high end it is going to hurt....a LOT and adding a handful of hit points aren't going to make a big difference.  As others have said, I think it also depends a fair bit on class, because there likely are several who would net better overall benefit from the 18 split.  My experience with my Monk has been that every point has made a difference.  Maybe my dice gods are more fickle and have us MEET AC on a hit or save often, so we can see there, the 1 point made the difference.

 Agreed, it would be a very rare build that getting the tertiary ability to 18 beat getting the primary to 20. At level 19 I took Tough as when you'd  not have spells and are immune to poison con offers little more than hp and tough gives twice as many.

I personally think that maxing out a stat is really only the ideal choice if you're playing a class where a resource of some kind is tied to that specific stats. For a wizard, a higher INT isn't just a boon to your spellcasting ability, but also determines how many spells you can prepare in a day. If you can get an INT of 20 early in the game, it opens up a ton of avenues for you. A bard only gets as many uses of Bardic Inspiration as their CHA modifier, which is a huge factor in their playstyle, especially for subclasses that give you alternative features to use with your bardic inspiration.

Personally, I'm glad that more features going forward have been keyed off of Proficiency bonus instead. It allows a steady progression in valuable resources without kind of "punishing" a player for not maxing out that specific stat. It allows more freedom to focus on well-rounded characters instead of feeling forced to juice one stat to the max as soon as possible.

I would say it depends on the class. Classes that depend on more abilities scores would benefit more from 18/18/18. Example:

Paladin: 18 STR, 18 CHA, 18 CON would be better

Hexblade: 20 CHA, 20 CON would be better

Quote from SeanJP >>

I would say it depends on the class. Classes that depend on more abilities scores would benefit more from 18/18/18. Example:

Paladin: 18 STR, 18 CHA, 18 CON would be better

Hexblade: 20 CHA, 20 CON would be better

 A Paladin depends very much on the role and concept but unless you are casting a lot of concentration spells I would put Con as the tertiary stat. If the paladin is also the face I would probably go 20/20/14, if charisma is being used mainly for aura's and spell slots for smites then I would probably go 20/18/16. 

Quote from Jegpeg >>

Quote from SeanJP >>

I would say it depends on the class. Classes that depend on more abilities scores would benefit more from 18/18/18. Example:

Paladin: 18 STR, 18 CHA, 18 CON would be better

Hexblade: 20 CHA, 20 CON would be better

 A Paladin depends very much on the role and concept but unless you are casting a lot of concentration spells I would put Con as the tertiary stat. If the paladin is also the face I would probably go 20/20/14, if charisma is being used mainly for aura's and spell slots for smites then I would probably go 20/18/16. 

 There are number of useful uses for CHA, some depending on subclass, not the least of which is the Aura.

If 20/18/16 is on the table that's something to think about. But I was under the impression from the OP that it would be 20/20/14 or 18/18/18 with mid-low remaining stats. Honestly I would 20/14/20 before I went 20/20/14.

But I'm sure either of us would be very happy with any of these :)

The more MAD your concept the less likely you are to have to go the 18/18/18 route simply to get anything. Realistically I have a sort of list of MAD concepts tucked away and typically roll ( no stat under 10) and when I get luck with really good rolls I take the next MAD character and use that set of rolls for the character. The next set of rolls can go for whatever I’m  going for initially that’s not so MAD.

From what I see in the starting comment, I don't really thin you need much help with coming up with ideas???