Role-playing Games Stack Exchange is a question and answer site for gamemasters and players of tabletop, paper-and-pencil role-playing games. It only takes a minute to sign up.
Sign up to join this community
I want to build a cool devil overlord as a villain for a group of three level 1 PCs.
How would I go about making a villain who's more powerful than them without going overboard?
Should I be scared to make it incredibly hard since there are only 3 party members?
Do I make a villain like I make regular D&D characters?
Or should I try and stick to the book mostly since I am new?
If you meant the BBEG (Big Bad Evil Guy - the main Villain of the Campaign), I will start with some frame challenging:
A BBEG that is going to be defeated by level 1 characters is probably underwhelming. That's an animated armor or bugbear level of threat. Anything your players can beat with just their lvl 1 features and spells probably isn't going to be a Villain level of enemy.
So, just make usual encounters. If you don't know how, it's explained on p. 81 of DMG or p. 56 of the free DM's Basic Rules PDF. From my experience the difficulties are easier than what they say - deadly usually means having to spend some resources (class features and spell slots), but unless your players are unlucky, deadly shouldn't be actually deadly.
Let them get some levels. By level 3 to 5 they might be able to defeat something that actually could be called a BBEG.
In the comments, goodguy5 suggested to introduce the Villain without combat. He gave two examples that you can read below, but I'll use the one from an official adventure - Strahd von Zarovich from CoS.
He is introduced almost as soon as the adventure starts, but he is only supposed to be fought by level 10+. He shows up in person some times, either to scare adventurers or to play them (or without them being the reason he shows up). If they engage combat against Strahd, he will charm them, make them fight each other and might try to turn one into a vampire, but he will not just kill them, because that would be unfun for him.
You can get some monsters, as the Bugbear mentioned, and give them some lore. If you can afford it, Lost Mines of Phandelver does exactly that and might teach you some things (the Starter Set is amazing for that).
Make the Animated Armor a guardian or a cursed soul. Make the Dryad corrupted by the destruction of a forest. Anyway, take CR 1 (at max CR 2) monsters and give them a background, a motivation, a story. That should be enough for 1st lvl.
If you want to create a new monster, p. 273 from DMG explains how to do it and calculate the CR in order to keep it balanced.
The DMG gives clear guidelines as to what is a balanced encounter to any party size and levels. You probably shouldn't have a BBEG that is defeatable by 1st level characters, but a "Boss" can be made taking usual CR1 monsters and giving them lore.
AFAIK, there is no guideline to what CR a PC would be (they use CR = PC when needed, but this is just gross, if a CR1 monster is supposed to be a decent encounter against FOUR level 1 PCs, there is no way a level 1 PC is equivalent to a CR 1 monster). Without experience, you probably shouldn't try to make a PC-like villain, as it is easy to make it overpowered or underpowered.
1 Related: How do I present an unbeatable encounter without frustrating my players? - Years later, I found out that, sometimes, presenting a strong villain early on (specifically Strahd, as in my answer here) may lead to player frustration because they may feel powerless. Still, I would like to mention that I have done it several times before and this was my first time getting this feedback, so YMMV, and I believe this answer is still useful.
In addition to the more technical answers, I'd point out that a villain doesn't have to be physically impressive. An enemy can be dangerous without a sword or spell (say, by poisoning a city's water supply). They can be tough to defeat without being physically imposing (say, by hiding in the sewers, protected by clever traps).
This is also an opportunity to create a nemesis at the point of origin. So he's not a "boss" yet. Maybe he is easily defeated, or his final trap/golem is the boss fight. The point is, he gets away, and you can bring him back in a future adventure, once he and the party have leveled enough to have anything resembling a boss fight.
Do I make a villian like I make regular D&D characters?
You do not, creating monsters is a totally different subset of rules that can be found in the DMG page 273. I've tried digging around in the SRD, and the rules for making monsters aren't there, so you'll really have to buy or borrow the DMG if you want to make monsters from scratch.
All your questions can be answered by reading the DMG, particularly the following:
Without a DMG, there's not much you can do about creating a creature from scratch.
Fortunately, there are several available creatures online that you can use. Just take a look at these NPC-villainables from the SRD, and choose one that is of Challenge Rating 1.
Two answers.
You don't need an anthropomorphic villain
Not all threats worthy of adventurers are creatures. In fact, some of the more believable low-level setups pit the PC's against an object, or a system.
Your villain can be defeated sideways
Perhaps the villain is a CR10 villain - much too strong for a party of newbies. This is, in fact, why he's the villain - nobody in the land is strong enough to oppose him. No city guardsmen, no street thugs, not even the most experienced huntsmen can stand up to him (maybe some have tried, and you can make it clear they failed spectacularly). This is the One-Punch Man of your part of the world.
But HP and AC are meaningless in the face of most ways a person might die. Such as;
There are lots of ways to kill villains stronger than you - and depending on your group, this could be an excellent way to introduce your players to non-mechanics-based thinking.
Ideally, a creature with a CR (challenge rating) of 1, should be handleable by 4 characters of first level. Thus would be challenging for 3.
The simplest method, would be to find a CR1 creature in a Monster Manual that is close to what you want class wise. Then, you can reskin it as the Villain you want.
For example:
In Volo's Guide there is a Kobold Scale Sorcerer (CR1) that could be used as a template for a 3rd level Mage.
Some useful online tools:
As other have stated; for level 1, you don't have a villain. Or more to the point, you don't have A villain.
Specifically, at level 1, the characters are just getting their bearing in the world. The chance of TPK is great with one role of the dice.
A true villain should give the air of "everything built up to this". How much build up are you going to have at level 1? Is the baker that sells cursed cakes to certain customers as revenge a "villain", or just a "bad guy"?
Instead you want to build a series of hurdles the players will go through. I gave an example in this question about bounty boards. The characters will do something simple that leads to deeper and deeper intrigue until they can find the mastermind that set everything in motion; the doppelganger that is morphed to be a king, or a devil that created a lawful religious order whose annual rites (and being performed soon) are actually a ritual to open a gateway to Hell.
Things always starts off small, but with the victory over the minor the characters have the skills and confidence to go after something worse.
The most important thing is to make sure the "villian" is not deadly, but merely dangerous. So your villain should not do a direct confrontation ("I leave you to my minions"), but instead they should have a plot that threatens either the party or a vested interest to the party.
Doing this serves two purposes; first, you can make the final "boss" fight level appropriate for the party. Second, the villain always gets away and can be a real villain in the future.
Not sure if this warrants my own post, but just bringing together what others have said...
There are two aspects to the original posters question:
For the first question it is as daze413 and ravery pointed out:
It is much simpler to stick with the books rather than creating your own monster which can be difficult to balance, though the DM's guide does have guidelines on this. Having said that, it isn't too difficult to 're-skin' an existing monster into something that seems entirely different to what's in the book (change the appearance, change a 'fire' attack into an 'acid' attack and so on, and you can easily change, say, a fire elemental into an acid blob monster!).
For the second question:
The DM's guide (page 81) has guidelines for a combat encounter as daze413 also points out. I would add that the Wizards Unearthed Arcana site also posted up an alternate way of calculating combat encounters (which I consider to be much simpler). Find this here.
Also, as Furiant points out, there is more to creating a memorable dangerous villain than just combat. There are plenty of ways a villain can be present, gloating about his goals while the PCs battle his minions (while occasionally lashing out with a spell or whatever to show exactly how dangerous he is). Such a villain will, of course, flee before the PCs get to him. (But also be prepared for your PCs to show surprising resourcefulness and defeat your 'undefeatable' villain! Always be prepared to change your carefully laid-out plots!).
To have a good villain, the players must really dislike the person, i.e. one needs to build up strong feelings. To do this one needs a) time for build-up and b) a reason for dislike.
The dislike can be come by through evil deeds. The DMG (p. 94 to 96) has tables with plentiful ideas for this. One very important point to consider is that the villain must have motivation to do what they do. Mysterious reasons or pure villainy are very counterproductive. Even for an archdevil, having him having been thrown out by a superior archdevil is a big step up because it delivers reasons why they sacrifice peasants (feed on their souls and acquire power to get revenge). The motivation should not be clear from the beginning of play, but it should be from the beginning of design.
Build-up needs time and the relevant pieces need to be introduced fast. That means that the evil deeds need to be obvious right-on. The villain also should be introduced directly except not as villain. This gives the players time to get to know them and their reasons for doing things. The fact that they are the culprit should not be revealed until much later, however. Reading literature on the creation of mysteries is probably helpful here. Pokémon Stadium for example introduced the villain early as simple NPC:
The villain is the nice mayor of Phenac, introduced early but revealed as villain much later.
Note that contrary to other opinions expressed here, a villain needs to be somewhat human to make the story interesting. This does not mean that an animal or an object cannot be a villain, but it has to behave like a human to be interesting. Further Reading: How to write a damn good novel by James N. Frey. A dragon or ghost or demon can easily be outfitted with feelings and goals like a human would have them. The same is – at least in theory – true for diseases and natural catastrophies, but it seems exceedingly difficult. Therefore, these cannot, a priori replace a villain.
Specific to a level 1 adventures: The scope In the description of the Tiers of Play (PHB p. 15), the scope of adventures for specific levels are suggested. For the first tier of play (including level 1) the description states, inter alia:
The threats they face are relatively minor, usually posing a danger to local farmsteads or villages.
The villain’s plot should be relevant in the respective scope. Such as the archdevil abducting the local farmers’ kids and planning to burn the whole village down in the final act. The scope of a story’s conflicts is usually not decisive for the story’s quality: Many people find many stories other than Dan Brown’s Inferno very interesting.
There a multiple techniques that can both make the final encounter more interesting and move the challenge away from pure combat. The former is relevant for all adventures, the latter is especially relevant for level 1:
There are some important aspects to make interesting fights. The most important are an interesting battlefield and the usage of multiple monsters. Both concepts have abundant literature on the internet. Also none of this is specifically relevant to level 1.
Of note though is the fact that putting together multiple monsters in one encounter is difficult with monsters from the MM and even with the rules for creating new monsters from the DMG; especially so at level 1. Therefore, other stat tables should be used that specify how many damage, hp etc. monsters should have if you want to use multiple. That way, the villain can make up part of the encounter while having a fierce monster or some minions. The monsters / opponents should be highly varied. Multiple identical monsters do not make a combat that interesting.
Literature with such stat tables are a) monster design guidelines by the Angry GM [Warning: very thorough article and contains strong language]: https://theangrygm.com/f-cr-theres-a-better-way-part-2/ and b) 5E monster maker by Giffyglyph [Very long document]: https://giffyglyph.com/monstermaker/
Specific design strategy: blocking standard tactics One way to make encounters more interesting is to block the usual go-to tactics of the players from time to time. If everyone always uses longbows and rapiers, give some monsters in the encounter a piercing immunity. This should not be overdone, however, and it is important that other tactical options exist. Since level one characters are fragile and do not have that many features, this technique needs to be used cautiously in such cases.
Specifically important to level 1: The standard deviation To have a challenging encounter at level one without risking an unlucky crit killing a character, the standard deviation on the encounter outcome needs to be kept low. In most roleplaying games, including 5E there is a standard assumption that the player characters will win and that the monsters / opponents are the underdogs. Increasing the standard deviation on the outcome favors the underdog, i.e. it works against the players in this case. The following article talks about football but explains the general concept well: http://archive.advancedfootballanalytics.com/2009/05/are-nfl-coaches-too-timid.html
Since level 1 characters have low HP they are especially vulnerable to outliers such as crits. Reducing standard deviation is therefore vital. To do this, the monster stats need to set suitably. For the offensive stats, this means high attack modifiers, relatively low damage, small damage dice, and potentially multiattack.
Calculations use anydice.com. For a CR 1 monster, 9 to 14 damage with a modifier of +3 is appropriate (DMG p. 274). Using a d20 for damage would give an average of 10.5 damage with a standard deviation of 5.77. Using 4d4 instead would give an average of 10 with a standard deviation of 2.24. Using the d20, we have a 5 % chance of rolling a 20 which is probably enough to kill most level 1 characters outright. Using the 4d4 (which is a very similar average damage), we can get at most 16 damage, which is less likely to kill a character outright. Also, the probability in this case to get more than 14 is less than 2.5 % as opposed to 30 % for a d20. Of course, getting rid of damage dice entirely and using static damage is even better in this perspective.
Another useful option is increasing the attack modifier and decreasing damage. Using the DMG table, a +5 modifier and 6 to 8 damage is also valid and will further decrease the standard deviation. Doing this will make the monster hit more often, dealing about 7 damage, which is more consistent than missing more often but dealing 12 damage on a hit.
Multiattack does, a priori decrease standard deviation because it decreases the chance that all or none of the attacks hit. It does have its dark side, though, since it increases the chances for critical hits, which are very problematic on first level.
Concerning the defense, AC should be kept low. Lower AC and higher HP decrease the standard deviation. Also, hitting the opponent gives a sense of progress which is important. The MM contains 400+ monsters of which 2 are below level 5 and have more than 17 AC. Using 15 or even 13 AC seems even more reasonable though.
Disclaimer: I am writing a different answer because I believe this answer goes on the completely opposite way of my original answer. The reason for this new answer is a bounty asking explicitly for answers that are not frame challenges. I still believe my original answer is the best I can give to the original asker problem and to most DMs having this problem, especially if they are new (as in the case of the question). I believe the answer below is a good solution if you are hard on the restraint of 1st level.
Furthermore, I will challenge the frame posed from the bounty itself before I answer it properly: If you are running a one shot, I strongly recommend that you start at least at 3rd level. In 5e, at 1st level, most classes don't even have their defining features or any combat options besides "I attack", and, maybe, sometimes, "I help my ally who deals more damage than me to attack (with advantage)". Take the Paladin as an example - the Paladin has 5 points of Lay on Hands, and, maybe somewhat useful, his Divine Sense. He does not get his Divine Smite, Spellcasting or Fighting Style until 2nd level. A 1st level paladin is a bad fighter with a holy cross in his shield.
This isn't even the problem when making a villain encounter, it is a problem through the whole session. In my experience, 1st level is great to build up characters for a long campaign and set a mood, but awful for stopping your play there. There is a reason you level so fast to 2nd level.
But now, what if you and your players really, really, really want to play a 1st level one shot?
Then...
Through this answer, I will give an example we actually ran, although at 3rd level, while tackling on the problems specific to 1st level and how to handle them, and presenting the general idea behind it. The setup was the following: A bunch of evil Drow cultists were performing a sacrificial Ritual to Summon the Goddess Lolth into the Material Plane.
Among the villains performing the ritual, we had a Priestess which CR would probably be around 17 - certainly deadly to a 3rd level party.
The goal in this encounter is not, and can not be, defeating the Priestess. That is ultimately an impossible goal. A good encounter is a clash of motivations, goals, objectives. (See this angry DM article about it - Disclaimer: as usual for Angry DM, lots of "strong words"). In my example, the Priestess wants the Ritual to succeed, while the adventurers want to stop it. It is not a one-on-one combat, and victory or defeat are not defined by one side dying.
So, during the whole encounter, the Priestess is ultimately busy with the Ritual. If attacked, she would probably cast Shield and keep focusing on the Ritual. If she ever stopped to attack the party, the Ritual would fail and her lifelong plans would go to dust - meaning the party has obtained victory, maybe at the cost of their lives, but a small price to pay for world's salvation. In the specific case of a one shot, as long as the players are okay with it, the ultimate resolution may involve the characters' death, and can be satisfying, although perhaps bittersweet, to know that the characters were martyrs that saved the world. So, this is my first point: since it is a one shot, do not fear characters' death so much.
Extending this line of thought, defeat is an acceptable outcome. Defeat happens. While D&D is mostly built around the assumption that the characters will win, it is also built around the assumptions that they will level and grow. If you are willing to throw away one of these assumptions, you may as well throw the other.
Obviously, the Priestess was not alone on her cult, and certainly would not perform the ritual alone and without anyone to help against potential intruders. This is where the characters can get level-appropriate encounters: through the minions of the villain. Given the Drow thematic, we had Giant Spiders and, in the setting, Goblins are essentially Drow's slaves, so we had a bunch of goblinoids to kill as well.
While the motivation belongs to the villain, their minions are effectively the ones that are fighting the party in the clash of motivations. In the example, the minions had to bring body sacrifices to a table, and perform a bunch of stuff I would prefer not to describe here, while fighting the party. After some time, the party understood that a major structure in the room was required for the ritual, and attacking that structure would delay or even stop the ritual. The minions are the ones responsible for stopping the party from freely attacking the structure.
Just make sure the minions are an appropriate challenge to the party. Here enters the first problem of a 1st level party: an appropriate minion challenge is, still, a little underwhelming. The powerful priestess has, as her minions... Four goblins. But if you and your party can run with that, then go for it. Otherwise, you can improve the looks of it. Perhaps the priestess has an army of 50 goblins, but 48 of them are too busy helping with the ritual, they are too brainless to understand that the characters are a threat and they are more scared about stopping the ritual and being slaughtered by their Drow overlord than they are of the characters.
As I mentioned previously, in my example, the ritual structure was a major part of the encounter. This already provides a somewhat meaningful decision for the players: should they focus on the minions or on the structure? This is even more important for characters at 1st level. They have very few possible decisions to make within their own features and character, so your encounter must provide meaningful decisions for the players. A 11th Wizard can think for minutes on what spell it should cast, so his features by themselves provide that, but a 1st level has Magic Missile and Sleep, and probably he is spamming Fire Bolt because he only has 2 spell slots anyway.
One way to solve this problem is to make the environment meaningfully interactive. This is very common in video games, where the way to defeat an unreachable boss is through the environment. In my example, the environment is a proxy to the villain, i.e., it represents the HP of the encounter. You can also make the environment a tool, say, by inserting flammable barrels that will explode in a Fireball-like way when touched by fire (the fighter can do that, riskly, with a torch, or the wizard can do that safely with a fire bolt). It may, as well, be a defensive tool. In 5e, spells require line of sight to be cast, and you have full cover if there is no line of sight between you and the center of an AoE spell. Maybe the powerful Villains indicates his intent to cast a Fireball in the center of the room, but the party has enough time (and clues) to hide behind the pillars. Such mechanics are, again, very common in video games, and can be used to provide meaningful decisions to the encounter.
Important consideration: if the environment is the proxy for the HP of the encounter, make sure the players understand it. One of the major flaws in the encounter when we ran it was that the players took a long time to understand the structure was meaningful, and that attacking it had any effect. Describe how the minions are clearly trying to protect that structure, how the villain gets annoyed by the party hitting it, anyway, convey this information the best you can. Same thing for using the environment as a tool. Make sure it is clear to the players that the barrels and the pillars are useful, not just a flavorful part of the scenario description which can be forgotten.
There are many ways the encounter can end. As previously mentioned, maybe the characters die and the villain succeeds. Awww, it sucks - but can happen, especially in a one shot.
The other way is through the characters accomplishing their goal. But now what?? The Villain could, now, full of anger, just kill the characters, with their plans frustrated. Unlike my previous scenario, where the villain is forced to give up his goal in order to kill the characters (therefore, their deaths have meaning - it is what saved the world), now the death feels a lot more frustrating, because it seems meaningless. They already had completed their goal, if they die now, it is for nothing. Sure, you can rationalize that as a consequence of saving the world, but it does not feel good. What can you do?
I am getting about 5 downvotes for actually suggesting this, but a Deus Ex Machina is a literary tool and can be used. Even Tolkien has his Eagles. Some deus exes may even be well built enough that they do not feel like a deus ex machina. In my example, the deus ex machina was a powerful (allied) Wizard sensing the disruption of the ritual structure and teleporting the party out of there. If you want a more conclusive mean for the example, you could
Say that the priestess is extremely exhausted from performing the ritual and does not have the strength or spirit to defeat the party, being an easy kill.
Lolth herself, who was half-summoned, gets frustrated with her servant failure and punishes her with death, before being banished away back to her plane.
Just be creative and as coherent as you can. If the party already has the fulfilling emotion of having their goal accomplished, from my experience, they will not get too bothered with a small deus ex machina to avoid the characters downfall.
So, this is my 50 cents on encounter design for a 1st level party. It still is not a conventional encounter or villain1, but, from my experience and in my opinion, is an engaging encounter that will let everyone have fun, which is the main goal as far as I am concerned.
1 Final Final Comment: By the way, you may notice that the threat (an incredibly powerful priestess summoning a goddess to the world) is considerably above the pay grade of 1st level adventurers, reason I say this villain is not conventional. I do not care. Both the Hobbit and Lord of the Rings are about a bunch of nobodies with one or two magic items defeating ancient dragons, one or two ancestral godlike angels and an infinite army of evil creatures, and D&D is literally founded on the basis of Tolkien's works, so do not tell me 1st level adventurers should be restrained to a Thief in The Shire.
There are extensive rules and plenty of good answers here regarding how to create monsters and villains in general, so instead I'm going to speak directly about the unique challenges of making a villain for level 1 players, compared to making a villain in general.
To get straight to the point, the problem is, well, hit points. Level 1 players don't have many of them. Even regular enemies stand a good chance of knocking down a level 1 player with a lucky crit or two. Having an enemy that deals even more damage will become overwhelming, fast. Having only a single villain making a low number of attacks makes this even worse, since in order to be a threat they will essentially be taking out a player every turn.
With that in mind, I would suggest that instead of increasing the damage of the villain compared to a normal monster, instead add status effects. Knocking players prone, knocking them away, blinding them temporarily, poisoning them, all of these things can help make the villain feel more dangerous than it actually is.
As an example, let's say that you have your players fighting some raiding barbarians, and as a capstone they'll fight the leader of the group, a half-ogre named Hask. At CR 1, a half-ogre would normally be a fairly easy fight for a group of 3-4 level 1 players. However, even it's baseline attack deals 14 damage on average, which is already more than some level 1 characters can handle. Increasing that damage even more could change those one-hit KOs into one-hit kills, not something we want.
Instead of killing players or even taking them out of the fight permanently, we want to make Hask's attacks feel more dangerous with status effects. There's a lot of statuses that we could use, but for a big strong bruiser like Hask, I think stunned makes the most sense. So, let's swap out Hask's battleaxe for a battle hammer, and add a rider to it: On hit, the target makes a DC 13 Con saving throw. If they fail, they are stunned for one round.
Just that isn't enough to make Hask a real threat yet though. With only one attack per round, the players will simply overwhelm him. He needs to have at least two attacks per round, but that presents a problem. If Hask hits the same player twice, there's a good chance he kills them outright, and there's really no reason he wouldn't, especially if he just stunned someone with his first attack.
However, we can actually solve this problem by leaning into Hask's large, brutish nature. We can "protect" the players by adding another rider to his attacks: knocking players back. No need for a saving throw on this one, simply say that anyone he hits is knocked back 20 feet and falls prone. Now if Hask hits someone with his first attack, it makes perfect sense that he would follow up by attacking someone else, since his initial target has been thrown away. It's entirely possible that the player hit will come back in swinging next round, but they'll have had a turn to respond and make their choice, so that's fine. If they get hit again, that's on them.
For higher level villains, I'd suggest adding in more things, such as legendary actions, legendary saves, lair actions, etc. However, for a low-level villain it's better to keep it simple. Do the above, and double Hask's HP to 60 to help him survive a couple rounds, and you've got a good level 1 capstone fight.
A useful way of thinking about it (that brings together much of what was said by others, but frames it differently) is that you don't need a CR 1 villain, so much as you need a villain of any level that creates a CR 1 problem the players need to deal with.
The best villain IMHO is one that is out of reach of the players for awhile, but for whatever reason has schemes they can dismantle that points them towards another of the villain's schemes. After gaining levels doing that a couple of times, the villain will begin to notice them, which should fill them with a nice sense of dread. By the time they reach the appropriate level and finally face off against the villain, they'll feel like they've really earned it.
To specifically address the issue of fighting a CR1 Devil opponent, I'd say your best bet is an Imp - but even then the Magic Resistance ability common to most Devils may sideline your party's spellcaster(s), and they'll be unlikely to be able to counter its' invisibility.
Let the players create the villain.
Present them with the usual set of encounters: bandits, orcs, goblins whatever. Make sure someone, anyone survives; have them swear undying revenge against the player characters.
A few sessions later, let them discover that the last set of challenges were actually set up by someone. They investigate and find out that "that guy" from the first session was actually behind it. At this time "that guy" is fully aware that the PCs will murder them in actual combat, so he instantly flees only stopping to taunt from somewhere safe.
Maybe a few more proxy murder attempts, then the villain makes some sort of desperate measure: demon pact, mass-murder-to-make-zombies, drinks a potion of vampirism; something to graduate from annoyance to villain. All he wants is to kill the PCs, if the rest of the world bites it as well, that's fine.
Keeping someone alive isn't that hard, it can be anyone. Maybe one of the goblins only plays dead, or one of the orcs is out hunting. One of the bandit captives were romantically involved with one of the bandits; it takes a special kind of PC to murder the just-freed girl who cries over one of the bandit corpses.
The important part here is that the players notice the survivor:
As the characters leave the goblin ambush, they hear something, look back and notice a shape rising, they notice the near mortal wound and missing hand; and he runs away. (Later the scars provide recognition.)
When they plunder the orc encampment, they notice an orc returning to it. Everyone stares at each other for a minute, until the orc breaks and runs. Someone comments on how unusual it is for orcs to have blue eyes.
Try to roleplay everyone, some characters stick, you find them fun and the story flows from there. Just keep your eyes open for potential stories.
You could make a low-attack, high health villain that uses an escape skill when they get low on health, and then reappears later in the campaign as a higher level character. The "high health" could come from the Wizard class's "Shield" spell and related defensive spells, and the low attack could come from the Wizard class's 1d4 attack damage.
