Do your DMs have raiders and larger predators behave intelligently?

[Amphetryon]

In my (possibly jaded) opinion and experience, having your NPC adversaries behave intelligently brings out the cries of 'Killer GM' and 'metagaming,' while having enemies attack stupidly and ignore tactical opportunities to flee or otherwise improve their odds brings out the cries of 'kid gloves' and 'forgot to read the Evil Overlord list.'

Pick which one you'd prefer to have the Players complain about.

[Slipperychicken]

Like the old Sid Meier's Pirates game where you actively wanted to bite off more than you could chew before you divided up the plunder so that you could get as large a share of the kitty as possible for your score.

Which if you like playing with the fourth wall, could be interesting if you pulled it off well, I suppose. Probably a bit niche both for those who'd want to pull such a move and those who'd have the necessary metaphysics that they'd want to use.

It makes a decent amount of sense from an in-character perspective, though, if crewmen are cheap and easy to hire (even after you get scores of them killed), their skills and experience are irrelevant, and you're extremely confident of your ability to bleed crew members while still winning fights.

Of course, it depends on how the loot is distributed. I know that sometimes it would work like this: the captain gets a fixed percentage (i.e. 26% of all loot, after expenses), first mate gets a smaller fixed percent, normal sailors get a fixed percent distributed between them (i.e. 30% of loot is distributed across all crewmen), and so on. Sometimes the captain gets loot measured in shares, which would encourage him to cut his underlings out of the equation (reducing the number of shares, thereby improving his own shares' value) when he could get away with it.

[Morty]

What does that have to do with anything? Whatever their leader plans, people whose motivation is to easily get rich by robbing helpless people are not going to stay cool and calculating while their comrades die. They're at least as likely to fear that it might be them who go down next rather as they are to be glad that there's people to split loot with.

[Knaight]

What does that have to do with anything? Whatever their leader plans, people whose motivation is to easily get rich by robbing helpless people are not going to stay cool and calculating while their comrades die. They're at least as likely to fear that it might be them who go down next rather as they are to be glad that there's people to split loot with.

Exactly this. The casual indifference only really makes sense in the context of those who have strong reason to believe that they will be among the surviving party. If, for instance, take a group of about 40 bandits, involving 15 hardened veterans with pretty decent arms and armor, and another 25 or so desperate peasants turned to banditry with some cheap weapons, no real armor, and minimal combat training. The 25 probably have some turnover - 40 might go into an attack, with about 35 coming out pretty routinely. Among the 25, this is a problem, and pretty terrible for morale, as the 5 dead are almost always coming entirely from them. Among the 15 veterans? There's really not that much cause for concern, the life is clearly working for them.

There's also the case of delusion helping here. When one of the 15 die, the other 14 can all convince themselves that they were never all that capable in the first place, and that they always properly belonged more to the disposable group that wasn't too great. It's the "sure, they died, but that couldn't happen to me" attitude. Conversely, more competent and better equipped members of the 25 might think that they are one of the hardened veterans, and as such not see the five compatriots of their group dying as cause for concern. This is particularly true as the line between them is likely to be pretty fuzzy, particularly as the peasant bandits that last longer start getting better equipment as battlefield salvage and pick up some experience.

[Slipperychicken]

What does that have to do with anything? Whatever their leader plans, people whose motivation is to easily get rich by robbing helpless people are not going to stay cool and calculating while their comrades die. They're at least as likely to fear that it might be them who go down next rather as they are to be glad that there's people to split loot with.

Notice how I mentioned this bit:

It makes a decent amount of sense from an in-character perspective, though, if crewmen are cheap and easy to hire (even after you get scores of them killed), their skills and experience are irrelevant, and you're extremely confident of your ability to bleed crew members while still winning fights.

If that condition is not satisfied, then needlessly getting crew members killed in battle does not make sense.

[Knaight]

If that condition is not satisfied, then needlessly getting crew members killed in battle does not make sense.

If that condition is satisfied, the crew is still going to be unhappy about it. Which causes problems.

[Rhynn]

If that condition is satisfied, the crew is still going to be unhappy about it. Which causes problems.

Especially given that pirate ships were democracies, and the captain only had (theoretically) absolute authority during battle.

The same probably generally applied for bandit bands, especially given that historrically many of them were just mercenary companies with no employer...

A leader who has even one disaster (such as 20% of the group getting killed in one day) is probably going to get replaced in a hurry, if they're lucky...

[NichG]

The thing to keep in mind is that members of a bandit gang are also a form of loot for the leader. You have to weigh whatever gains acquired from a risky attack that resulted in the death of one of the gang against all future gains that gang member might have enabled you to pull off. So 'damn the survival rate, full speed ahead' behavior is going to dominate more in people who are looking for that 'one big score', because the future gains are irrelevant, but people who are going after smaller prizes on a regular basis are going to be much more risk averse.

The cost of actually replacing a member is likely to be quite large, because if you have to hire one then you're paying them - possibly per job rather than based on a share - for their strength (whereas the initial group of bandits were probably gathered by looking for people in desperate circumstances and will just take a minor share), and if you have to wait for another appropriate person to come along then you're at reduced strength for that interval (which correspondingly means you can't make raids that are quite as big).

For example, a bandit gang living in the vicinity of a city of 2000 people. If the gang has 20 members that's a full 1% of the city's population, so there probably isn't a big pool of replacements waiting in the wings to hire into their losses, otherwise they'd be called a 'military' instead of a 'bandit gang'. If they're distributed demographically then they basically would have something like 11 Lv1s, 5 Lv2s, 2 Lv3s, and a Lv4 leader. If they just use an extreme show of force to scare non-local travelers into paying a 20gp toll to enter the city, the merchants will probably pay up without too much trouble (thats cheaper than hiring the guards necessary to fight off the bandits, even at the low wage rates in the DMG for hiring NPCs). Lets assume they're particularly smart, and they ignore travelers who are obviously well-armed or have anything significantly magical on them. If they get one traveler a day (which means that the city would roughly have a 15% temporary resident fraction, probably on the small side), thats enough for everyone in the bandit gang to earn a gold piece a day, which puts them well above peasant wages, and for very little actual work.

The total CR of the bandit gang is 10 if they all fight at once, but since thats significantly higher than the CR of the individual members its failing to account for certain problems that tend to emerge at large CR gaps (lack of things like flight or ways to deal with magic). Still, lets assume they decide to go after a 'big score' - a group of four Lv4 PCs. This is before flight becomes an issue, and they're a CR+6 encounter for the group so they have some chance despite the inherent slant in the system to favor PCs (better ability scores, better wealth, etc). They can probably expect to lose something like 6 of the Lv1 guys in the fight, and possibly even some of the higher level gang members depending on how the PCs focus fire (if the leader is obvious, expect him to go down in the first round due to standard PC tactics when dealing with ragtag groups). What they get in exchange is about 11000gp in cash after selling off the PCs' gear. This is certainly better than 20gp/day but actually its just about a year and a half's total income. And of course, this is assuming that none of the PCs escape, use up their wealth in consumables, etc. And that the transition from charging tolls to killing heroes of the realm doesn't create an arms race with the local power center. And that they didn't misjudge the CR of their victims, or pick on people who had some spell they didn't have an answer for, or whatever.

So if you have a bandit group that has been operating in an area for more than two years, its probably the slow and steady kind.

[Slipperychicken]

If that condition is satisfied, the crew is still going to be unhappy about it. Which causes problems.

Also notice when I said "when he could get away with it".

It makes a decent amount of sense from an in-character perspective, though, if crewmen are cheap and easy to hire (even after you get scores of them killed), their skills and experience are irrelevant, and you're extremely confident of your ability to bleed crew members while still winning fights.

Of course, it depends on how the loot is distributed. I know that sometimes it would work like this: the captain gets a fixed percentage (i.e. 26% of all loot, after expenses), first mate gets a smaller fixed percent, normal sailors get a fixed percent distributed between them (i.e. 30% of loot is distributed across all crewmen), and so on. Sometimes the captain gets loot measured in shares, which would encourage him to cut his underlings out of the equation (reducing the number of shares, thereby improving his own shares' value) when he could get away with it.

You guys need to work on your Spot checks. You keep missing all my qualifiers.

[dps]

What does that have to do with anything? Whatever their leader plans, people whose motivation is to easily get rich by robbing helpless people are not going to stay cool and calculating while their comrades die. They're at least as likely to fear that it might be them who go down next rather as they are to be glad that there's people to split loot with.

Which, if you want to talk realism, brings up another problem. Bandits aren't going to be attacking an adventuring party anyway--they're going to be going for easier targets.

[Morty]

For example, a bandit gang living in the vicinity of a city of 2000 people. If the gang has 20 members that's a full 1% of the city's population, so there probably isn't a big pool of replacements waiting in the wings to hire into their losses, otherwise they'd be called a 'military' instead of a 'bandit gang'.

That's silly. Everyone knows that in fantasy lands, gangs of bandits, brigands, raiders and other such criminals are at least as numerous as the people they prey on.

Which, if you want to talk realism, brings up another problem. Bandits aren't going to be attacking an adventuring party anyway--they're going to be going for easier targets.

Yes, there's that too. Still, adventurers are likely to bring the fight to them.

You guys need to work on your Spot checks. You keep missing all my qualifiers.

We notice them, they're just not relevant. Whether or not the leader is fine with losing people is a secondary concern - the primary concern is that when a band of heavily armed, magical murder-hobos starts cutting their comrades down in droves, members of a criminal gang are going to leg it.

[ElenionAncalima]

My problem with the - What if the leader is a sociopath who knows he will survive and wants the others to die to get more money - argument is that it is just one possible scenario. That one possible scenario doesn't justify having all bandits fighting to the last man in every scenario. It doesn't even justify why the minions would continue fighting.

Regarding the OPs original point, I think it can be a good thing to run enemies intelligently...but the real key is to run them interestingly. We have seen numerous bandit scenarios proposed in this thread:
-A sociopathic leader who has hired disposable minions
-Group with hardened veterans and a rotating roster of new recruits.
-A group that preys on the weak and got in over their heads by attacking a group that was stronger than them.
-A close knit group of bandits with a fraternal bond.
-A recently elected leader who needs to win the approval of his fellow bandits.

I think it is a mistake to decide that one of these is the way that bandits work. Anyone of those scenarios, if roleplayed well, would be far more interesting than fighting 10 nameless, faceless bandits, killing them all, taking their stuff and moving on with your travels.

[Rhynn]

My problem with the - What if the leader is a sociopath who knows he will survive and wants the others to die to get more money - argument is that it is just one possible scenario. That one possible scenario doesn't justify having all bandits fighting to the last man in every scenario. It doesn't even justify why the minions would continue fighting.

Yeah, it's all basically an irrelevant side-argument that has nothing to do with the fact that, as a rule, neither bandits nor predators* are willing to sustain extensive casualties or great personal injury, because it runs counter to their primary motivation, i.e. survival.

* Or really anyone who's not a crazy monster, some kind of berserker, a constuct, magically compulsed, exceptionally disciplined under specific orders, etc. etc. - a bunch of side cases and exceptions that have no bearing on the point.