This is a method of character design and mechanical resolution that is abstracted away from player awareness and left entirely in the GM's hands. It is primarily intended for use in a play-by-post format or other asynchronous types of play. It accentuates narrative play publicly, but still relies on the robust Hero System game engine for resolution privately, handled out-of-band by the GM.
Have the players describe their characters via numerically expressed Traits, adding up to some GM-determined number of total Trait Points. A good rule of thumb is to divide the total Character Points of the characters in the campaign by 10 to determine the Trait total; thus, if the GM wanted to run with 150-point Hero System characters, they could tell the players they have 15 Trait Points to spread around (an extended example is provided below to illustrate this in practice).
However, this is just a numbers game, and any ratio or scheme can work as long as the GM is consistent.
The GM then stats the players' concepts using the broad descriptions provided as their guide. The GM may or may not allow the players to see the final write-ups, as they prefer.
It is recommended that the GM keeps the Hero System write-ups private (at least until the campaign is over) to avoid arguments and encourage the narrative aspect of this method. In fact, a GM might choose to never reveal that they used the Hero System under the hood to help them run the game, as it is largely irrelevant to the players. But ultimately, it is up to the GM whether to reveal their Hero System wizardry or keep such sordid details behind the curtain.
During play, the GM and players narratively describe their characters' intentions for the next Turn and work out any necessary clarification. Then, after the appropriate window of time has passed, the GM accumulates all the declared intents and resolves the outcomes off-screen using the Hero System mechanics. They synthesize/summarize that resolution into the next chunk of narrative text (eliding the intermediary mechanical cruft) for the next GM post to move the adventure along. Rinse and repeat as necessary to completion.
After play starts, rather than assigning XP to players, instead have the players describe what areas they would like to see their characters show improvement in, which the GM interprets into the characters' mechanical implementations.
In this example, Bob the GM is about to run a standard 150 Character Point (75+75) Fantasy Hero campaign via play-by-post and has decided to try this format. Bob uses the recommended "divide by 10" method to determine Trait Points, and thus tells each prospective player for his campaign that they have 15 Trait Points with which to define their characters via labeled Traits, and each Trait must include a concise description.
Bob doesn't want the PCs to be entirely one-trick, so he further requires that each PC must have at least three Traits, and no Trait can start with a rating higher than 6 (i.e., approximately 40% of a character's total capability). Bob also doesn't want the PCs to be spread too thin, so he further states that starting characters can't have more than ten Traits.
Bob's setting has a handful of playable fantasy races, and under the hood, these map to Race Packages. Bob indicates to potential players which races are allowed in his setting, and that some require allocation of Trait Points to opt into.
For instance, Bob's Elf, Dwarf, and Halfling Packages all cost approximately 20 Character Points, so he indicates they each require 2 Trait Points for a character who wants to be of that race. Bob's setting also features an exotic Giant-kin race with a more expensive Race Package costing approximately 40 Character Points, so Bob indicates that it requires 4 Trait Points for a character to be of that race. And so on.
All of these decisions are arbitrary and campaign-specific; for a different campaign, Bob might make different choices.
Four players respond to the campaign announcement: Joe, Fred, Mike, and Amy. They defined their characters as follows:
Joe is playing the wizard Andronicus, who Joe wants to be a brainy but capable gentleman adventurer.
When designing this character, Bob might spend approximately 50 points on things that make the character "smart" and knowledgeable, approximately 60 points on his magic, approximately 20 points on agility (DEX, SPD, Run, etc.), approximately 10 points on being tougher than average (which could be CON, BODY, STUN, PD, ED, etc.), and finally approximately 10 points on social connections.
Fred is playing Uaga the Bold, a barbarian which he defines as follows:
When designing this character, Bob spends approximately 60 points on making Uaga a beefcake, approximately 60 points on making him a killer, approximately 20 points on skills like survival, tracking, stealth, shadowing, PS: Hunting, etc., and approximately 10 points on looking good and Seduction.
Mike is playing Gibs Lightfinger, a halfling pickpocket and general thief. You've defined a Halfling package that costs around 20 points; thus, to be a halfling costs 2 Trait Points. Gibs is defined as:
At this point, it is hopefully pretty clear that Bob would translate this into approximately 40 points in thief-y abilities, approximately 30 points in stealthy abilities, approximately 20 points in interaction and PRE, and so forth.
Finally, Amy is playing Vailia the Pure, Cleric of Dionicus the White. She defines Vailia as so:
And... it is hopefully evident at this point how Bob might go about spending points to make this character.