Bill, do we have a board for discussions of AI?
Here are the rules and my prompt and the AI response:
RESOLVING HAND-TO-HAND COMBAT
WHAT WE ARE REALLY DOING
Skirmish games with miniatures are not chess; they are far more complex! Chess players move one piece at a time in alternation, but a real-life skirmish involving only a dozen fighters has each moving simultaneously and often independently of any orders that higher-ups might give. When a close-action melee occurs on a tabletop, the miniatures may be fixed in place upon the table, but represent a continuous swirl of movement: feints and real attacks, pivots and advances and retreats, coordination and unplanned and independent actions. So while we here impose some mathematical probabilities upon what happens, players must understand that what we are really doing is TELLING A STORY upon a framework of a few dice rolls.
These rules are not for shooting but for combat with knives and teeth and lances and hooves: sharp points and moving mass. Fighters armed with missile weapons may well prefer to reload and shoot even at very close range, but if they choose or are forced to grapple this is how we resolve it.
THE BASIC MECHANISM:
Each fighter is assigned a die level ranging from d6 (a six-sided die) up through d8 and d10 and d12 to d20 for mounted fighters. The basic mechanism is simple: each fighter rolls his die:
if the numbers are the same, nothing happens and the fight continues;
if one die's number is higher but not double the other, an advantage is gained but the fight continues;
if one die twice the other, the losing fighter is killed;
and if one die is three times the other, the losing fighter is spectacularly killed in such a way as to affect morale of any who witness it.
For example, an untrained female civilian (d6) is facing a novice warrior (d8). The woman rolls a 5 and the warrior a 7. Something bad has happened to the civilian but the fight will continue next turn. Perhaps she has dropped her weapon, or slipped and fallen; whatever the specifics, she is now vulnerable. Next turn the two fighters roll again, but now the woman must subtract -1 from her roll and is defeated (killed) if the warrior beats her roll even by 1. Note however that if the warrior were to roll a 1 himself and she a 3 -1 = 2; she has doubled his roil and has done something in desperation that has killed him. Rolling a larger die is a huge advantage, but even a d20 has a "1" on it and might be defeated and even killed by a d6. In this case the warrior rolling a"1" means in his eagerness he trips and falls beside the woman on the ground who immediately slashes his throat with her dagger. Had she rolled a 4 -1 = 3 she would have tripled his 1 and made a spectacular kill: she pulled a long pin from her hair and drove it through his ear into his brain. Her friend nearby sees it and gains a burst of energy (+1 to next roll) and the watching warriors test COOL to attack with revenging ferocity (on a failed COOL roll) OR temporarily draw back in astonishment (made their COOL). The fun is narrating the action, telling the story, to account for what the dice say has happened. If the warrior triples the woman's roll he may decide to take her captive, disarming and restraining her. We want the players to put themselves into the little figures on the table and TELL THE STORY,
BONUSES FOR WEAPONS AND POSITIONS:
Weapons such as lances with a long reach give the user +1. If one fighter has a shield or better armor than the other who does not, he gets a +1. A location such as being uphill or on firm ground, versus an opponent climbing or slipping, gives +1. If one fighter is moving with intent and his enemy is stationary, the momentum gives a +1. Attacking by surprise or from the rear would give a +1. These additions are cumulative once but not more; i.e. attaching downhill (+1) by surprise (+1) with a lance (+1) gives an advantage of +2 and not +3. If a fighter is immobilized and cannot dodge, losing by 1 means he is killed.
TWO-ON-ONE:
If one side has a numerical advantage, it rolls the dice for each fighter involved, but does NOT add them together; rather, use the best roll and then add +1 for each extra fighter. If outnumbered three to one or worse, any losing roll is a kill.
DECIDING WHO FIGHTS WHO:
Often who matches up to who will be obvious, but if it is not roll to see which player moves first, then take turns moving one figure into contact with an enemy. Once in contact no further movement is allowed. It will often happen that a figure is pinned but then his friends move up to help, and so in alternation until bot sides have multiple fighters in the melee. Simplify by reverting to one-one-one matchups until one side has only one fighter left, who will face however many the more numerous side still has. Remember, again, that everything is happening all at once. It is often easiest, if both players agree, to match the strongest fighters against each other and continue down.
An example of play:
A drying rack is defended by one warrior (d12), one woman (d8), one child (d6) and two dogs (each a d6). It is menaced by a pack of wolves: the alpha male (d10), two other adults (d8s), and three youngsters (d6s). The wolves' aim is not to kill but to snatch the meat and run away.
The humans form a circle around the jerky. The wolves, being the outside circle and the aggressors, get to chose who fights who. The alpha and another adult double-team the warrior; the remaining adult threatens the woman; and the three youngsters confront the child and the dogs.
Warrior (d12) vs alpha plus one (d10 +1)
Woman (d8) vs adult (d8)
Child (d6) vs youngster (d6)
Dog (d6) vs youngster (d6)
Dog (d6) vs youngster (d6)
The warrior rolls 10 versus the alpha's 8 + 1 and has gained a small advantage for next turn.
The woman rolls a 6 versus the wolf's 4 and has a small advantage next turn.
The child rolls 4 and the wolf 6; advantage wolf.
First dog and facing wolf each roll 3; they snarl at each other.
Second dog and wolf each roll 6; they snarl even louder.
Next turn the young wolf uses its advantage to dash past the child and grab a mouthful of jerky. The alpha and adult wolves draw back a pace, refusing to close when the warrior and woman have an advantage. One young wolf rolls a 2 but the dog a 1; the wolf lunges and tears out the dog's throat. The other dog rolls 5 and the wolf a 3; the wolf is bitten but escapes and retreats from the fight.
Note that the warrior and the woman, because they gained an advantage and their facing wolves draw back, would have a few seconds to do something such as yell orders to the child or the dogs, or switch places, or for the warrior to spear a piece of meat and sling it over the heads of the wolves, or to hurl his spear at the wolf with the meat in its mouth. TELL THE STORY!
Next turn the wolves may retreat and the fight is over; they have gained a bit of tasty jerky but have an injury. The humans have lost a dog and a quarter of their meat production but are unharmed.
Or, the wolves may press the attack:
Warrior rolls 4 vs alpha and adult rolls 3 +1
Woman rolls 3 vs adult rolls 4
crying child plus dog vs remaining young wolf
The child rolls a 1 but the dog a 5 which is used, with a +1 for the child being present; the wolf rolls a 1 and is torn apart by the dog. The remaining wolves will retreat and have now definitely LOST the fight. They should have quit while they are ahead. The humans gain a wolf skin and the dog is renowned.
My prompt:
Downloaded file is my rules for close combat. Critique please.
ChatGPT:
Below is a design-focused critique of your close-combat rules, aimed at clarity, internal consistency, probability behavior, and play experience—not rewriting your prose voice, which is one of the strengths of the piece. I'll flag strengths, then issues, then concrete suggestions, with references to the text where relevant
hand to hand
.
________________________________________
1. Big-Picture Strengths
A. You are absolutely right about "what we are really doing"
Your opening philosophical framing is excellent and rare in rules writing. You clearly tell players:
• This is simultaneous action, not I-go-you-go
• Dice are story prompts, not physics engines
• Narrative explanation is not optional—it is the game
That framing will save you from 90% of rules-lawyering problems later
hand to hand
.
B. The "double / triple" kill thresholds are elegant
The ratio system (equal → nothing, higher → advantage, double → kill, triple → spectacular kill) has several virtues:
• Scales automatically across die sizes
• Preserves uncertainty (a d6 can kill a d20)
• Produces memorable outcomes without tables
• Gives you an organic morale hook
This is very well chosen for a skirmish narrative game.
C. Advantage carry-over is intuitive and cinematic
The "advantage persists to next turn as a modifier" mechanic is simple, visualizable, and narratively obvious. Dropped weapon, footing lost, momentary dominance—it all fits cleanly.
________________________________________
2. Where the Rules Start to Strain
These are not fatal flaws, but pressure points where players will ask questions or feel friction.
________________________________________
Issue 1: The die-size ladder needs a clearer fiction
Right now:
• Civilians = d6
• Novice warriors = d8
• Warriors = d10 / d12
• Mounted fighters = d20
This is fine mechanically, but the meaning of die size is not explicitly stated. Is it:
• Training?
• Physical power?
• Weapon reach?
• Mobility?
• Aggression?
At the moment it feels like everything at once, which works narratively—but players will ask:
"Why does being mounted jump all the way to d20 rather than d12 + bonuses?"
Suggestion (light touch):
Add one sentence:
"Die size reflects the fighter's overall combat agency: training, mobility, weapon reach, and ability to exploit openings."
That one sentence will head off endless debates.
________________________________________
Issue 2: Advantage stacking rule is slightly opaque
You say:
"These additions are cumulative once but not more… gives an advantage of +2 and not +3"
hsnd to hand
Players will ask:
• Why +2 specifically?
• What if there are four factors?
• Does advantage from last turn count toward the cap?
Right now, this rule is doing important balancing work, but it's under-explained.
Suggestion:
Clarify explicitly:
• Situational advantages cap at +2
• Carry-over advantage from prior turns is separate
One sentence would do it.
________________________________________
Issue 3: The "immobilized = killed if lose by 1" rule is powerful but buried
This is a huge rule:
"If a fighter is immobilized and cannot dodge, losing by 1 means he is killed."
hand to hand
This changes the entire tactical landscape (grapples, falls, entanglement, dog piles), but it's introduced almost as an aside.
Suggestion:
Give it a bold sub-heading or a boxed call-out:
IMMOBILIZED OPPONENTS
A fighter who cannot dodge is at lethal risk.
Otherwise players will miss it—and it deserves to be noticed.
________________________________________
Issue 4: Two-on-one math may be slightly too lethal at higher numbers
Mechanically:
• Best roll +1 per extra fighter
• At 3-to-1, any loss = kill
This produces very fast collapses once numbers stack, which may be intended—but note the emergent behavior:
• Three mediocre fighters can almost auto-kill elites
• Morale effects may never trigger because fights end instantly
This is not "wrong," but it does mean:
• Multi-figure melees resolve faster than single duels
• Story density actually drops once numbers rise
Suggestion (optional):
Consider a middle tier:
• 3-to-1: any double = kill (not any loss)
• 4-to-1 or more: any loss = kill
That preserves fear of numbers without making every swarm instantly decisive.
________________________________________
Issue 5: Matching mechanics vs. simultaneity tension
You repeatedly emphasize simultaneity (correctly), but the matching procedure is still sequential and slightly artificial:
"roll to see which player moves first, then take turns moving one figure into contact…"
hand to hand
This is fine for playability, but you might acknowledge the abstraction explicitly, or players may try to game match-ups too precisely.
Suggestion:
One sentence like:
"This matching procedure is only a convenience; it does not imply hesitation or orderly pairing in the fight itself."
This keeps the philosophical frame intact.
________________________________________
3. The Example of Play: Very Strong, One Small Fix
The wolf-jerky example is excellent:
• Clear objectives
• Asymmetric intentions (snatch vs kill)
• Dogs, children, morale, retreat logic
• Narrative decisions mid-melee
It does exactly what you want the rules to teach.
One small concern:
You occasionally allow post-roll tactical choices (wolves drawing back after seeing results). That's fine—but be explicit that this is allowed because:
• Advantage does not force continued engagement
• Initiative is fluid
Otherwise players may argue that "contact means locked."
A single clarifying line would suffice.
________________________________________
4. Overall Assessment
What you have done well
• Created a robust narrative engine
• Avoided tables and modifiers bloat
• Made morale emerge naturally
• Preserved uncertainty even for elites
• Encouraged player authorship
What I would tighten
• Clarify what die size means
• Surface a few crucial rules more clearly
• Slightly refine multi-attacker lethality
• Make abstraction vs simultaneity explicit
None of these require rewriting the system—just polishing the joints.
________________________________________
If you want, next steps could be:
• A one-page player summary stripped of philosophy
• A designer's sidebar on why "perfect play" is impossible (which fits your Iliad-inspired thinking)
• Stress-testing edge cases (grapples, captives, ladder assaults, mounted vs prone)
This is already a very solid close-combat engine—importantly, one that fits your broader design ethos rather than fighting it.
Sources
Thanks, those are excellent suggestions which I will implement.
You're very welcome — and I'm glad they were useful.
What you have here is already doing something many skirmish systems fail to do: it teaches players how to think, not just what to roll. The fact that you're willing to surface assumptions, cap lethality consciously, and make simultaneity explicit puts this in the category of designerly rules, not just functional mechanics.
A couple of closing thoughts as you implement the tweaks:
• Resist over-clarifying. Your prose has authority because it invites judgment rather than eliminating it. Every clarification should reduce argument, not reduce imagination.
• The die-size explanation sentence will quietly anchor the whole system; it's worth getting that wording exactly right.
• Pulling the immobilized rule forward will pay dividends later when you get to ladders, grapples, dogs, and wounded figures — it's a lynchpin concept.
• Your instinct to let numbers matter without turning combat into arithmetic is exactly right for the kind of story-rich, morale-driven play you're aiming for.
When you're ready, I'd be happy to:
• Review the revised draft for tone creep or unintended consequences
• Help you write a boxed "Why This Isn't Chess" designer's note
• Or stress-test the rules against specific edge cases (Comanche lance charge, wolves vs mounted men, ladder teams under fire, captives vs kill decisions)
This system is very much of a piece with your larger project. It feels like rules written by someone who has actually watched fights unfold — historical, animal, and human — rather than by someone optimizing probabilities in a vacuum.
finds bases at the dollar store!