My strength is as a strength of 10 because I am average

[whswhs]

I think I’m going to disagree with Roger’s suggestion that assigning a specified series of numbers, such as 13, 12, 10, 9, and 8, to a set of characteristics is one step away from straight point build on a scale of randomness versus design. It seems to me that both methods are completely free of randomness. The former method is more constrained, but it’s not constrained by randomness, but by a restrictive rule.

That said, of my three considerations,

• The average set of rolls produces characters with a mean of 63 total points, which may not be heroic enough for a particular genre.
• Even if the average roll is good enough for the genre, you may have gotten bad rolls and have only 49 total points, and your character won’t be any use.
• Even if your character has a suitable total point value, the distribution may not support the character concept you want (a gadgeteer with Intelligence 5, or a brawler with Strength 7 and Dexterity 8).

I’m going to say that I’m satisfied to address the second by giving players freedom to choose between random rolls and some form of pure design. For the first, I think that “roll six times and discard one roll” gives enough of a boost in the random approach, and that raising the point total for point build correspondingly keeps things balanced. I think that I don’t want to do a specified series of numbers, both because it may block a character concept and because it gives too little variety of characters; instead I want to provide a point total that can be spent fairly freely.

So I’m going with either

(i) roll 3d6 six times, and then assign five of the rolls to the characteristics;

or

(ii) set each characteristic to 8, and then allocate 15 points to raising desired characteristics.

That seems to me, on one hand, to make superheroes who gained their powers randomly reasonably close to ordinary human beings, but on the other to make them a bit better than average, which both makes them viable as adventuers and fits the romantic convention of the genre. And it leaves room for an exceptionally talented and/or trained normal human being to buy up characteristics with “Heightened (Characteristic) A,” or for a scientific genius to take “Heightened Intelligence A” specifically. So I think it will work for the concept I’m playing with.

All the discussion here is greatly appreciated; it really helped me think through what I was trying to accomplish and what methods were available.

[Agemegos]

@whswhs

The nonrandom approach is certainly better if you want to play a character who fits a concept. The random approach is more of a challenge: “Here’s a role, let’s see what you can make of it!”

I suppose you’re right, but even so, rolling a number of character stats independently does not efficiently or reliably assign or even allow you a role. If you wanted to do that it would be much better to roll your character class or professional template, or deal each player a class or role from a special deck of cards.

The random rolls approach works better if you will make multiple random rolls and they have a chance to average out, which isn’t usually what we’re doing in an rpg, where you have some measure of long-term commitment to playing your character.

The wise and handsome Professor @RogerBW has told us many times that RPGs need plots and characters that are suitable to their peculiar form, and cannot be relied upon to do well with plots and protagonists copied from even the great successes and stalwart principles of other media. The same thing is true of game mechanics. Fundamental properties that make chess and Go, whist and poker, gin and dominos, hazard and knucklebones, darts and shuffleboard, charades and spin-the-bottle immortally successful game designs and the forebears of numerous descendants are not necessarily translatable to RPGs.