Defenders require bad AI from monsters

[Rechan]

It also depends on the monster.

Monsters with no survival instinct are not going to run. Zombies, golems, etc - its in their "programming" so to speak to fight to the death, because they're neither smart enough to understand being outmatched, and they have no instinct to prompt them to flee.

An intelligent monster who's boxed in (has to deal with several opportunity attacks and is slower than the party) just has no chance of running on foot away. Running is suicide. If it's intelligent, and it realizes this, then it might fight to the death out of spite. The ol 'I"ll take one of these punks, or go down swinging.'

An important thing to note is that some monsters might not run, but simply charge right past the defender to get to something weaker. Either because that monster just likes beating on weaklings, or it understands attacking the fighter is a futile effort. In fact try this: the monster makes a move action to get away from the fighter. Cue combat superiority. Then the monster uses its standard action to charge. Or, monster shifts back, cue combat challenge, then uses its standard to charge. Sure, it's taking an attack, and a -2 to hit, but it's going to be much more threatening.

This also goes for the paladin's challenge. Some monsters sure can take the damage (this goes double for elites/solos). The only monsters that should be afraid of the Paladin's challenge are undead, who are going to take radiant damage.

 

[Elric]

Secondly, how the hell would you know you are at 5% odds of winning? Or at 50% for that matter? Without an extended observation of the opponent, you can't tell.

How can an ork tell at a glance whether a warrior is +6 or +10 to hit? Whether the wizard is capable of just magic missile and minor tricks like most low level wizards NPC or whether this guy is capable of casting flaming sphere and ruin your day?

It can't, at least not reliably. But that doesn't mean it thinks it will definitely win the fight if it has (approximately) rational expectations but doesn't have much information. As I said before,

All that matters is the monster's expectations. Just because the monsters have no information about the party doesn't mean they think they'll win. If the monsters have approximately rational expectations, then they'll sometimes be overconfident and will sometimes be underconfident, but those will average out so that they're approximately correct on average. So the monster's actual chance of victory might be 1%, and sometimes the monsters will think it's higher than this, and sometimes they'll think it's lower.

Given that the monsters know there is a possibility that they'll lose a fight if the five of them charge at the invaders of their hall, they'll generally go get reinforcements a few rooms back and take the invaders on 15 to 5 instead of 5-5.

So it's okay for adventurers to be incapable to realize that Shadowfell Keep is too tough for them but monsters should know when a party of adventurers outclass them?

No, that was actually the exact point of anecdote. You can justify the monsters being overconfident about easily killing those five adventurers in the entrance hall if a bunch of groups of five adventurers have come in and been easily dispatched before. But this isn't a satisfying rationalist explanation, because while the monsters now have correct expectations, the previous adventurers must have been incredibly stupid.

For example, there are lots of identical groups of PC Level-6 characters in the world going around (stupidly) attacking the Shadowfell Keep that the PCs are currently attacking; the monsters win those fights almost every time without needing backup. When the PCs come in, the monsters assume it's the usual truly wimpy dudes, and don't want to split the treasure, so they charge at the PCs. You can see that explanation is a little strained, because while the monsters had reasonable expectations, this require many other adventurers out there to be incredibly dumb.

So... in your world guards are just for show? They figure that if someone comes, they must be stronger than them and they run?

They’d probably go to get reinforcements almost every time. If their only options were “run, without getting reinforcements” and “fight”, then we’d end up in a perfect Bayesian Nash equilibrium where some weak parties attacked, all strong parties attacked, and the defenders sometimes fought them and sometimes ran away, such that the weak parties are indifferent between attacking and not, and the defenders are indifferent between fighting and running (this assumes most parties are “weak”, where weak means weak relative to the kobolds).

This is essentially example #2 in Frank Karsten’s great article on game theory applied to magic: the gathering.

History has taught us that those who starts fight are hardly always capable of seeing it through. So no, just because a bunch of adventurers crashed the monster's door, they shouldn't automatically expect that they are outmatched and run. Odds are very good that the last time the same thing happened, the adventurers got their ass kicked. If it wasn't the case, the monsters would have been booted from their lair already.

This is the “there are other dumb adventurers out there, which is why the monsters are justly overconfident” story I mentioned above. Again, it’s not really a rationalist explanation as mentioned above. This isn’t to say that non-rationalist explanations are wrong. Rational expectations itself is quite implausible; rationality in general is a useful approximation.

But it shows that if you start from a strong premise of rationality, it’s hard to get the preponderance of “fight without getting reinforcements” that D&D assumes is typical. I don’t think this is a serious problem; D&D isn’t trying to simulate an entire fantasy world, after all. It’s just a game that uses multiple encounters a day, so things play out to give you multiple encounters per day.

 

[Rechan]

It's also worth noting that in many circumstances, the PCs are invading a dungeon, which is the monster's home. Most humans will fight vehemently, even to the death, to defend their homes. See: cornered rat.

And, I rarely see this done, but animals will fight to the death to defend their young. Or at the very least, will attack unrelentingly. I can't recall the last time I've seen a monster attack PCs due to the PCs being too close to the young, but dang, now I want to do it.

 

[Danceofmasks]

Well, think about the issue from the monsters' point of view.

Since losing fights is likely fatal, in their experience, they've won all their fights (or close to), and approach fights expecting to win .. just like the PCs.

Now, in our world, game theory says you ought to run. 'cos if the:
War goes against you, what good are you going to do?
War goes for you, why risk yourself when your side would have won anyway?
War is a close one, everybody loses.

But in D&D, one monster will look across at his childhood buddy, and can estimate that he has about 5 times as many hp.

 

[Mustrum_Ridcully]

War goes for you, why risk yourself when your side would have won anyway?

Who fights might lose.
Who doesn't fight lost already.

No, that was actually the exact point of anecdote. You can justify the monsters being overconfident about easily killing those five adventurers in the entrance hall if a bunch of groups of five adventurers have come in and been easily dispatched before. But this isn't a satisfying rationalist explanation, because while the monsters now have correct expectations, the previous adventurers must have been incredibly stupid.

Why did they have to be stupid? People do not run around with their character sheets neatly flying over them explaining how much of a fight they can put up.

The adventurers might have stumbled into the monsters on accident - they got surprised, and were killed.
Another adventuring group might have know that there were monsters to expect, and prepared themselves well. Unfortunately, that still wasn't enough, and they lost, too.
At this point, the monsters are fairly confident that they can beat what is coming to them. They also have risked their lives twice for their lair, which means they have a strong psychological investment in it.

And now a bunch of heroic PCs come around that just looks like another adventuring group, trying again to drive them out? They know how to deal with pesky intruders. They might have even prepared a few new tricks!

 

[Bumamgar]

In my POL-based world the monsters are used to running the show, so they go into pretty much every encounter expecting victory - PCs or not these are the same humans and dwarves they've been devouring for years. To use your analogy, the PCs will likely be the 1st real "party" to have ever raided Shadowfell Keep.

This is dead on.

In 4e DnD, the adventurers are the HEROES. In fact, they are the ONLY heroes in the entire world. The whole world literally does revolve around them, and there is no expectation that the monsters and NPCs have ever encountered anything like them before. These are the future gods and immortal rulers of the world, so that should give you an idea of how rare they really are.

Even the 'adventurers' who the monsters killed in the past who provided them with all that loot, weren't heroes like the PCs.

 

[GoLu]

I think the best rule is that the players shouldn't think the DM is making dumb decisions to help them out, or making convoluted decisions to screw them over. As long as the monster behavior is making sense, and the group is having fun, whether the defender gets triggered or not doesn't seem to matter. If you default to any one position every time, I think it would be boring and predictable for your players, and that's bad.

This is exactly my take on it. Monsters don't have to make tactically perfect decisions. They just have to be clever enough (yet also foolish enough) to seem plausible to the players. 'Suicidal lemmings' is just as inappropriate a style as 'omniscient hive mind'.

For for getting reinforcements... sure. Let them do it. It's clever, plausible, and entirely reasonable. As it turns out, that's a trick I've seen in a few 4e adventures already. A single encounter might be several rooms, and the monsters will get reinforcements from the nearby rooms. Alternatively, an initial encounter might have a monster flee and try to bring in a second encounter. I believe the infamous Irontooth was one of those.

 

[Mal Malenkirk]

No, that was actually the exact point of anecdote. You can justify the monsters being overconfident about easily killing those five adventurers in the entrance hall if a bunch of groups of five adventurers have come in and been easily dispatched before. But this isn't a satisfying rationalist explanation, because while the monsters now have correct expectations, the previous adventurers must have been incredibly stupid.

Do you have any idea how many stupid people there are in the world?

We've been warring since the dawn of time and 80% of the time the result of the battle was a foregone conclusion. How moronic do you think the Egyptian's leadership was to start the six day war, in hindsight?

Even the Axis in WWII can be seen as a bunch of stupid idiots from that angle. They purposefully started a fight they couldn't really win. Surprise and tactical brillance got them a few years where they seemed on top of the world and provided for some of the most harrowing battle in history, sure. But the bottom line is that they had willingly picked a fight with an assortment of countries that grossly outmatched them in terms of industrial production and manpower. Germany, Italy and Japan vs the rest of the world... Do you have any idea how much more soldiers and tank the allies could produce in the long run? Which is exactly why after a series of stunning victories, they still hadn't won the war and then got their ass kicked in the second half.

It's just like a D&D band of monsters ambushing a bunch of PCs, thinking surprise will make this an easy fight but without realizing that the opposition has far more staying power than they suspect. When they think the battle should be over, it is just beginning. Looks pretty stupid in hindsight, but that's what happened all the time in history.

It's like that line from Gladiator when Maximus and his aide are looking at the assembled barbarians. It goes something like:

-''Don't they know that they are beaten?''
-''Does any man know when he is beaten? Would we?''

And then they proceed to slaughter the barbarians. The roman empire did execute countless slaughter like this one at its height. These were one-sided affairs that should never have occured by your logic.

They’d probably go to get reinforcements almost every time. If their only options were “run, without getting reinforcements” and “fight”, then we’d end up in a perfect Bayesian Nash equilibrium where some weak parties attacked, all strong parties attacked, and the defenders sometimes fought them and sometimes ran away, such that the weak parties are indifferent between attacking and not, and the defenders are indifferent between fighting and running (this assumes most parties are “weak”, where weak means weak relative to the kobolds).

That's why I said fight and start the alarm.

Beside, all those game theory is balderdash once the fight start.

It's the kind of things leaders consider before deciding to fight or not. And as demonstrated by history, the leaders still often choose what seems like stupidity. In their defense they usually have imperfect informations and in D&D's specific case the PCs are oftens the one who starts the fight anyway. Beside, just because one has the quality to win leadership doesn't mean one has the quality required to lead. Two very different things.

The troops on the ground, they think of nothing at all. There is no game theory when PCs crash the gate; There is only fight or flight and these monsters have been conditionned to fight. It's that simple. No soldier in history has ever considered game theory when being shot at. He shoots back; that's what he's trained to do. His field commander might order a reatreat before long, but even he will need some time to assess he situation (remember that the PCs tends to be fewer than the monsters, it skews evaluation) because he hasn't been trained to give ground at the first challenge either. There is a reason he is occupying that area and you don't concede it lightly. For example; you don't concede the gate!

 

[Elric]

And now a bunch of heroic PCs come around that just looks like another adventuring group, trying again to drive them out? They know how to deal with pesky intruders. They might have even prepared a few new tricks!

Like not standing in different rooms in groups of 5 so that the PCs can beat encounters 2-4 with 5 minutes rest breaks in between. Instead, after the first group of guards at least sounds the alarm, they could get into one group, and crush the adventurers 15-5 when they ventured into the next big room of Kobold Hall.

Do you have any idea how many stupid people there are in the world?
...
And then they proceed to slaughter the barbarians. The roman empire did execute countless slaughter like those at its height, one-sided affairs that should never have occured by your logic.

Indeed, it can be fairly hard to come up with rationalist explanations for war, as I said in my first post on the subject. That doesn't imply that there aren't any non-rationalist explanations for war, and some of these wars can also be explained with rationalist explanations.

That's why I said fight and start the alarm.

So the monsters sound the alarm or send at least one of their number to get reinforcements. However, the PCs are able to take a short rest (5 minutes!) after the first fight in Kobold Hall, move on the the second fight, take a short rest (5 more minutes!), move on to the third fight, take a short rest (5 minutes!), move to the fourth fight, take a short rest (5 minutes!), and then confront the final fight. That's a lot of time in which encounters 2-4 could have decided to stop waiting in separate rooms to get beaten one by one.

This is dead on.

In 4e DnD, the adventurers are the HEROES. In fact, they are the ONLY heroes in the entire world. The whole world literally does revolve around them, and there is no expectation that the monsters and NPCs have ever encountered anything like them before. These are the future gods and immortal rulers of the world, so that should give you an idea of how rare they really are.

Even the 'adventurers' who the monsters killed in the past who provided them with all that loot, weren't heroes like the PCs.

Even if the adventurers are the only "heroes", low levels PCs clearly are not much more powerful than humanoids of their race who regularly engage in combat. Compare a (Players-Handbook) level 1 Dragonborn Fighter to the Dragonborn Soldier (level 5) in the MM, for example. Monsters shouldn't be so overconfident that they don't bother to get together in a large group to defeat the invaders after an initial alarm is sounded/someone goes to get reinforcments.

Imagine trying to play Kobold Hall with the DM as the adventurers and the PCs playing all of the kobolds. Do you think you'd end up with 5 distinct encounters, (5 minutes apart each!), or would the PCs mass their forces after an alarm/going to get reinforcements?

 

[Mal Malenkirk]

So the monsters sound the alarm or send at least one of their number to get reinforcements. However, the PCs are able to take a short rest (5 minutes!)...

Well, I'll give you this one. The five minute rest is a new 4e conceit that I don't really use.

My PCs rest for all of 30 seconds and that's if it take that long to reach the next encounter.

Basically, once the hectic fight is over, they check temselves for wounds, take a deep breath (and regain encounter powers!) and then resume hunting the monsters. If they haven't successfully killed everyone before they could sound the alarm, they run into monsters in the next room and the twenty seconds delay since the end of the previous fight barely explains why they had the time to grab a weapon and gather in a more sensible formation.

A four encounter dungeon like, say, the Room of Eyes in the Thunderspire labyrinth would probably be cleaned out in about three minutes to four minutes of real world time.

Imagine a bunch of commandos sweeping a terrorists hide out.