Jane rolls 8, 11, 14, 10, 17, and 15.
Jill rolls 13, 10, 13, 13, 13, and 10.
Joe rolls 11, 11, 13, 8, 12, and 10.
Jason rolls 11, 17, 16, 11, 13, and 15.
Everyone can then pick which of the 4 sets they want to use. Everyone picks Jason because yowzers. The power level of the characters is certainly higher than normal, but the DM can easily adjust for that. You got your RNG, everyone got to pick the stats they wanted, and someone at the table got to be a big dang hero!
(Those are actual rolls I just randomly generated.)
The problem with this approach is you get higher ability scores than 4d6, drop lowest, already generates.
Now, if you want that, then it
isn't a problem of course.
FWIW, the "standard array" for 4d6, drop lowest is actually 16, 14, 13, 12, 10, 9.
If you have a "standard" party of 4 players, each rolling, the "highest (standard) array" is instead: 16, 15, 14, 13, 11, 9.
Comparing to the standard array (15, 14, 13, 12, 10, 8), we see 4d6, drop lowest increases the highest and lowest only (15 to 16, 8 to 9) and the middle scores remain the same. Having four players all roll 4d6, drop lowest and then take the
best set among them
also increases all the middle scores (14 => 15, 13 => 14, ... 10 => 11). The table below shows them in comparison:
Of course, if you have more players, the rolls will generally be higher. Also, if one player happens to get really lucky (like a player in my first d20 game, he rolled 16, 18, 17, 14, 13, 13), then your whole party will be very powerful...
So, nothing wrong with such methods, just food for thought.
