I'm gonna soon DM my first game in 10 years. I have zero experience in DMing in D&D 5e. At each session, four players will seat around the table.
My future players already said and repeated they disliked the standard array because it offers no way to get a stat at 20 at level 4, if they take the first ASI (and thus no feat). And they like to roll. In every other game I played with them, we rolled, and we like that.
On the other side, I'm looking for fairness among the group. I don't want that one player has 2 scores with 18 and another player has only a 14 as their best value.
We never minded being a tad OP as we're mostly in for the story, not for the combat, even though we like it. Also, given that what my players will see are their own stats, potentially going to 20 at level 4, I can use this as a a psychological factor and throw encounters that are just a tad stronger at them as well. They'll be happy with their stats, but still get similar challenges.
I've read on the internet that some DMs take the standard array and add 2 to each score. Now this feel really, really OP, so I've thought about adding only 1 instead. This has the advantage of being able to reach 20 at level 4. However, my players will then not roll. I don't really know how OP this is. In average, it gives +3 mod.
I homebrewed a bit myself and thought why not let them all roll their own array (4d6, keep 3), then I let them decide which of the rolled arrays they take, but they all take the same array, and they must all agree. There is no fallback in case they dislike each array. This has the advantage of letting them roll, I'm satisfied as everybody has the same base scores and on top of that, I'm potentially putting them in front of a difficult choice: take this min-maxed-like 17, 14, 11, x, x, x array or that MAD-friendly 15, 15, 14, 12, x, x array. They can also be thankful for the one player who rolled 18, 17, x, x, x, x, or be all sorry because none of them did a great roll. Since the players choose, they're responsible for it.
I didn't forget the point-buy, I'd allow 33 points (instead of 27) to buy and they can go up to 16, but buying from 15 to 16 costs 3 points. This feels like it's below the stronger array in the first option, for instance. To a min-maxer, it gives +2 stat mods in crucial stats, but to a thoughtful player, it potentially gives +4 stat mod. Pros: 16 is a guaranteed in. Cons: no rolls.
Now, all those methods really feel like my players will be above the average campaigns players (which, in turn, are above the average guy in the in-universe world). I know that. But I would like to know in what measure.
Note: I'm not asking for advice about those alternatives (though they're welcome), but to evaluate -- or tell me how to evaluate -- how OP any of these options will be.