How long before VTTs support TTRPGs with AI GMs?

[MoonSong]

BTW, if I'd asked it to create guardrails and set a safe word required to break the veil, told it to cut down on superfluous rolls, anything else, it would have. Refining the experience is as easy as asking for what you want.

I'm not suggesting it's perfect, but it's certainly better than being in a game with a bad human DM, right? Let's be honest here, it's entirely playable as a solo D&D experience. I can't think of a better solo published D&D product.

Here's the point. IMO if you are playing solo, at that point why not write a novel with random dice if you feel like it? or grab a few dolls and play pretend with them...?

 

[SableWyvern]

I'm not suggesting it's perfect, but it's certainly better than being in a game with a bad human DM, right?

Not in any meaningful sense.

I would much prefer to play with a bad GM who is a friend and doing their best in good faith, than spend time trying to game with an LLM. I could have plenty of fun with the former.

If you're talking about a bad GM I don't like, running a bad game I don't enjoy, well then any comparison is pointless -- whether one is better or worse than the other is irrelevant, as they're both bad enough that I won't waste my free time participating.

You're welcome to like the LLM, but you're not going to convince me that I should like it. I am impressed at what they can do, because it is impressive. I just don't find it useful to me in any way.

 

[kermit4karate]

Ongoing excerpt from the 5e game with ChatGPT today. First example of combat.

The storm thrashes harder through the chapel rafters. Lightning flashes, throwing stark light across the broken altar where the Moonstone Shard rests. It pulses faintly now, no longer binding the child — but something else stirs.

From the shadows of the collapsed nave comes a hollow scraping sound, like bone dragging over stone. A figure emerges — skeletal, armored in rust, its eyes burning faint blue to match the shard. Behind it, two more shapes rise, skeletal sentries that had been buried in the rubble until your unbinding woke them.

Bram growls, hefting his axe.

“Finally! Something to hit.”

Maribel’s voice trembles but steadies as she grips her holy symbol.

“Saints preserve us… the shard’s ward held them too. They come for it!”

The lead skeleton points its corroded sword at you, jaw clacking in a silent challenge.

---

Combat Begins​

Initiative order (rolled quickly):

• Fizzbang (you)
• Bram (Fighter)
• Skeletons (3)
• Maribel (Cleric)

You act first.

The ruined chapel is about 30 feet wide. The skeletons are 20 feet away, approaching from the shadows. The shard lies on the altar just beside you.

---

Your turn, Fizzbang.
What do you do?

You said:​

What are my prepared spells and cantrips?

ChatGPT said:​

Here’s Fizzbang’s spell setup as a level 3 wizard (rules-accurate):

---

Spell Slots​

• 1st level: 4 slots
• 2nd level: 2 slots

You can prepare 6 spells each day (INT 16 → +3, plus wizard level 3 = 6).

---

Cantrips (known at level 3: 3 total)​

• Fire Bolt (ranged attack cantrip, 1d10 fire damage, range 120 ft)
• Prestidigitation (minor magical effects, flavor/mischief utility)
• Mage Hand (spectral hand, utility spell)

---

Prepared Spells (6 total)​

Here’s the loadout I set up earlier (balanced for offense, defense, utility):

1st level (4 slots):

• Mage Armor (already cast at the start of the day, AC = 13 + Dex = 15)
• Magic Missile (auto-hit force darts, 3 × 1d4+1 damage, split as you choose)
• Shield (reaction, +5 AC until start of next turn)
• Detect Magic (utility, concentration, sense magical auras)

2nd level (2 slots):

• Misty Step (bonus action, teleport 30 ft)
• Scorching Ray (3 rays, each ranged spell attack for 2d6 fire damage)

---

Summary​

• Cantrips: Fire Bolt, Prestidigitation, Mage Hand
• Prepared Spells: Mage Armor, Magic Missile, Shield, Detect Magic, Misty Step, Scorching Ray
• Slots: 4 × 1st level, 2 × 2nd level

---

With skeletons closing in (20 ft away), Fizzbang can go offensive (Fire Bolt, Magic Missile, Scorching Ray), defensive (Shield if attacked, Misty Step for positioning), or tactical (grab the shard, reposition, or command your allies).

Do you want to attack immediately, or do something clever with positioning/interaction before combat gets bloody?

You said:​

I'm going to move toward the shard and cast Fire Bolt at the lead skeleton.

ChatGPT said:​

Round 1 — Fizzbang’s Turn

You dart to the altar, the Moonstone Shard glowing faintly at your side. The lead skeleton, its rusted armor scraping, raises its sword — but you’re quicker.

You hurl a searing Fire Bolt.

---

Fire Bolt Attack​

• Range: 120 ft (they’re only 20 ft away, so well in range).
• Attack roll: d20 + 5 (INT +3, proficiency +2).

d20 = 14
+5 = 19 total
Hit! (Skeleton AC = 13)

---

Fire Bolt Damage​

• 1d10 fire damage.

d10 = 9

Result: The Fire Bolt slams into the lead skeleton’s ribcage, shattering bones in a burst of flame. The skeleton staggers but remains standing, its eyes burning brighter.

---

End of Fizzbang’s Turn​

• Lead skeleton HP: Badly scorched, bones cracking (still active).
• Two skeleton allies: Unharmed, closing in.

---

Next in initiative: Bram (the fighter).
He charges forward with a roar, axe swinging.

Would you like me to fully control Bram’s combat actions (as DM), or do you want to give him tactical directions in battle?

You said:​

Yes, please. Fully control both NPCs.

ChatGPT said:​

Perfect — from here on out, I’ll fully run Bram and Maribel as NPCs under my DM control. You just play Fizzbang.

---

Round 1 — Bram’s Turn​

Bram charges the lead skeleton (the one you scorched). His war axe arcs down in a brutal chop.

• Attack: d20 + 5 → 17 + 5 = 22 (hit).
• Damage: 1d8 + 3 → 6 + 3 = 9 slashing.

The skeleton crumbles apart under Bram’s strike, bones scattering across the wet flagstones. One down.

Bram spits on the rubble.

“That’s how you deal with bone-rattlers!”

---

Round 1 — Skeletons’ Turn​

Two skeletons remain. They charge forward, weapons raised.

• Skeleton #2 attacks Bram with a shortsword.
  
  • Attack: d20 + 4 → 12 + 4 = 16 (Bram’s AC = 17). Miss! The rusted blade clangs off his shield.
• Skeleton #3 goes for you, Fizzbang! Its shortsword thrusts toward your chest.
  
  • Attack: d20 + 4 → 18 + 4 = 22 (Fizzbang AC = 15 with Mage Armor). Hit.
  • Damage: 1d6 + 2 → 4 + 2 = 6 piercing.

Fizzbang takes 6 damage (HP: 16 → 10).

---

Round 1 — Maribel’s Turn​

Maribel raises her holy symbol, light flaring.

“Back to the grave with you!”

She uses Turn Undead (Channel Divinity). Both skeletons must make Wisdom saves (DC 13).

• Skeleton #2: 7 (fail).
• Skeleton #3: 12 (fail).

Both skeletons screech soundlessly, drop their weapons, and flee toward the far end of the chapel, compelled by radiant terror. They cannot willingly approach Maribel for 1 minute.

---

End of Round 1 Recap

• Skeleton #1: destroyed by Bram.
• Skeletons #2 and #3: fleeing (Turned).
• Fizzbang: wounded (10/16 HP).
• Shard: still glowing beside you on the altar.

---

Round 2 — Fizzbang’s turn.
The skeleton that struck you is retreating, back turned, weapon dropped. You have a chance to press the attack, grab the shard, or do something else entirely.

What does Fizzbang do?

You said:​

Can I grab the shard and still manage to attack the skeleton nearest to me?

ChatGPT said:​

Yes — that’s perfectly doable this round!

Here’s how it works rules-wise in 5e:

• Grab the shard: Picking something up from the altar is a free object interaction (doesn’t cost your action). You can scoop the shard into your hand or pouch without trouble.
• Attack: You still have your action to cast a spell or cantrip.

---

Round 2 Options for Fizzbang

1. Grab shard + Fire Bolt → blast one of the fleeing skeletons (ranged spell attack).
2. Grab shard + Magic Missile → guaranteed hits; could finish a fleeing skeleton outright.
3. Grab shard + Scorching Ray (2nd-level) → risky overkill, but you could annihilate both.
4. Grab shard + Defensive choice (Shield reaction if attacked, Misty Step reposition, or just hold action).

---

The shard pulses faintly in the rain as the skeletons scramble toward the broken nave.
Do you want Fizzbang to attack while grabbing the shard, and if so, which spell do you use?

You said:​

OK, I grab the shard and cast Fire Bolt at the nearest skeleton!

ChatGPT said:​

Round 2 — Fizzbang’s Turn

You snatch the Moonstone Shard from the altar. It feels cool in your hand, humming with faint lunar energy. The glow brightens slightly as you touch it, as if resonating with your presence.

Without hesitation, you spin and hurl a blazing Fire Bolt at the nearest fleeing skeleton.

---

Fire Bolt Attack​

• Attack roll: d20 + 5
  d20 = 11
  +5 = 16 total
  Hit! (Skeleton AC = 13)

Damage​

• Fire Bolt: 1d10
  d10 = 8

---

Result​

The bolt slams into the skeleton’s spine, sending shards of bone and fire scattering. The creature collapses in a heap of smoldering remains.

• Skeleton #2: Destroyed.
• Skeleton #3: Still fleeing, clattering toward the far end of the ruined chapel.

---

End of Round 2​

• Bram roars with laughter:
  
  

  “Nicely done, Fizzbang! Burned him to ash!”

• Maribel lowers her holy symbol, voice steady:
  
  

  “One remains. Don’t let it escape — turned or not, it will return if not destroyed.”

Rain hammers down through the ruined roof. The shard pulses faintly in your hand.

---

Round 3 — Fizzbang’s turn.
Only one skeleton remains (fleeing, 30 feet away).
Do you:

• Pursue and finish it,
• Hold position with the shard, or
• Try something else with your magic or wits?

You said:​

Shouting "Let's go after it!" I run after the last skeleton and cast Fire Bolt again.

ChatGPT said:​

Round 3 — Fizzbang’s Turn

You clutch the shard tightly, shout, “Let’s go after it!” and sprint across the slick flagstones, robes whipping in the storm. The last skeleton clatters away down the ruined nave, but you close the distance enough to fling another searing Fire Bolt.

---

Fire Bolt Attack​

• Attack roll: d20 + 5
  d20 = 13
  +5 = 18 total → Hit! (Skeleton AC = 13)

Damage​

• Fire Bolt: 1d10
  d10 = 10

---

Result​

The bolt ignites the skeleton mid-stride, and it collapses into a blazing heap, bones scattering across the chapel floor. The last of the undead guardians is destroyed.

---

End of Combat​

The storm begins to ease outside, thunder rolling farther away.

• Bram plants his axe in the rubble, breathing hard but grinning.
  
  

  “Hah! That’s more like it. Fizzbang, you’ve got some fire in you—literally.”

• Maribel wipes rain from her brow, looking at the shard in your hand.
  
  

  “That relic is not just a trinket. I fear it has drawn notice… and will draw more.”

The chapel grows unnervingly quiet. Only the shard pulses faintly, like a heartbeat, in your grasp.

---

Do you want to:

1. Examine the shard immediately (Arcana check, no combat pressure now).
2. Let Maribel consecrate it with a ritual.
3. Pocket it and leave quickly before more danger arrives.
4. Search the chapel ruins for anything the skeletons may have guarded.

What does Fizzbang do next?

 

[Epic Meepo]

Call center bots are expert systems. They have to be programmed manually.

Fair point. But I'd say an AI GM is just one step removed from that; an AI GM would be an expert system that runs game rules paired with a LLM that generates adventure content and NPC dialogue.

(You'd want an expert system like a VTT to handle the game rules whenever the LLM isn't needed to improvise adventure content or dialogue. Otherwise, you're just playing make-believe with an LLM that can hallucinate incorrect game mechanics at any time.)

 

[kermit4karate]

This is part 4 of the ChatGPT D&D 5e game I started yesterday on this thread for people curious about how far along AI has progressed in the past couple years.

If you're just showing up to this thread, this is a simple demo of me playing a solo D&D adventure with a modern AI as the DM. I'm not suggesting AI is wonderful or proving/validating/defending anything philosophical about AI by posting this, merely sharing an example of where the technology stands at this point in time to help you judge for yourself without all the background chirping.

-----

You said:​

I quickly pocket the shard, say "Maribel, hurry!" to Maribel, and step outside beside Bram. "What is it, my friend? What do you hear?"

ChatGPT said:​

The Moonstone Shard disappears into your robes with a faint hum, its light muffled but not extinguished. You stride to Bram’s side at the chapel’s broken doorway, rain dripping from the shattered arch above.

Bram narrows his eyes into the treeline, knuckles white on the haft of his axe.

“Footfalls. Not beasts. Two… no, three sets. Heavy. They’re circlin’, tryin’ to box us in.”

The night outside is black save for flickers of moonlight breaking through storm clouds. Shapes move among the skeletal trees, cloaked figures weaving between the trunks. The wet mud muffles them, but the occasional glint of metal catches your eye.

Maribel’s chant falters for a breath as she raises her voice.

“Fizzbang, I can end the ritual now, but it will be incomplete. The chapel’s sanctity will hold for days, not years. Shall I stop?”

Bram spits into the mud.

“Forget the stones — better we live to fight another day. What’s your call, wizard?”

---

You must decide quickly:

1. Order Maribel to complete the full ritual (10 minutes), risking an ambush while you and Bram defend her.
2. Tell her to end it now — weaker consecration, but you can move immediately.
3. Leave the chapel unfinished, retreat into the storm with your companions.
4. Prepare an ambush: you and Bram set up just outside the doorway, using the bottleneck of the ruined chapel to your advantage.

What does Fizzbang choose?

You said:​

"Maribel, you better get out here!" I say. "We have company." Turning my attention to the cloaked figures at the tree line I shout, "You stand at the precipice of pain and death, children of the damned! If you hope to sleep with the living again, you will turn back now! You will find only blood and doom here!"

ChatGPT said:​

Thought

Your shout cuts through the drizzle like a thrown blade. Thunder rolls as your words hang in the wet air.

I’ll resolve how the cloaked figures react — this is a meaningful, uncertain moment, so I’m rolling an Intimidation check for Fizzbang (you already roleplayed the line).

• Intimidation roll: d20 + CHA modifier (+1).
• Roll result: 17 → 18 total.

Outcome​

Your voice, the crack of thunder, and the shard’s faint, uncanny glimmer make an unexpectedly fearsome combination. The three cloaked figures freeze where they stand, half-hidden in the trees. For a tense heartbeat they are statues — wet cloaks hanging, boots shuttering in the mud.

One of them, the tallest, lifts a hand slowly and pulls back his hood. A young man’s face is revealed, streaked with rain and terror rather than bravado. He glances at his two companions; the small trio exchange nervous looks. They were expecting an easy ambush or a frightened few to intimidate — not a wizard who answers with thunder and poetry.

He spits out a shaky curse and mutters something that could be a command, but the command dies on his lips. The tallest shakes his head, and the three take a half-step backward, forming a loose triangle as if reconsidering the plan. They do not flee — not yet — but their forward hunger has been checked. They are weighing whether to press an attack or withdraw into the woods and try again later.

Bram huffs a laugh, tightening his grip on the axe.

“Good show, Fizzbang. That scared the teeth out of ’em.”

Maribel steps to the threshold, voice steady and low:

“They hesitate. We might learn why they came — or they might draw reinforcements. We must decide quickly.”

---

What happens next​

The cloaked trio are still a threat but currently slowed by uncertainty. You have the initiative of the moment — a few immediate choices that will shape how the encounter unfolds. (No need to ask; pick an action and I’ll resolve it.)

Options you can take now (you’re in the doorway with Bram; Maribel is ready behind you):

1. Demand parley — call them out and try to learn who sent them or why they came (Persuasion/Intimidation follow-up; they may answer truthfully or lie).
2. Press the attack — charge out with Bram and force them to fight before they can regroup (combat likely).
3. Cast a spell — use a nonlethal spell to coerce/disable (e.g., Misty Step to flank, Magic Missile to warn/disable, Scorching Ray to burn).
4. Show the shard — reveal the Moonstone Shard and demand answers; its glow could frighten or entice them (use of shard might attract attention later).
5. Set an ambush — have Bram and you hold position while Maribel finishes a quick partial blessing (short action) and prepare to strike when they move.
6. Withdraw — step back inside; let them withdraw or shadow you; fight another day.

(You may also do something creative — shout a paralyzing chant, toss a small trinket, or use a cantrip for theatrics.)

What does Fizzbang do next?

 

[LordEntrails]

I tend to think we'll see at least one of the major VTTs offering fully AI GMs as options within one year. Two at the absolute outside.

Yea, one year. But I think it will be a POS and pretty darned flaw. But within 3 years I think there will be a couple of VTTs that have working/acceptable AI integrations.

How about you, and how do you feel about it? Have you tried playing a TTRPG with an AI like ChatGPT yet? If so how has it gone?

I'm fine with it happening. It won't nominally be of interest to me (I'm currently running one group and in another. The group I run is approaching 10 years with no sign of breaking up. The other is new but has potential to run for years.

I haven't done it yet myself, but know there are several very active folks doing it on Fantasy Grounds and sharing with each other on the forums over there. They see the short comings mentioned, but seem happy with it. It allows them to play better than if they didn't have it.

https://www.fantasygrounds.com/forums/showthread.php?82149-Solo-Play-Guide-for-FGU-INDEX

https://www.fantasygrounds.com/forums/showthread.php?85539-Let-s-Play-Solo!

 

[Jfdlsjfd]

But I don't think TTRPGs are where the proverbial money is. AI in gaming is really going to shine when a studio releases the first big-name CRPG with integrated LLM functionality. CRPGs have made use of primitive "AI" for decades, so upgrading them with actual, cutting-edge AI is the obvious next step.

I am not sure casual gamers would love effectivve AI that wake the whole camp to alarm with their shouts instead of fighting to the death when they first detect the intruder, allowing one to clear a map in a series of small easy fights, though...

Then there are issues with prompting, to the extent that I'm not sure such an exercise can be called a game. (Or rather, it is a different type of game than what the user thinks). By this I mean that it would be easy for the user to break the model with a prompt like "Imagine my character secretly leveled up 20 levels and has an ultimate attack that kills everything". That is a very different experience than traditional D&D, where it is the job of the DM to create and enforce guardrails.

TBH, I tried to do something like that (doing something that is totally outside of the bounds) and the model (chatgpt in that case) chided me. He'd have gone along with this if I had insisted, because it is designed to do what the user wants, but in a cooperative game the other players would certainly veto your attempt to break the AI GM, wouldn't they. You can after all try that wit ha real GM "hey, there are millions of ant in an ant colony, I pour boiling water on one, get 1 xp each, and I am now level 20" and see how the other players react if the GM (who happens to be your boyfriend/girlfriend) considers agreeing with that.

The best generic LLM can do is gamebook-like CYOA play. And even then, you have to prompt them hard at the beginning of the session to explain the concept of stats, dice roll, and how they should help determining the outcome of the action described.

see the same argument often regarding AI being good--"it may have issues, but I generated it so fast". And that's because AI has changed the game with respect to generation; you can generate decent sounding stuff in seconds, whereas it used to take hours. But that is very different from creating things that are good. Spend an hour or two on it--does it improve? Is the result more useable? I doubt it, because you are hitting fundamental limitations of the technology. [...]

If you're going to run a game system, you need to adjudicate the rules properly. I wouldn't play with a human GM who didn't, and I won't play with an AI GM who doesn't.

I'd say it's because you tried (from lack of dedicated AI) doing it with a general purpose LLM. It can't replicate correctly the rules of D&D because it doesn't know them. He might have been trained upon them, but the details are lost in training and the exact wording of the game books aren't present in the model. Other AI, designed to connect to a specific database, offer better results (that's how legal or medical AI can actually give you correct references while generic LLM can't). If a company was going to create an AI GM, it would certainly spend more than an hour or two to (1) be connected to a database containing the rules (2) don't allow players to change the rules unless they are the game host (who might need to introduce house rules) and yes, it would improve the result. Wether it's economically worth it remains to be seen, but I think CRPG will be the first to do that, since they need to implement the rules of their game in the first place.

 

[kermit4karate]

I am not sure casual gamers would love effectivve AI that wake the whole camp to alarm with their shouts instead of fighting to the death when they first detect the intruder, allowing one to clear a map in a series of small easy fights, though...



TBH, I tried to do something like that (doing something that is totally outside of the bounds) and the model (chatgpt in that case) chided me. He'd have gone along with this if I had insisted, because it is designed to do what the user wants, but in a cooperative game the other players would certainly veto your attempt to break the AI GM, wouldn't they. You can after all try that wit ha real GM "hey, there are millions of ant in an ant colony, I pour boiling water on one, get 1 xp each, and I am now level 20" and see how the other players react if the GM (who happens to be your boyfriend/girlfriend) considers agreeing with that.



I'd say it's because you tried (from lack of dedicated AI) doing it with a general purpose LLM. It can't replicate correctly the rules of D&D because it doesn't know them. He might have been trained upon them, but the details are lost in training and the exact wording of the game books aren't present in the model. Other AI, designed to connect to a specific database, offer better results (that's how legal or medical AI can actually give you correct references while generic LLM can't). If a company was going to create an AI GM, it would certainly spend more than an hour or two to (1) be connected to a database containing the rules (2) don't allow players to change the rules unless they are the game host (who might need to introduce house rules) and yes, it would improve the result. Wether it's economically worth it remains to be seen, but I think CRPG will be the first to do that.

You've got some inaccuracies here. First, it's trivial to create boundaries during play now. You just set a safe word and agree to the parameters you set for using it. You tell the AI not to allow you to break character or override a ruling unless you use the safe word first -- so, unless you say "banana," or whatever else you set. Instruct the AI that you want it to end the game immediately if you try to get it to change a ruling without using the safe word.

Obviously, it's still just a tool (for the time being), but it's a tool that listens to you. One does have to voluntarily agree to follow their own rules. The AI isn't going to fight you for supremacy (yet).

You do bring up a point about general purpose LLMs versus purpose-built LLMs with custom datasets. Those do exist and can be superior to general LLMs for certain tasks, but the general purpose LLMs like ChatGPT can do/play any game for which the rules are publicly available (published online like 5E is). It knows the rules and follows them at least as closely as the average human does.

 

[Jfdlsjfd]

You've got some inaccuracies here. First, it's trivial to create boundaries during play now. You just set a safe word and agree to the parameters you set for using it. You tell the AI not to allow you to break character or override a ruling unless you use the safe word first -- so, unless you say "banana," or whatever else you set. Instruct the AI that you want it to end the game immediately if you try to get it to change a ruling without using the safe word.

Sure, that's part of answer to "does it gets better if you spend one hour or two over it?" and what I covered in the "initial prompting to explain the goal of the game". [you need to make sure the LLM will create conditions for the game to end]

You do bring up a point about general purpose LLMs versus purpose-built LLMs with custom datasets. Those do exist and can be superior to general LLMs for certain tasks, but the general purpose LLMs like ChatGPT can do/play any game for which the rules are publicly available (published online like 5E is). It knows the rules and follows them at least as closely as the average human does.

In my experience it was quite... erratic on what the dragonmark of making was able to do, despite it being described online. It was clearer on rules that are more frequently presented. It failed to be forthcoming when adjucating whether advantage was warranted or adjucating DCs, but it could fall under the "as closely as the average human" proviso.

 

[The Firebird]

TBH, I tried to do something like that (doing something that is totally outside of the bounds) and the model (chatgpt in that case) chided me. He'd have gone along with this if I had insisted, because it is designed to do what the user wants, but in a cooperative game the other players would certainly veto your attempt to break the AI GM, wouldn't they. You can after all try that wit ha real GM "hey, there are millions of ant in an ant colony, I pour boiling water on one, get 1 xp each, and I am now level 20" and see how the other players react if the GM (who happens to be your boyfriend/girlfriend) considers agreeing with that.

I picked that as an edge case. But it's easy to imagine guiding the LLM more subtly. "I wonder if there is any cover I can use here..." for example. If it isn't running a fixed world, you have a lot of power to create things on the fly. Some players like that; it would work much better with narrative games, I imagine.p

Other AI, designed to connect to a specific database, offer better results (that's how legal or medical AI can actually give you correct references while generic LLM can't). If a company was going to create an AI GM, it would certainly spend more than an hour or two to (1) be connected to a database containing the rules (2) don't allow players to change the rules unless they are the game host (who might need to introduce house rules) and yes, it would improve the result. Wether it's economically worth it remains to be seen, but I think CRPG will be the first to do that, since they need to implement the rules of their game in the first place.

Point taken, but general purpose LLMs are quite good about pulling references if you ask for them these days.

It knows the rules and follows them at least as closely as the average human does.

This does not appear to be the case based on your examples of play. Unless the people you have played with make up many variant rules?