The problem with most tabletop systems is that they're designed to play out games in real time and in person. Doing one attack per day (and missing!) will drag down the pace unbearably. But usually you want some kind of conflict resolution.
So I propose a minimal mechanics system which allows conflict resolution, character building, and yet minimum mandatory mechanics based on the posting.
The idea here is that the players will just roll for conflict resolution once to decide the ending of the battle or whatever. And then they write up their freeform posts in a way to build up towards that ending. The results are abstract enough - so if you're doing a bar brawl, you can still play out kicking, shield bashing, throwing mugs at the enemy, and both getting non-critical cuts to each other. The system just decides who gets the killing blow. It doesn't even have to be fatal, it can be a knockout blow.
Threat represents the ability of the character to do critical, life-threatening damage. Things like bruises, cuts, and non fatal might not count.
Winning a threat roll means that the character has landed a significant blow on the enemy.
Weapons and armor increase threat, not damage in any form. Weapons might increase threat level or negate the effects of armor. Armor might act as a multiplier, so someone with Threat 3 in heavy armor might have Threat 6 instead.
Example of use:
A lady knight and a mage meet in combat. The knight has Threat level 3 and the mage has Threat level 5. The knight has some rather heavy armor, but the mage's magic negates it. The chance for the knight to win is 3/(5+3) = 37.5%
The attacker makes a roll, which decides who is the winner of the battle. Then they freeform it out to get to the conclusion of the battle.
Both of them do their combat. The knight thrusts her sword at the mage, only to have it melted down by his magic. The mage then melts the rest of her armor away and tosses some fireballs. Perhaps the knight might have successfully dodged those fireballs and kicked him in the face if she won. Or perhaps she fails to dodge them if she fails the roll. It's up to the two players to decide. And the game gives more flexibility for what they want to play out rather than regular D&D.
Represents number of critical blows character can take. This includes hero luck like a bullet to the heart being softened by a book, or taking an arrow to the knee.
Most characters have toughness of 1.
Heroes (PCs/villains) might have a toughness of 2.
Legendary characters might have 3 or more.
Every point of toughness requires an extra roll to knock off. Because of this, I don't recommend putting in too many toughness points because they'll slow the game down drastically. They also have a very strong impact on the balance of the game which I'll go into later.
It doesn't necessarily mean that physically stronger characters have higher toughness; this is more of a heroic luck check. But a big tough ogre who can take a spear through the gut might have a higher toughness rating.
Take the same example as above, a knight of Threat 3 and mage of Threat 5. However, the knight is a heroine and has two toughness points. The mage casts a powerful fireball near her, which she barely dodges, burning the rest of her clothes off. Most other mooks might falter here, but being the hero that she is, she keeps fighting.
They do a second roll where the knight has 37.5% odds. If she fails, she might still be knocked out from his various magicks. Otherwise, she might have successfully knocked out the mage despite not having weapons or armor.
Statistically, because the knight had two rolls, her odds of winning the entire encounter goes up to 61%. But if she took that first toughness hit, it weakens her for future encounters or present encounters with multiple enemies.
Secret attack or ambush first, if applicable. Rogues shine here. The ambusher doesn't lose any toughness even if they fail.
Otherwise do a skirmish roll, if applicable. Will cover this later.
Conflict. Players show threat level here and can make decision to flee or cancel. This represents duels to first blood or light sparring. If both sides agree to have a full conflict, they can do a roll.
Make a roll to knock down a point of toughness each time.
The conflict ends when one side runs out of toughness points.
Both sides roll their skirmish threats against another character. If they succeed, they knock out a toughness point. It's possible for both to shoot the other critically at the same time.
A skirmish uses different threat levels. A ranger or trained archer might have a higher skirmish threat level.
Penalize threat levels if the skirmishers have less than 3 shots, and perhaps give them a bonus or extra rounds of skirmishing if there are more than 10 shots.
Take note of armor piercing as many ranged weapons easily go through chain mail.
If the game takes place in a more modern setting where firearms are common, just use the normal conflict resolution, because you don't want to keep rolling misses.
Add up threat levels on both sides. Mages might need a threat level increase for their armor piercing. A good general or leader, or team discipline, might also increase the threat levels.
Each person uses their own toughness roll. So a first round might knock out the mooks on one side and leave the heroes standing.
Alternatively, the rolls can be made one-on-one, if the situation of the battle allows it.
Another alternative for large scale battles is wiping out soldiers on one end based on the odds. So let's say one side has 200 threat and the other side has 300. The ones with 200 threat would have 60% of their army wiped out and the one with 300 threat would have 40% of their army wiped out.
Of course, this isn't a very realistic resolution as many generals have won despite overwhelming odds.
Of course, this doesn't just have to be applied for violence. It can apply to common themes here like seducing someone highly reluctant, or intimidating/coercing them. It can be used for opposed rolls where the results are binary, like pickpocketing or gathering information.
I haven't actually been able to test and get the right numbers for things, but maybe people can bring up any issues with the proposed system first before I put together something.