Problem-Solving Combat: Breaking the Cage of "Roll Initiative"

I don't want to take the stance that this is an objectively better way to do things, but there's something in the water (maybe hydrogen, maybe oxygen?) that makes humans write toward that stance. I could go edit through everything and make sure I'm not, but that is a recipe for not actually publishing any blog posts. So take this as a blanket disclaimer: This is what I've found works for me, to evoke the kind of engagement I want with and from my players, compared to how I've seen and played other game systems run by other GMs. If it sounds good to you too, cool, see how you can integrate it into your games, and look forward to the indeterminate future when I release some public form of Cunning Knave.

The Cage of "Roll Initiative"

D&D-like combat rules are a black hole for what I find fun about OSR style gameplay. 

Yes, we all plan ambushes when we can, and try to out-think enemies.

But when we do Roll Initiative, a cage comes down over each player, trapping you in a very well-defined set of options - to the death.

Having default types of actions within combat discourages creativity. And If a player spends too much time thinking of a clever plan (which is harder than it should be since it's restricted within the framework of the combat rules), they're pressured into taking one of the default actions.

Strict turn order discourages teamwork. If one PC calls for the retreat, the enemies will likely act before the rest of the PCs can, which in the majority of cases means there's a good chance that one or more won't be able to get away. This fact pre-emptively discourages fleeing, along with plans involving more than one PC, across the board.

Yes, there are things like 5E's Help action, but that's almost worse - it converts a good idea (how are you helping?) into a mere modifier for a future action someone else is able to take, instead of, like, actually doing something interesting.

"But making combat deadly and standardized - even boring - helps push players to avoid it altogether, a central tenet of OSR play! And it's easier to ensure it's deadly by making rules that ensure it's deadly." This proposition isn't incompatible with combat being a Bad Idea. It's just that when we inevitably do end up in combat, I don't want to leave the "OSR style challenge" mindspace and shift into a Roll Off.

Of course all of this probably sounds familiar to anyone who groks Powered by the Apocalypse games. It's something I've always found that PbtA and OSR games have in common; an emphasis on presenting qualitative, rather than quantitative problems, in an unrestricted solution space. This is certainly true of OSR gameplay in general, but when it comes to combat, the vast majority of OSR rulesets seem to hew very closely to the wargame-rooted systems of the original games.

Wat do?

I think, if nothing else, Problem-Solving Combat can be distilled into a handful of guiding principles that can apply to most any RPG, and that's probably all you need.

  • Don't Roll Initiative - Just describe the situation and ask "What do you do?"
  • Don't Enforce an Order - Let everyone discuss a plan, team up on actions or establish a group tactic, then call for the rolls that make sense, in an order that makes sense.
  • Don't Just Attack - Use enemies to do interesting things that threaten the PCs.
  • Don't Stay Still - Present new challenges as every round evolves the situation.
  • Don't Stay Abstract - Concrete details are opportunities to seize.

It's tempting to say "well, if you want to treat combat the same as you do non-combat stuff, just don't distinguish it. Use the rules from non-combat stuff for combat stuff." I think, at least for me, this is a trap. Having a discrete mode signals to the players that their lives are at stake. And it is still nice to have some kind of framework for arbitration, rather than leaving it completely freeform. A procedure, for guidance, something to keep hold of as you navigate the dizzying tactical infinity.

On the other end of the scale, while permitting an open solution space, even the Move systems of PbtA games is too rigid a framework for my taste (for OSR style gameplay).

There are also systems that present combat as an open-ended challenge, but reduce them to a single roll. But Combat is such a rich source of interesting problems to solve that this seems like a loss.

So, like I do, I wrote a hack. Here is an excerpt of the WIP, which I'm calling Cunning Knave.

Combat in Cunning Knave

[For context, "Resolve" is essentially both Level and HP (Monsters' Resolve is their HD), and PCs begin with 1 Resolve. You can assume the system is otherwise very similar to Knave, for current purposes.]

If violence has become inevitable, a round of Combat begins.

The referee will describe the situation and what the enemies appear to be doing. The players are then free to discuss and plan any course of action, with the referee clarifying feasibility and likely consequences of failure, but not exact difficulty or potential complications.

Once the plan is clear, the referee will then call for any Rolls that are necessary for each PC’s actions in an order makes sense, applying advantage or disadvantage where relevant. Players may clarify details before rolling to attempt to gain advantage or negate disadvantage.

Just like outside of combat, very good plans or straightforward actions don’t require Rolls. And any intent can be attempted; disarming or capturing, driving off by intimidation, de-escalating to parley, retreating, and any approach can be utilized, not just attacking with weapons but utilizing other items or aspects of the environment, perhaps gaining advantage.

Once all the effects of the current course of action have been resolved, a new round begins; the referee will describe how the situation has changed, and the players may discuss their next actions.

Each round of combat should result in significant changes to the situation; combat is chaotic and quick and should not last more than a few rounds, often resulting in surrender or retreat after two or three. Try to avoid a straightforward exchange of blows, instead presenting a challenge unique to the type of enemy, their equipment, and the environment that might not be resolved by simply attacking or defending.


A physical attack requires an Opposed Roll against their Armor Target. A success reduces the enemy’s Resolve, usually by 1, but potentially more with particularly effective plans. An enemy at 0 Resolve is at the PC’s mercy, possibly dying.


Failing a Roll to attack, or taking particular risks in a hostile encounter may result in the PC being attacked. If a PC is attacked, they make an Armor Roll. A failure causes damage, depleting their current Resolve, usually by 1.

Traps and other types of incidental harm may also be considered an attack, and in certain dire circumstances, a PC may take damage directly with no chance to resist with a roll.

Attacks are usually physical, but could be mental or even emotional. These would trigger an Ability Roll other than Armor, such as Charisma or even Resolve itself.

At 0 Resolve, the PC is out of action until they can be safely attended to.


Wicked foes, traps, or other environmental hazards may outright kill a PC while they are out of action or otherwise helpless, if their companions cannot prevent it. If there is some chance that they could avoid certain death, the referee may call for a Resolve Roll to survive. If this is successful, the PC survives, but also takes a Dire Wound, which fills an Item Slot and reduces Resolve by 1, permanently.

Attacking foes directly is very dangerous. Look for safer ways to exploit the particulars of your enemies, equipment, and environment. 

I'll be following this post up with a concrete example of combat from a playtest with further commentary, and possibly a third with some more procedural advice on how to turn combat into a series of interesting challenges.

Related Resources

As with a lot of RPG blogging in this day and age, none of this is revelatory or original; I've just felt compelled to assemble a particular combination and reinterpretation of stuff I've ingested over the years. Among those influences:

I encourage you to comment below, rather than elsewhere. I even welcome "But this one system here does this one thing like you want!"

Long live the Blogosphere!


  1. I urged the death of initiative here:

    I've been playing with "we go" (as opposed to you go I go) aka phased real time combat for years. It is strictly superior to you go, I go, to a degree where I find it surprising traditional initiative is still so popular.

    Glad to see other folks adopting it!

    1. Nice, I'm sure I have come across that post in the past and it influenced me as well. I'll be sure to point folks to it if they find the method in Cunning Knave too loosey goosey.

  2. Maze Knights was already leaning in this direction, so I'm very interested to see how your rules shape up. Maze Knights makes dungeon crawling squad based, where each player controls 2-4 characters that each have one unique ability. Going one at a time is prohibitively slow, so I'm planning on allowing the players to simply plan out synergistic combo moves and then execute them all at once, rolling checks as necessary.

    1. Nice, that sounds cool. So looking forward to Maze Knights!

  3. An inherent problem is that combat is often not open-ended. This means that with fair refereeing, players will be heavily incentivised to take the action that brings them closest to the winning condition (typically enemy fleeing or dead). Combat is therefore a quantitative problem, by definition.

    Example: say that you have a 1/10 chance of killing the opponent with an attack, and a 1/3 chance of driving it off with a clever plan. If your group has 3 ppl or less, they should go for the plan, otherwise attack.

    This is also true for the rules you suggest: unless the group can think of a plan that will net more than one damage per character, within a reasonable amount of time, their best bet is still attacking. And if there is a plan, unless it is unique to that situation, the players should always default to the same plan until they can think of a more effective, and then default to that, etc. Meaning that "we do plan A" is equivalent to "we attack". Unless... you have varying initiative order, preventing the players from relying on the same plan each time.

    Personally, I do group initiative. But I'm looking forward to your play report.

    1. It sounds like we have different playstyle assumptions, or I haven't sufficiently explained myself. Definitely check out the play report when it goes up.

    2. This is quite a late reply, but I wonder whether part of it would be to shift thinking about combat as a succeed/fail state where there ARE open-ended outcomes. Open-ended outcomes make non-combat challenges interesting. Maybe this is an argument for more GMs to consistently bring in dilemmas and alternative goals in combat: defeat the goblins AND stop the wizard from fleeing. kill the frog-demon AND don't let it eat too many hirelings.

    3. Good point and interesting thought!

  4. Seems odd that you omit group initiative as an option, given how well-traveled that territory is. I think that the B/X initiative rules solve many of the problems you mention here.

    1. I'm super tempted to answer in a sarcastic manner, which means I probably shouldn't reply at all. So I'll just say, please refer to the first paragraph of this post.

    2. Of course it's how you do things. You're the one writing the blog post. I don't believe I missed that point.

      It's just that you use the term OSR so often in this post, I kind of thought some consideration for how many OSR games handle initiative would crop up somewhere. Anyway, clearly not the day for a debate.

    3. Yeah this wasn't a blog post about initiative rules? :shrug:

      You bring up a good point about using the term OSR though, I'm not following my own advice from
      I might go back and replace some references here with more accurate terms if I have the energy.

    4. I can see how that assumption could be made though, given the title of the post! Blargh. This is the kind of thing that makes me spend months on a blog post and never posting it, though, so.

  5. In the system I’m trying to use I don’t use armor rolls, I used opposed 2d6 rolls. If you fail your attack, essentially the counter attack succeeded. This resulted in a few other changes: armor is now extra hit points since it’s not needed for the vs roll, for example. A thing that I’ve been wondering, however, is spell casting. Cast a spell, roll vs, fail, and you take damage. In your case, it seems similar: the wizard who fails in their attack gets counter attacked and has to do the armor roll. Are you OK with that, or do you treat spells differently? Maybe not having the kind of D&D attack spells would also solve the problem, of course.

    1. If the wizard is unwisely on the front lines at the time, sure! All of these details should be adjudicated in a fiction-first manner though. If I were that wizard, I'd be sure I was in a fairly defensible position first.

      That said, yeah - the magic system I'm using in Cunning Knave is similar to Knave in that most spells are not "attack spells".

    2. Yeah, must remember that, too. Less looking at dice and more fiction-first. Interesting blog post in any case. Thanks! 😄

  6. A couple things for anyone keeping track here:
    1. The followup post with an example of play is up Hopefully it clarifies a few confusions that this post creates due to lack of context.
    2. Macchiato Monsters seems to be the closest to this style of a combat system that I've seen; in particular, its Running Combat in 6 Steps section contains most of the essence of what I'm trying to elicit here. The full system still has a few too many restrictive assumptions for my taste; Cunning Knave will have more inherent freedom in its procedure.

  7. Sorry for the necromancy (my specialty btw), how is your experience with this style of combat? Anything that is weird, dont work as well as intended, etc? I see a lot of people talking about this style, but very few actual playtest experience with pros, cons and weird about it.

    What does it does well? What does it does wrong?

    OSR classical style say that the GM is a referee narrating a unforgiving world with clear procedures. You have to know and survive the procedures, you live in a "real" world with it´s own laws. These procedures try to minimize the arbitrarily of the GM role, so he can be judgement free of character death.

    Personally, i like this no-initiative style, but i'm having a hard time consolidating it with the "imparcial referee" role of the GM.

    1. I go into some of that in

      To be honest, I have not experimented much more beyond that, as I have not been running games during most of the pandemic. If I start running again, I definitely want to feel out this method more.

      I like the concept of a rigidly proceduralized game style, but I don't think it has to be the only OSR-influenced GM style. This style takes another core aspect of the OSR, problem-solving, and further extends it. Even though these problems created mid-combat may be "arbitrary" relative to the players, they can still be somewhat objective challenges which follow from, and must be overcome within, the constraints of the established fiction.

      It does place a heavier burden on the mind of the GM, and I hope to provide some tools to aid this (perhaps procedures, to bring that aspect back in more).

      When it comes to death, Cunning Knave is still quite lethal with its tighter HP rules. At level 1 it comes down to judging whether a good blow against a character could mean death. If so, they die. If a bit of leeway is desired, the GM might allow a new challenge for them to recover.


Be kind, approach disagreements with curiosity, enrich the conversation!