Be a GAME-MASTER, not a DIRECTOR

[MichaelSomething]

You can't teach experience.

So you're saying we need to create a DM dungeon of some type where they can grind DM exp???

 

[The-Magic-Sword]

Don't be a novelist, either. TRPGs are not (usually, at their best) linear authored fiction; most of the techniques and approaches from linear authored fiction will not apply. (This includes TV, comics, and video games, as well.)

I'm starting to realize this is good advice for players as well, a lot of frustration comes from expectations that we know the ending and that the story will have an arc like canned fiction.

 

[prabe]

I'm starting to realize this is good advice for players as well, a lot of frustration comes from expectations that we know the ending and that the story will have an arc like canned fiction.

Yeah, it's a good thing for the players to understand the differences in the media, at least as much as it is for the GM to. I think it's especially relevant for players if their expectations are constructed around pop-culture things (in non-TRPG media) the GM doesn't know. I pretty routinely tell the players at tables I'm running that if they're making a pop-culture reference they should just presume I won't get it.

 

[tetrasodium]

I'm starting to realize this is good advice for players as well, a lot of frustration comes from expectations that we know the ending and that the story will have an arc like canned fiction.

So very much yes. It can be really frustrating to watch a player who is clearly unhappy with or tired of their PC for whatever reason turning down offers to help set them up with a changed different or new PC because they have some kind of death grip on the idea of "finishing [PC name]'s story. Same with a player who has their heart set on playing out a totally setting/campaign/system inappropriate PC's story.

 

[prabe]

So very much yes. It can be really frustrating to watch a player who is clearly unhappy with or tired of their PC for whatever reason turning down offers to help set them up with a changed different or new PC because they have some kind of death grip on the idea of "finishing [PC name]'s story.

I guess how much that would bother me would depend on how clearly the player was done with the PC. I've finished a lot of novels I didn't really enjoy, after all--completion urge (especially when it comes to stories) is a real thing. You want a new PC, and you want to wrap up your current PC's story? Sure, lemme see what I strings I can push and what buttons I can pull; you probably gotta help, though.

 

[Celebrim]

It's hard for me to evaluate a video such as this...

Honestly, I've tried to avoid commenting on the video because I think I can evaluate it and well, and I think I can sum up the video with, "I wonder if he wrote a script for the video."

I get so tired of this sort of, "It doesn't have to be work!" crap. I can watch Mercer and see the sort of qualities I associate with high skill GMing. I can listen to Seth Skokowsky and immediately tell this is a guy who has been in the trenches and leveled up as a GM. I have watched Ginny Di grow in maturity and understanding as a GM as she gains more and more nuanced takes. And then someone with no credentials as a GM and absolutely no evidence that they can GM and tons of tells that he just doesn't get it, decides to tell novice GMs that they shouldn't pay attention to the people with actual skills and spends the first 7 or 8 minutes of the video saying literally nothing of consequence and the rest giving bad advice. It's a video that reminds me of one of those click bait advertisements for weight loss products or how to succeed with women where they go on and on and on about what they are going to tell you without actually telling you anything because they really don't have a product.

I really don't know what is worse: if there was a script for this video or there wasn't.

You do not want to put yourself in the same position as your players have trying to plan or evaluate or strategize or improvise on the fly. Because one of the only things worse than a session where the players are facing choice paralysis or really have run out of inspiration is when you as the GM are facing choice paralysis and have run out of inspiration. It's bad enough when your players are blind and are groping around for direction or just going off on tangents, but if the GM is blind and groping around for direction and going off on tangents then the players are (speaking from experience) in hell - and not in the good "Paladin in Hell" sort of way.

I'm really not even sure the speaker has ever GMed a game if he wants to give as an example that the villain doesn't have enough movement rate to beat the players to the foozle and so they get there first, because does he have any freaking idea how much preparation it takes to know exactly when an NPC leaves point A and arrives at point B without winging it such that it's effectively a teleport? Like that level of attention to detail you just don't improve. That's Tolkien making accurate maps of his world and making calendars and calculating movement rates so that he can track all of this moving pieces accurately level of attention to detail and preparation.

Let me give the untutored a clue. If you write down what the goal of your NPC is, you've done absolutely no useful prep at all. About the only thing more useless than that is writing down that your NPC is funny or charming.

The product he pitches at the end of the video sure looks like prep to me, albeit the previews look like useless prep and give me more "ick" than the infamous prostitute table in the 1e DMG.

 

[Gorgon Zee]

Well, also a software Dev, and also have read the Nethack code, but don't agree with you here. You seem to have a very different idea of what it means to have special handling.

I'm not sure how you can think that, as my statement was about "specific code for each scenario". Your statement and example is about rules for a single action. You are arguing about a single action; I am arguing about a set of actions.

Let me iterate: I believe that coming up with a scenario by combining known rules in a way that was not previously planned for is improvisation.

I was thinking rather than Nethack can't suddenly provide hew regions to explore that it hasn't been programmed to provide

And neither can you or I come up with new regions to explore that we have no knowledge or understanding of how to provide. I think maybe this is your key logical error -- for a computer to improvise you require it to create things it knows nothing about, but for people you are content that they combine things from knowledge they already have.

If you apply the same definition of improvising as you do to netback, then you you should be able to fulfill this request:

Create a new region for the game @Gorgon Zee is currently running, including descriptions of the environment and things that tie into their player character's backgrounds.

I believe this fits your description of what Nethack cannot do (and so cannot improvise, according to you). So for your definition to hold water, you have to show that it's actually possible to do at all!

 

[Celebrim]

I'm not sure how you can think that, as my statement was about "specific code for each scenario". Your statement and example is about rules for a single action. You are arguing about a single action; I am arguing about a set of actions.

Fist, this seems like one of those cases where bad definitions are leading to confusion. I'll attempt as best as I can to concede your definitions for the purposes of communication, but ultimately I don't think they are coherent. I don't think "a set of actions" in this context has a lot of meaning. In Nethack actions are taken discretely. There isn't really any meaningful concept of "a set of actions" except in the mind of the player. Players can exercise creativity by performing a series of actions but they can neither improvise (in the sense of doing something wholly new) nor can the program itself improvise. Further, by this point in the code's refinement, pretty much any sequence of actions has been discovered and pretty much any circumstance that a sequence of actions could create has been studied and thought about and if warranted special handling for the circumstance is added. For example, entering or leaving a shop by means other than the door is handled. For example, if you dig a hole in the floor you can leave the shop and won't be charged with theft unless you carried out an object or an object fell through the floor with you. That your pet can carry out an object out of the shop without flagging you for theft isn't improvisation, but designed behavior.

Let me iterate: I believe that coming up with a scenario by combining known rules in a way that was not previously planned for is improvisation.

But even by your own definition, the computer can't do that. For one thing, the computer has zero understanding of the sense of anything it does, and that puts it in a very different class of language users and tool users than a human. The computer can't handle any proposition in any combination except what it was programmed to handle because it doesn't understand the sense of any proposition. A human can propose to both jump and kick in the same action. In nethack you can't even input that proposition much less expect the computer to handle it. A human could resolve the improvised "jump kick" action because it understands the sense of both actions and attempt to handle.

And neither can you or I come up with new regions to explore that we have no knowledge or understanding of how to provide. I think maybe this is your key logical error -- for a computer to improvise you require it to create things it knows nothing about, but for people you are content that they combine things from knowledge they already have.

You seem to be attempting to argue that there is no difference in the improvisational skills of a human and the nethack program. But the key logical error you are making here is the nethack program knows nothing. It has no sense of anything. So it can't combine anything from knowledge it already has because it has no knowledge. Thus your point about you can't improvise unless you know something while it is true also misses the point. I actually know something. Nethack has zero knowledge.

To see what I mean, let's not focus on what a human and a computer program both can't do but what a human can do that a computer program could not. And the obvious example is suppose that Nethack was a pen and paper game using the rules of Nethack. Then the human GM taking the computer's role could decide to add a Kobold town as a branch on the second floor, and an orc town as a branch on floor 7, and a dragon's lair branch between floors 15 and 19, and 4 different branches of Hell reached via a new river Styx set of maps, and any other things by combining elements from the game in new ways. He could also choose to add new monster variants and rules extensions. He can improvise all these things because he actually understands what an "elf" is and what game they are playing. The nethack computer program has zero understanding of the meaning of what it does.

If you apply the same definition of improvising as you do to netback, then you you should be able to fulfill this request:

Create a new region for the game @Gorgon Zee is currently running, including descriptions of the environment and things that tie into their player character's backgrounds.

I believe this fits your description of what Nethack cannot do (and so cannot improvise, according to you). So for your definition to hold water, you have to show that it's actually possible to do at all!

Do you see why this is spurious? I agree that you can't improvise without understanding, but that doesn't in any way impact my argument. I don't need to show that it is possible to improvise without understanding. I just need to show that there is a difference in the degree of understanding. For example, I was fully willing to agree that a LLM has some understanding in some sense and so can improvise to some degree. Exactly how much understanding a LLM has and whether it understands either the sense or meaning of a word is a matter of debate. I don't think we know and we are just beginning to understand how the apparent understanding of a LLM emerges. But the idea that there is no difference in the improvisational skills of a program like Nethack or Balder's Gate and a human GM is ridiculous, and the idea that handling a series of creative move choices using the rules is improvisation by the GM rather than the player is equally ridiculous.

 

[Gorgon Zee]

the computer has zero understanding of the sense of anything it does

It's an extreme position, but if this is axiomatic to you, then there is no point in any further discussion. If you don't think that sematic webs, knowledge graphs or all the tools that are used to capture understanding are not valid, then I don't think we have common ground. You are essentially asserting that AI cannot exist, which really makes any conversation about computer intelligence, understanding or improvisation a non-starter to you.

You seem to be attempting to argue that there is no difference in the improvisational skills of a human and the nethack program.

Not at all. You are making a strong statement: Nethack never improvises. You are attempting to characterize my position as an equally strong statement ("there is no difference"), whereas actually what I am saying is much weaker: Nethack can improvise.

I agree that you can't improvise without understanding, but that doesn't in any way impact my argument.

But since you believe that computers have "zero understanding", very simple logic says that of course you must believe that computers cannot improvise by exactly this statement.

I was fully willing to agree that a LLM has some understanding

... contradicting your previous statement that computers have "zero understanding"

The idea that there is no difference in the improvisational skills of a program like Nethack or Balder's Gate and a human GM is ridiculous,

Yes it is. Which is why I never stated that.

Anyway, I think we are far enough off on a tangent for this thread. Happy to continue, but maybe start up a new thread called "can computers understand" or something similar. Or we can just call it a day.

 

[Snarf Zagyg]

While I do my best not to be a director, I do sometimes get frustrated when players seemingly go out of their way not to engage the adventure. I ran an Angel (of Buffy fame) campaign years ago, and I had a player who really loved dragons so I included one in the campaign. We spent sessions building up the character's relationship with Sir Kay (of Arthurian fame) with plenty of foreshadowing of a dragon to come. When the big day came and the dragon appeared, the PC pretty much said, "Nah, not interested," and went to go do something else instead. Kind of pissed me off a bit.

The player hates while the master baits.