The Player vs DM attitude

[haakon1]

I just want to point out that this sucks in movies and literature as well.

And computer games. I love the Total War series. But when you've been allied with some NPC country for umpteen turns, given them money for umpteen turns, helped them when they're attacked, used them as your little friend that you give excess territories too, and their strength is a Canada to your US, why the heck does the game have them declare war on you with a piddly ineffective attack like blockading a port?

IMHO, it's poor AI programming with the priority on "give them a crack fix of action" rather than "make countries behave in their own best interest". But I guess they were just going for something different than I was. I have to assume "their king must have gone bonkers" to suspend my own disbelief and keep playing.

 

[ExploderWizard]

I don't play 4e. I don't adhere to the notion that the rules override the GM's common sense and rulings.

It's the other way around, to me: The GM's common sense and rulings override the rules if need be. That's why you have a flesh and blood GM in the first place, someone who is able to make the game and its rules work for the benefit and pleasure of all involved, and if this means nuking a rule, a bunch of rules, or heck, the whole rulebook to get there, so be it.

Gotta do some spreadin.

This approach to play isn't connected with any particular rule set. Heck, I do play 4E and run it in the same style as I have the previous editions. The rules are general guidelines that serve well enough until they start to produce something stupid. No set of rules can forsee every possible situation, so we have GMs to pick up where the rules leave something to be desired.

 

[Doug McCrae]

The ANTITHESIS of this approach is to treat NPCs as tools of the movie director to move the action along, rather than characters in their own right. If the NPC barber in the Western suddenly turns into a psychopath who wants to cut the PC's throat -- even though he was a normal person with no warning signs or previous incidents -- just because the DM decides it's time for a fight, then you're going to get paranoid players.

In the real world, psychopaths are very good at pretending to be normal. All these interviews you see afterward where the neighbour says, "It's funny, he didn't seem like an axe-wielding maniac when I spoke to him over the fence. And he was great with the kids."

 

[ExploderWizard]

In the real world, psychopaths are very good at pretending to be normal. All these interviews you see afterward where the neighbour says, "It's funny, he didn't seem like an axe-wielding maniac when I spoke to him over the fence. And he was great with the kids."

While this is true, having too many instances of psychotic townsfolk just happening to crop up will mean that what would normally be considered paranoid behavior by the PC's will become simple common sense and behaving like a normal person will seem downright suicidal.

 

[Dausuul]

IMHO, it's poor AI programming with the priority on "give them a crack fix of action" rather than "make countries behave in their own best interest". But I guess they were just going for something different than I was. I have to assume "their king must have gone bonkers" to suspend my own disbelief and keep playing.

Now, now, give the programmers a break. You can recognize that the AI's behavior is stupid, and if the programmers were looking at your screen they could recognize it too, but teaching the AI to recognize when it's being stupid is much more difficult.

The "AI randomly declares war" and "AI launches ineffectual attacks" problems are endemic in computer strategy games because it's insanely hard to program a computer to navigate a complex, fuzzy situation like international diplomacy (even simulated international diplomacy). See ExploderWizard's comment:

The rules are general guidelines that serve well enough until they start to produce something stupid. No set of rules can forsee every possible situation, so we have GMs to pick up where the rules leave something to be desired.

In a computer game, there's no GM to recognize and correct when the "rules" under which the AI operates are producing absurd results.

You want verisimilitude and intelligent foes, put down your mouse and pick up your d20.

 

[haakon1]

I find the fantasy world to be a more satisfying construction when it doesn't level in lockstep with the PC's.

I agree, in principle. In practice, however, I'm in the habit of setting up adventures where PC's are likely to face stone giants or ancient dragons when they are 1st level. Might be interesting, but I've never tried -- I mostly just run adapted modules that more or less level appropriate.

I do, however, keep the random soldiers, guards, bandits, wolves, etc. the same power level regardless of the PC's level. Once an 8th level party got wolves as their random encounter while traveling in the wilderness, and it was interesting to see how quick and easy it had become -- gave them an idea of their relative power growth against the world.

 

[haakon1]

Now, now, give the programmers a break. You can recognize that the AI's behavior is stupid, and if the programmers were looking at your screen they could recognize it too, but teaching the AI to recognize when it's being stupid is much more difficult.

But computer CAN play chess well. I assume Total War is actually simpler, and the programmers could make the AI's play to win/survive/do well, if they wanted too, but instead, they've got me in the game designer's Skinner box, and they think I want simulated strategy play from the opposing AI's that REALLY works on the unofficial definitions of the balk rule in baseball -- "When in the opinion of the umpire, the game has become boring, he calls a balk", substituting "random unwise attack" for balk.

"Create an opponent capable of defeating Data." That's what I'm looking for from an strategy AI.

Strategy games lose their luster ("the throne room becomes a prison") when you realize the AI is not trying to win, just trying to "keep it interesting". Maybe I should stick to 1:1 scenarios where it's a straight up fight, like Gary Grigsby's World at War?

In a computer game, there's no GM to recognize and correct when the "rules" under which the AI operates are producing absurd results.

You want verisimilitude and intelligent foes, put down your mouse and pick up your d20.

No argument here!

 

[Lanefan]

I do, however, keep the random soldiers, guards, bandits, wolves, etc. the same power level regardless of the PC's level. Once an 8th level party got wolves as their random encounter while traveling in the wilderness, and it was interesting to see how quick and easy it had become -- gave them an idea of their relative power growth against the world.

Or not, if the wolves rolled well and the PCs didn't.

Lan-"weak foes giving the party fits happens amazingly often, in fact"-efan

 

[ExploderWizard]

I agree, in principle. In practice, however, I'm in the habit of setting up adventures where PC's are likely to face stone giants or ancient dragons when they are 1st level. Might be interesting, but I've never tried -- I mostly just run adapted modules that more or less level appropriate.

I don't generally center adventure objectives around foes far too difficult to defeat. There will sometimes be adversaries that might be really tough to fight and combat could be costly. There will be ways to mitigate the difficulty by thinking players though.

For the world at large I let the terrain, proximity to civilization, and other conditions determine what types of creatures are where. Adventurers moving through these areas run the risk of encountering something trivial, challenging, or outright deadly depending upon where they go and random chance.

 

[Dausuul]

But computer CAN play chess well. I assume Total War is actually simpler...

Hell no! Most computer strategy sims are VASTLY more complex than chess. Chess is a simple decision tree, practically custom-made for computers to be good at. In principle, you could write a very simple algorithm, throw enough CPU cycles at it, and get an infallible chessmaster that always makes the best possible move in any situation. (In practice, of course, you'd need more CPU cycles than there are atoms in the universe. But the point is that chess is "solvable," which makes it easy to program for.)

In a strategy sim, there are far more variables and possible moves. And there's hidden information, random elements, multiple opponents... all manner of fuzzy hard-to-pin-down stuff that makes a poor computer's head hurt, while we humans juggle it with ease.

The reason chess seems "more complex" than Total War is that chess requires a human to think like a computer, while Total War requires a computer to think like a human. So we, being human, think chess is the harder game.

Plus, computer chess games have had a whole lot of extremely smart people working for decades on improving their play. And those people are working from a theoretical background (evaluating whether knight + bishop is worth more than rook + pawn, that kind of thing) that's been built up over centuries. You can't expect a couple of programmers working on a deadline, on a game whose rules are in flux from day to day, to compare to that.