Developing the Sandbox: The Campaign Calendar

This post is part of my series on developing the sandbox. You can read other posts in the series here. The series will, when necessary, go into detail on the development of my sandbox for Dungeon Module B1: In Search of the Unknown, but will also discuss sandbox development more generally. You can read play reports from my B1 campaign here.

A man of disputable wisdom once said that "YOU CAN NOT HAVE A MEANINGFUL CAMPAIGN IF STRICT TIME RECORDS ARE NOT KEPT." This has since become a bit of a meme in the TTRPG community. It is an emphatic declarative statement written in all caps by a larger than life figure in the hobby whom a lot of people enjoy ridiculing, often for good reason. But is there truth to it?

Kinda! Honestly, I've actually never kept a calendar for any of my D&D games. They still felt pretty meaningful to me. Nonetheless, at the end of my longest-running campaign ever (a three and a half year 5e game that took the player characters from 3rd to 20th level and beyond), I felt myself regretting not having a calendar. 

How long had it taken for the characters to make their ascent to power? How long had they known each other? When did they first meet that henchman? How many seasons, moon's turns, and holidays had passed them by? I didn't know.

I was running the campaign by the seat of my pants. I randomized the weather every day. It was an NPC's birthday or the day of some festival when I needed it to be. If a PC became a lycanthrope, only then would I bother to figure out when the next full moon was. Dungeons were restocked, new monsters moved into the region, and factions moved against one another and towards their goals when I felt like they should.

It worked for me. I didn't need a calendar, but I felt myself wanting one all the same.

One of the benefits of running a small sandbox (like the one I created for B1) is that you stand a reasonable chance of, at some point (possibly even before play begins), actually finishing prepping the whole thing for play. Technically this is possible for larger sandboxes as well, but it's a more Herculean task, and you are more likely to burn out before play even begins or to fall behind during play (if, like me, your interest waxes and wanes naturally as the campaign goes on), likely resulting in a lesser quality experience for you and your players.

The sandbox is never actually finished, because running the sandbox is its own thing that requires maintenance between sessions. The benefit of having most of your sandbox prepped and ready to go is that you have more time between sessions for those more ancillary pursuits. When you're not rushing to detail the owlbear den the players intend to clear out on their next expedition, you can instead spend that time restocking the dungeon, updating your outdoor encounter tables, and figuring out if those NPCs succeeded in their most recent scheme.

This is also a great time to work on your calendar. Now, the calendar might be one of those things you want to have prepared ahead of play. Perhaps you want to generate a month's worth of weather. That's pretty reasonable. But I've run nearly two dozen sessions of my B1 campaign, and I'm only just now feeling the need to create one of these bad boys.

What made me want a campaign calendar? Well, part of it was that I just thought it would be neat. That's reason enough. Part of it was remembering the regret I felt at the end of past campaigns. A big part of it was that the party started to accumulate an absurd number of NPC hirelings and paid them in advance, and the way that I was tracking the length of their employment was totally impractical. 

So, I made a calendar.

If you like, you can go on rootring and search "calendar", read all the advice, and come up with something that works for you. Like making a dungeon, there is tons of advice on this subject, and it's impossible to summarize without conducting a meta analysis.

Here's what I did. I got a calendar template from Google Sheets (I used the first option, which is just a sheet for each month of the year and a cell for each day of the month). This is a 2026 Earth calendar. There are no fancy names for the months of the year or the days of the week. You can do this if you want, but it isn't important to me, so I didn't do it for now (I remember really liking the vanilla fantasy names for the month presented by Shamus Young on his Twenty Sided blog...21 years ago?!).

First, I went back through my play reports and figured out when the campaign began. I had already decided that the campaign began in Spring, because that seemed like a reasonable time when adventurers would embark on adventuring activities. For simplicity, Spring is March, April, and May, Summer is June, July, and August, Fall is September, October, and November, and Winter is December, January, and February (this will be important when determining the weather).

I looked at when my first session took place in real life - March 12th. Perfect! I decided that was the in-game start of the campaign as well. Then, I went through my play reports, looking for any mention of the party resting or time passing. I also noted when the party brought on their hirelings and for how long. I ended up with this:

After 20 sessions and 6 months of play, just 13 days had passed in game. Wow! You could go more in depth, noting which sessions occurred during which days and what specific activities the party engaged in - and I might do that still - but it felt unnecessary at the moment.

Since one of the main reasons I wished to start work on the calendar was to track the length of NPCs' employment, I went into the future on the calendar and marked the days when each hirelings' employment would expire. At this rate, it will be another 5 or 6 sessions before the party's porters' and mercenaries' contracts need to be renewed, and 18 or 19 sessions before the crew of the party's river barge can retire, but it's useful information to know. Time passes weirdly during a campaign. A single day might be several sessions, and the next session might cover several days.

I also liked the idea of pre-rolling weather for a month or so. In the real world, you can often tell when it's likely to rain in the coming days, and I want to be able to foreshadow that (the players might not want to embark from town if it seems like they're going to get a torrential rainstorm the next day). It also allows me to skip rolling for weather at the table, and makes me less likely to forget to do so. So, I rolled weather for the rest of the month (as well as the following month, since we're nearing the end of the current one).

I used these hex flowers to determine the weather:

I like these because they're evocative, they only require a single roll for each day, they're based on the seasons, and each day's weather affects the next day's weather. Depending on your system of choice and how weather mechanically impacts things like overland travel or encounter chances, you'll likely have to do some interpretation of the results, but it works reasonably well. In addition to marking daily weather on my calendar, I noted if the weather imposed penalties on overland movement or modified outdoor encounter chances for that day.

Since I am using a real-world calendar, I could also look up what the weather was actually like on each day in some area of the real world comparable to my setting, but in practice I found the information I was looking for too difficult to find, and the information I did find too fiddly for my uses. It would be fun to use a system where the weather can change throughout the day, as it does in real life, but I personally don't need that.

Another thing I thought would be fun and necessary to track is the phases of the moon. This information can be based on the real world and is pretty easy to find, but you can also just make it up. Pick a day during your first month that the moon is full, then have it be full again every 29 days or whatever. Or decide that full moons in your setting are more or less common if you like. You'll have to come up with that yourself. You can also track other celestial events like eclipses and supermoons the same way.

Why is the moon important? Well, don't tell my players, but my sandbox has lycanthropes in it. I want to know when there's increased lycanthrope activity. The player characters might in turn become lycanthropes, and I want to know when they're going to transform and wreak havoc, or do whatever it is werebears do (for example). The AD&D DMG has this whole table breaking down the percentage chance of a cursed individual transforming into a werebeast based on the current phase of the moon:

Clearly you will need to accurately track the phases of the moon to use this thing properly! Also note that you will need to know for how long the character has been a lycanthrope. Another win for the campaign calendar!

You might also decide that the local Evil temple does some nefarious stuff during the lunar eclipse or that magic is more powerful during that time. That seems like a fun application of this tech.

Another fun thing you can do with your calendar is track settlement events like holidays, festivals, NPC birthdays, or whatever. If the players plan for their characters to spend a week in town, it would be nice if they knew that it was the friendly wizard NPC's birthday so that they can arrange to get her a gift. If they're planning to venture out into the wilderness but they learn that tomorrow is the annual Cheesemonger Parade, they might decide to stick around an extra day to attend the festivities and rub elbows with important contacts in the cheese industry.

I roll a d20 for each day on my calendar, and a result of 18+ is some day of significance in the settlement. I like this table from the 2014 5e DMG:

This is one of the many unsung tables of that oft-maligned book, which is honestly full of useful little stuff like this. You'll have to ignore some stuff - seasonal events like equinoxes and solstices, midsummer/winter festivals, fertility/harvest festivals, and celestial events which you're already tracking by other means - but overall this is pretty good stuff!

I wouldn't bother determining events like this for every settlement in the sandbox. I simply rolled for the largest settlement - the one the PCs are most likely to spend extended amounts of time in. You can probably safely assume that the villages, hamlets, and thorps which surround your town celebrate more or less the same events as the town itself.

There's also no reason to limit this to settlements only. Do the denizens of the dungeon celebrate holidays? Is there some mutually recognized armistice day where they come together and parade through the dungeon? Do monsters in their lairs celebrate birthdays? Is a tournament being held at the fighter's stronghold?

You might also want larger events going on in the background - wars and famines and natural disasters and stuff. Joel at Silverarm Press wrote a post which highlights a neat table from uh...Oriental Adventures, on which you can roll for a single yearly event which in turn influences which monthly events are likely to occur. This is pretty neat tech. I haven't tried it out yet, but it's worth a look if you want like, a war happening somewhere "off screen" (or maybe even on it). Warren D. of I Cast Light! also reworked the table on his blog.

Just as you might do with weather or the phases of the moon, you can also pull yearly/monthly events from real life. Simply pick a real-world year that corresponds roughly to the period during which your campaign takes place, Google it, and look at the Wikipedia article to see what happened that year.

The calendar is also useful for tracking faction events. I know that the wizard that rules my town has a tenuous truce with the cleric who rules her own stronghold. The wizard also has a mutually friendly working relationship with a subordinate wizard who wishes to build a stronghold of her own. The cleric wants to kill the subordinate wizard. Will the cleric achieve her goals? Will the truce deteriorate? When does that happen? You will probably want some way of running factions.

You might also use the calendar as a simple timer. One of my biggest bugbears in adventure design is when some time-limited event is happening in the setting (a monster is growing in power, a city is being overrun with plague, or a bad guy is working on a nasty ritual), and the conditions of those ongoing events don't change unless the player characters intervene somehow. 

There's something to be said of a "powder keg" scenario where everything exists in a tenuous equilibrium, requiring some "X factor" (the player characters) to move things decisively in one direction or another, but that doesn't mean everything has to remain static. If some event is going to occur if given enough time to gestate, you should track it somehow. No more Big Bad Evil Guys waiting until the player characters show up to complete their rituals!

There's a party of adventurers camped out in the wilderness, clearing land to build their stronghold. When do they finish? The wizard in town is plotting against them, trying to recruit a band of ogres to root them out. When does he get the ogres to act? There are brigands who have usurped a stronghold and keep its rightful ruler prisoner. When do they execute her? Set dates for all of these things on your calendar. If the player characters take actions to aid or thwart these plots, adjust the timeline on your calendar.

These are the broad strokes. The campaign calendar is of near infinite utility, and there are many things both big and small which it can be used to keep track of. Here are some examples of what else you might track:

• When does a disease or parasitic infestation run its course or kill the afflicted?
• When do followers arrive at a player character's newly built stronghold?
• When does an NPC spy complete a mission and, if the mission fails, when can the characters safely employ another spy on the same mission?
• When does a slow poison kill?
• When does a character or NPC finish crafting an item (magical or otherwise)?
• When does a sage find the answer to a question (or discover that the question is unanswerable) and when can another question be asked of the sage? When does the sage finish acquiring a new minor or major field of study?
• When do henchmen respond to recruitment attempts?
• When does an unfaithful cleric finish atoning for their sins?
• When must an Evil cleric reassert control over their undead minions?
• When does an injured player character make a full recovery?
• When do effects like psionically induced comas, magical enchantments, or temporary insanities end?
• When does a character who mixed an oil of slipperiness with an oil of etherealness return from being lost in the Ethereal Plane? 
• When is a soul no longer trapped in an amulet of life protection, when does a character finish reading a manual of bodily health, and when can a character again invoke the Goat of Terror, and other magical item effects?
• When does an assassin finish their study of poisons?
• When does a character finish their training and level up?
• When is construction of a font or basin for holy water completed?
• When is spell research completed?
• When does an Evil artifact inflict its malevolent effects upon its possessor?
• When has a character known a hireling or henchman long enough to warrant a loyalty score increase?
• When does a character finish training a griffon as a mount and when do they finish training to fire a bow from its back?
• When does the next peasant rebellion occur? When does the rebellion evolve into a civil war?
• When do the characters have to pay taxes?
• When does a cleric or magic-user finish creating a golem?
• When is a paladin's warhorse or a conjured demon released from service?
• When is construction of a stronghold (and mazelike dungeons thereunder) completed?
• When does a character die of old age?

(This list was created by combing through the AD&D DMG with Ctrl+F, keywords "days", "weeks", "months", and "years". There are undoubtedly a great number of many other things you can track as well!)

The calendar is not just for the DM's benefit. Just as players can master a dungeon's layout and use their knowledge of such to plan optimal routes for exploration, treasure extraction, and avoidance or ambushing monsters, they can master the calendar and use it to their benefit. If the DM knows when it will rain, when the hirelings' contracts expire, and when the NPC's birthday is, this should be telegraphed to the players in some way so that they can make informed choices.

Do they wait another day in town to let the weather pass? Do they prioritize their next expedition, before their hirelings demand more money? Do they make sure to be back in town so they can attend the wizard's birthday? There's a full moon tomorrow night. Should they plan to patrol the streets of town and lay a trap for the werewolf? When are the owlbears out hunting versus sleeping in their lair? What can they do before the royal ball to improve their chances of getting an invite?

These are all more or less things you will have to track as a result of player actions - they are less about developing the sandbox than they are about running it. You can wait until your game is underway and until you absolutely need the calendar before creating it (as, admittedly, I did), but hopefully I have demonstrated why creating the calendar upfront can be beneficial as well.

Creating a Sandbox for Pokémodule B1: In Search of the Unown

Hey, I haven't created a sandbox on the blog in a while. I just finished up Pokémodule B1: In Search of the Unown, an Into the Oddish conversion of Dungeon Module B1. I made a sandbox for my version of B1 (which I've been running in AD&D for a few months now), so why not make one for the Pokémon version?

Now, I could just take my existing sandbox for B1 and convert it to Into the Oddish, but that's no fun. I love making sandboxes, and am always looking for an excuse to do so. So, let's make a new one.

The first issue I have to address is how the sandbox should feel. Do I want to design this like a Pokémon game, with predefined, somewhat linear routes with little patches of tall grass where wild Pokémon are encountered and which you can sometimes navigate around? Should every forest and cave have an entrance and exit, connecting different regions of the map? Should my towns and cities be designed like settlements in a Pokémon game? Or do I want to design this like a typical AD&D sandbox but replace all the AD&D stuff with Pokémon stuff?

My approach to creating content for Into the Oddish has been to treat it not like a TTRPG that feels like a Pokémon game, but like a Pokémon game that feels like a TTRPG. What I mean by that is that Into the Oddish already isn't like a Pokémon game. You aren't a Pokémon trainer catching Pokémon and battling other trainers. You're a wild Pokémon (or a left-behind trained Pokémon) exploring the ruins of a society. That's just D&D but everything is Pokémon now.

Overland travel, outdoor encounters, and dungeons play the same, but instead of seeing a dragon flying overhead in the distance, it's a Charizard (or a gang of teenaged Charizards). The Charizard might still come down and incinerate you or demand you surrender your Pokédollars or TMs, just as a dragon might demand gold or magic items. There are still dungeons, but they're abandoned gyms, power plants, mansions, ghost ships, etc. inhabited by wild Pokémon instead of monsters. A "stronghold" is just a fortified location that's been commandeered by some singularly powerful Pokémon (I really like the idea of "strongholds" as standalone wilderness "gyms"). When I run a forest in Into the Oddish, I won't run it like in a Pokémon game - it's like in a D&D game, but all the goblins are Caterpie.

I am also always harping on how AD&D has these super versatile worldbuilding tools that you can fit into whichever TTRPG you personally prefer. Usually I'm talking about D&D-likes, but I bet it can apply just as well to Into the Oddish. Let's find out, eh?

I'll start with a small regional map, 5 hexes by 5 hexes. Just like with my original B1 sandbox, I'll assume that the region is predominantly a cold climate forest. Using some variation of Welsh Piper's Hex-Based Campaign Design I get the following:

Next I drawn in topographical lines, which lets me chart the course of rivers like so:

More rivers is always better, so I added a few coming in from off map in addition to those flowing from the depicted hills. Water-type Pokémon can presumably swim along the waterways, and those who know Surf can carry others with them. Of course, I don't see any reason why Pokémon can't hitch a ride on a boat the same way adventurers in a D&D game would.

I know that Quaqsqueton is on a wooded crag, and there's only one wooded hills hex (03.02), so it has to go there. Since that hex is already stocked, I'll roll 24d100 for the remaining hexes. I use a random number generator that doesn't allow duplicates and get this:

That's two single dwellings (2 and 3), a thorp (4), a stronghold (13), and two ruins (15 and 16). I'll need some additional info about the stronghold and the ruins, so I roll 2d100 for the stronghold and d100 for each ruin. I get 89 and 71 for the stronghold (large walled castle with keep, character-types), and 92 and 24 for the two ruins (a tomb and a shrine). I won't interpret those rolls literally, but I at least know that the "stronghold" is large and is ruled by a powerful Pokémon, and that neither of the ruins are cities. 

Now I can place each location on the map:

My key as of now:

• 01.03: Single dwelling
• 01.04: Large stronghold, character-type
• 01.05: Tomb ruins
• 02.03: Single dwelling
• 03.02: Quagsqueton
• 04.04: Village ruins
• 05.04: Thorp

While Pokémon presumably don't build roads, there are of course roads (or routes) connecting locations in the Pokémon world. So, I add roads:

At a glance, it looks like the two single dwellings pay homage to the ruler of the stronghold, who in turn controls the road to the tomb ruins. There is a road through the forest, but to travel from one end to the other, one must pass through the village ruins. Just as in my original B1 sandbox there is no road to Quasqueton because of its intended remoteness (in defiance of the guidance in the AD&D DMG), there is no road to Quagsqueton on this map. The player Pokémon have come into possession of some special map which points to the dungeon's long lost location, allowing them alone to exploit it (for now).

Next, I place wild Pokémon lairs. Just as I do with AD&D, I'll roll d10 for each hex which doesn't already contain a location or a road. Plains and water hexes have a 10% chance to contain a lair, light forest and hills have a 20% chance, and heavy forest has a 30% chance. I roll up a lair in the fourth plains hex (03.01), one in the first forest hex (02.04), and one in the heavy forest hex (04.05). My map looks like this:

Now, how do we make all of this work for a Pokémon game? 

I'm thinking the tomb ruins could be something like the Pokémon Tower in Pokémon Red and Blue, while the ruined "village" is like, the ruins of your classic Pokémon starter town - Old Pallet Town or something like that. If I had rolled a ruined city instead, that would be more like a city in a Pokémon game, with an old gym, a dilapidated Pokémon Center and Pokémart, maybe the ruins of a Silph Co. building or something like that.

The "stronghold" is much like a stronghold in D&D - a seat of power for some high-level Pokémon. It is probably a "gym" of sorts. If this were AD&D, I could just roll up an NPC using the tables provided and riff on that. This isn't AD&D, so I'll have to get creative.

One of the first things I did when converting B1 to Into the Oddish was to sort each of the 151 Generation I Pokémon into "levels" 1 through 10 (according to their average base stats), the same way monsters in AD&D are sorted. This allows me to easily convert an existing D&D module by switching a monster of a given "level" out for a Pokémon of a similar level. Since Into the Oddish's Pokémon statistics are derived from their game statistics, this works quite well. The list I made for Into the Oddish is here. This is what I used when converting B1.

Since a Pokémon gym leader's "ace" Pokémon is usually a fully evolved form or a Pokémon that does not evolve in its Generation (not universally true, but true at least for Kanto, which is all I particularly care about), that seems like a good rule of thumb to go by. Looking at my list, the lowest level candidates are Level 4 Pokémon - Beedrill, Butterfree, Farfetch'd, Lickitung, and Porygon. Conveniently, this corresponds to what a "medium level" NPC is in AD&D (level 1d4+3, or 4 to 7).

Normally when I determine the level of some random NPC in AD&D, I roll d20. 1 to 13 is levels 1 to 3 (1d3), 14 to 17 is levels 4 to 7 (1d4+3), 18 to 19 is levels 7 to 12 (1d6+6), and 20 is levels 9 to 20 (1d12+8). So I can just exclude the first category and instead roll a d7 (weird number, but it's fine). Since this is a large stronghold (as compared to a small or medium one), and because I want something cooler than a Lickitung, I'll roll 2d7 and take the highest result (for a small stronghold I would take the lower result, and for a medium stronghold I would simply roll once). I ultimately get a 6 (the other roll was a 5, so the result would be a Pokémon level 7 to 12 regardless).

I roll d6+6 and get 12. Any result of 10 or higher is a level 10 Pokémon. I look at my table and roll d14 and get an Exeggutor, which is great for a forest but maybe a bit weird for a cold climate. That's fine though. Pokémon doesn't make a lot of sense if you think about it too hard. I can easily change this to a tropical climate if I want to.

Stronghold rulers in AD&D have 1d4+1 henchmen of a level equal to their level divided by three (rounded to the nearest whole number), which is doubled if the NPC is above 8th level. Since my Exeggcutor is technically 12th level in AD&D terms, their henchmen Pokémon can be up to 8th level. I go looking for suitable Pokémon (I decide this is a Grass-type gym rather than a Psychic-type gym) and get a Gloom, a Parasect, a Vileplume, a Tangela, and a Victreebel. This is similar to how Pokémon gym leaders have lesser Pokémon the player must defeat before facing their ace.

I can do something similar to determine what important Pokémon inhabit each settlement, determining character-type NPCs (in AD&D terms) as usual and simply switching them out for like-level Pokémon. The only tricky part of this is that AD&D assumes that 98 to 99% of the population is made up of noncombatants. That leaves you with, at most, around 900 NPCs in your largest city who are "character-type" NPCs, which is still a lot, but you only really need to detail a fraction of those (135 would be high level or very high level, and only four or so would be 20th level, so you could just stick to those as faction leaders and call it a day). 

In contrast, all Pokémon fight. A city of 60,000 would have something like 3,000 (5%) "Tier 4/9th to 20th level" Pokémon, and somewhere around 21,000 (35%) who might be "gym leaders" (4th level and up). That clearly doesn't work. 

There are no towns or cities in this particular sandbox, but I do have a "thorp" which is home to 50 inhabitants. If I just say that 5% of them are "Tier 4" Pokémon, then that means two or three 10th level Pokémon, any one of which could be legendary. That doesn't really make any sense to me. And what of the "single dwellings"? Should each of the 1d12 Pokémon there be diced for, with a 5% chance that one is in the same category as a legendary?

Here's what I'm thinking. Larger settlements like towns and cities are more like the "cities" in the Pokémon games. An AD&D town is like one of the small "cities" in a Pokémon game (Pewter, Cerulean, Vermillion, etc.) and an AD&D city is like one of the big cities in a Pokémon game (i.e., Saffron, Goldenrod). When I prep for AD&D, I break these large settlements into districts, so I'll do the same here. Each district is home to a special location - like a gym, the S.S. Anne, Silph Co., the Safari Zone, etc. Each of those locations can be ruled by a gym leader-type or otherwise prepped like a dungeon with a similar distribution of Pokémon of different levels.

Smaller settlements like hamlets and thorps are Pokémon-formed communities which came together after the collapse, while villages are analogous to small human settlements like Pallet Town. They don't have gyms or Pokémon Centers. The Pokémon who live there are all 1st to 3rd level "rabble" who aren't generally prone to fighting but will defend themselves if they must (you don't want to mess with a community of 50 Pokémon, even if it's just a bunch of Pidgey and Rattata). Since these smaller settlements are essentially one small "district", maybe there's one "gym leader"-type who acts as community leader and informs what types live there but isn't actually a gym leader, since these communities don't have gyms.

A single dwelling might be the isolated former home of a specific trainer, and the relatively small number of Pokémon there will be of the type typically used by that trainer class. Since there are very few Pokémon here, you can really flesh them out and not worry about it being too wonky statistically speaking because who knows what type of trainer or how powerful they were.

Wild Pokémon lairs are easier. I can just roll on the AD&D outdoor encounter tables and switch out whatever monster I roll for a like-level Pokémon. For the light forest lair, I got wolves (3rd level), which I switch out for Cubone. The plains lair is trolls (7th level), which I switch out for Electabuzz. The deep forest lair is giant lynx (4th level), which I switch out for Nidorina/ino. It isn't perfect, but it sort of works and is better for me than just kind of making up whatever. Your mileage may vary.

In this post, I'm only going to bother statting out individual Pokémon - whichever ones are the most powerful in their given location. There isn't any clear guidance for what "level" these Pokémon should be for the sake of determining their movesets, so I kind of just look at what moves they learn before hitting the level at which they evolve and pick the ones I want.

My individual Pokémon also occasionally have TMs and HMs in place of their level up learnset, as well as some healing items. How do I determine who has what? Well, since my Pokémon are analogous to character-type NPCs in AD&D, and AD&D has instructions for determining if such NPCs have magic items, I can use those same procedures. Basically, I use this:

How? Well, if, for example, I'm detailing a "level 4" Nidorina, I'll check for items as if she's a level 4 AD&D NPC. Let's say I get a positive result on the first roll and a negative on the second, which means she has two items from "Table I". I ignore that and just say she has two items. For one of them, I'll look up what TMs and HMs Nidorina can learn and roll a d50 or something until I land on one. If its a move I want her to have, I'll replace some other move she would otherwise know with that, and she'll have one additional single use battle item like a Potion or a Heal. If it's a move I don't want her to have, I'll give her two battle items instead.

Which items? Eh, whatever. If it's a Potion, I'll pick whichever level of Potion seems appropriate for a "4th level" Pokémon (probably a Super Potion). What other item might she have? Well, she's a Poison type, so she's vulnerable to Psychic-type moves which means, because of how Into the Oddish works, she's susceptible to the Confusion Status. So let's give her a Full Heal to cure that in case it happens. A Water type would probably have a Paralyze Heal, a Grass type would have an Antidote, a Burn Heal, or an Ice Heal, and so on.

Oh, and remember, gym leaders always give TMs! So I like to give all my gym leader Pokémon at least one TM move of their specialized type and include that TM as a reward for defeating that Pokémon in battle.

Now let's go through and detail each location, in order.

01.03 FISHERMAN'S CABIN: This riverside cabin and boathouse was once the home of a Fisherman. Since the humans disappeared, the Fisherman's Cloyster, Captain, has established an abode in the river along with other Water-type Pokémon, including five Shellder and two Krabby. They harass any Pokémon attempting to use the road to travel south.

Captain (male Cloyster): STR 14, DEX 11, WIL 11, 9 HP, Armor 3. Water/Ice type. Resists Ice (x2) and Water. Weak to Electric, Fighting, Grass, and Rock. Items: Full Restore, Paralyze Heal. Moves (d20):

• 1-10 Withdraw: Increases his Armor by 1 (up to 5) for the remainder of the encounter.
• 11-15 Bubblebeam: Deals highest 1 of 2d8 Water-type damage to the target. If any damage dice match, the target has Disadvantage to Dexterity saves until they pass a Dexterity save.
• 16 Blizzard: Deals highest 1 of 2d12 Ice-type damage to the target(s) (Blast damage). If any damage dice match, the target suffers the Freeze Status.
• 17-20 Spike Cannon: Deals 1d8 Normal-type damage to the target. On a Critical Hit, if the additional damage die rolls its highest possible value, roll an additional damage die. If that additional damage die rolls its highest possible value, roll an additional damage die. Ad infinitum, if the damage dice keep rolling their highest possible values.

Captain is motivated by greed and self-interest. He is unfriendly towards player Pokémon (-2 to Willpower saves to influence him), demanding a tithe from any who pass through his territory or expect hospitality. He and his Pokémon control passage downriver from this point on. 

Belladonna (01.04) has attempted to recruit Captain against Aegon, but hasn't managed to convince Captain that the odds are in their favor. There is some enmity between Captain and the Pokémon of Tamer's Cabin (02.03) over control of the river, although this has not resulted in bloodshed yet.

01.04 PINECREST LODGE: In the Age of Humans, this location was a secluded resort hotel and lodge for the wealthy looking to get away from the hustle and bustle of the city to relax amongst nature. It is an imposing complex with all the amenities associated with such a place - a grand timber and stone main lodge, guest cabins, hiking and skiing trails, hot springs and a spa, a Pokémon Center and Pokémart, and staff housing and maintenance buildings. Most Pokémon found here are of the Grass type, though Bug, Ground, Poison, Rock, and Water types are found as well.

The lodge's owner and proprietor, Myrtle, was herself a gym leader. Her Exeggutor, Aegon, was her strongest Pokémon. Since the humans disappeared, Aegon has gone somewhat mad and has claimed rulership of the Lodge. He is a capricious despot. One of his heads is greedy and cruel, another is gracious and benevolent, and the third is constantly mediating between the two. He is prone to mood swings and changes of heart, so he inspires great fear and obsequiousness in his subjects.

Aegon is unfriendly (-2 to Willpower saves to influence him) towards player Pokémon, granting them hospitality only if considerable homage is paid. 

He tightly controls the road to the Frost Witch Barrow (01.05), granting access only to those who greatly please him or vow to deliver its valuables to him. He greatly fears the Jynx there and can be convinced to finance expeditions by player Pokémon to roust her from her lair.

Aegon (male Exeggutor): STR 14, DEX 8, WIL 16, 17 HP, Armor 1. Grass/Psychic type. Immune to Ghost. Resists Fighting, Grass, Ground, Electric, Psychic, and Water. Weak to Bug (x2), Fire, Flying, Ice, and Poison. Items: Antidote, Burn Heal, Full Restore, and Ice Heal. Moves (d20):

• 1-7 Hypnosis: Rolls 1d20. On an 8 or higher, the target suffers the Sleep Status.
• 8-14 Stomp: Deals 1d8 Normal-type damage to the target. On a Critical Hit, the target flinches and misses their next turn.
• 15-17 Mega Drain: Deals highest 1 of 2d8 Grass-type damage to the target and restores his own HP by half the damage inflicted on the target.
• 18-20 Psychic: Deals highest 1 of 2d10 Psychic-type damage to the target. On a Critical Hit, the target has Disadvantage to Willpower saves until they pass a Willpower save.

Aegon is served by five "henchmen":

• Moss (male Gloom): Aegon's steward. He is quiet and dutiful. He is the first of Aegon's inner circle that visiting Pokémon are likely to meet. He is fiercely loyal.
• Morel (male Parasect): Master of fungal gardens and Bug Pokémon. He is slow, methodical, and unsettling. He prefers a quiet life and likes Aegon well enough. Belladonna has attempted to recruit him and his Bug Pokémon to rise up against Aegon, but she hasn't convinced him to act.
• Azalea (female Vileplume): Keeper of the conservatory and royal advisor. The oldest after Aegon himself. She is chamberlain and counselor, inclined towards neutral, balanced advice. She attempts to temper Aegon's less desirable qualities and has no great love or hatred for him.
• Briar (female Tangela): Aegon's spymaster. She rarely speaks and can appear almost anywhere by emerging from the overgrowth. She is fiercely loyal to Aegon and is always listening for talk of subversion.
• Belladonna (female Victreebel): Captain of guards. She is power-hungry and quietly marshalling allies against Aegon, but has little in the way of supporters aside from her own subordinates.

If formally challenged by Stage Three Pokémon, Aegon will be joined in battle by these "henchmen", though Belladonna is likely to change sides if the battle isn't going their way. If lesser Pokémon challenge him, he will force them to defeat his henchmen before he accepts the challenge. He grants the Evergreen Badge and the Mega Drain TM to any Pokémon who manage to defeat him.

01.05 FROST WITCH BARROW: A dungeon for Stage One Pokémon. Once a tranquil burial site for trainers visiting Pinecrest Lodge (01.04) to pay their respects to dead Pokémon, it has since been occupied by Sable, a powerful Jynx known to Pokémon in the region as the Frost Witch. It is a cold, lonely, partially flooded, presumably haunted ruin inhabited by Magikarp, Psyduck, Slowpoke, Drowsee, and other 1st through 4th level Pokémon who either enjoy the chilly environs or the dread atmosphere.

Sable (female Jynx): STR 7, DEX 14, WIL 12, 12 HP. Ice/Psychic type. Immune to Ghost. Resists Ice and Psychic. Weak to Bug, Fire, and Rock. Items: Burn Heal. Moves (d20):

• 1-4 Lovely Kiss: Rolls d20. On a 6 or higher, the target suffers the Sleep Status.
• 5-13 Thrash: Deals 1d10 damage to the target, then deals 1d10 damage to the target again on the following turn. After the second attack, she suffers the Confusion Status.
• 14-16 Blizzard: Deals highest 1 of 2d12 Ice-type damage to the target(s) (Blast damage). If any damage dice match, the target suffers the Freeze Status.
• 17-20 Psychic: Deals highest 1 of 2d10 Psychic-type damage to the target. On a Critical Hit, the target has Disadvantage to Willpower saves until they pass a Willpower save.

Sable is proud, manipulative, patient, and vain. She is unfriendly (-2 to Willpower saves to influence her) towards player Pokémon, attempting to coerce or deceive them into surrendering any valuables they have. 

She is antagonistic towards King Aegon and has entertained Belladonna's attempts to sway her towards usurping the Exeggutor, but feels she has not yet been offered suitable compensation for her role in ousting the Lodge's ruler. She lords over the Cubone of Cubone Cave (02.04) and exacts tribute from them every moon's turn.

02.03 TAMER'S DUNGEON: This secluded forest dwelling was once home to a Tamer. The Tamer's Golduck, Tracker, has since taken over, establishing a home alongside others of the Tamer's Pokémon, including a Persian and three Sandshrew.

Tracker (male Golduck): STR 12, DEX 13, WIL 10, 15 HP, Armor 1. Water type. Resists Fire, Ice, and Water. Weak to Electric and Grass. Items: Paralyze Heal, Super Potion. Moves (d20):

• 1-8 Confusion: Deals 1d6 Psychic-type damage to the target. If any damage dice match, the target suffers the Confusion Status.
• 9-13 Fury Swipes: Deals 1d4 Normal-type damage to the target. On a Critical Hit, if the additional damage die rolls its highest possible value, roll an additional damage die. If that additional damage die rolls its highest possible value, roll an additional damage die. Ad infinitum, if the damage dice keep rolling their highest possible values.
• 14-15 Hydro Pump: Deals highest 1 of 2d12 Water-type damage to the target.
• 16-20 Surf: Deals highest 1 of 2d10 Water-type damage to the target(s) (Blast damage).

Tracker has been hardened by his trainer's cruel nature. He believes that might makes right, and only the strong survive. He is indifferent towards player Pokémon. With a suitable offering, he may be convinced to transport player Pokémon down or upriver using Surf.

Tracker is uninvolved in the power struggle taking place at Pinecrest Lodge (01.04). He dislikes the Pokémon of Fisherman's Cabin (01.03) due to their claim to the river which borders the two's territory, though no blood has yet been shed.

02.04 CUBONE CAVE: This wooded cavern bordered by rivers is home to a clan of twelve Cubone. They howl mournfully at the moon at night, and their cries are particularly terrible during the full moon. It is on these nights that the Cubone meet the Jynx of Frost Witch Barrow (01.05) at the river bordering their territory to pay homage in exchange for another moon's turn of peace and solitude.

Cubone: STR 7, DEX 5, WIL 5, 9 HP, Armor 1. Ground type. Immune to Electric. Resists Poison and Rock. Weak to Grass, Ice, and Water. Moves (d20):

• 1-7 Bone Club: Deals highest 1 of 2d6 Ground-type damage to the target. If any damage dice match, the target flinches and misses their next turn.
• 8-20 Growl: The target’s attacks are Impaired for the remainder of the encounter.

The Cubone are unfriendly towards player Pokémon (-2 to Willpower saves to influence them), suspecting that they are agents of the Witch dispatched to torment them.

03.01 ELECTABUZZ DEN: This rocky burrow is home to a pair of male Electabuzz. They emerge during thunderstorms to seek out lightning strikes and absorb the electrical energy.

Electabuzz: STR 12, DEX 16, WIL 11, 12 HP, Armor 1. Electric type. Resists Electric and Flying. Weak to Ground. Moves (d20):

• 1-5 Quick Attack: Moves first in combat this turn regardless of Dexterity save results. Deals 1d4 Normal-type damage to the target.
• 6-12 Screech: The target's Armor is reduced by 2 for the remainder of the encounter.
• 13-17 Light Screen: Any Dragon-, Electric-, Fire-, Grass-, Ice-, Psychic-, or Water-type attacks against them are impaired for the remainder of the encounter or until any Pokémon in the encounter uses Haze.
• 18-20 Thunder: Deals highest 1 of 2d12 Electric-type damage to the target. If any damage dice match, the target suffers the Paralysis Status.

The Electabuzz are good natured but very energetic, and their excess energy can often tend towards misplaced aggression. They are especially aggressive during storms. They are neutral towards player Pokémon and usually attack only if provoked or if they've very recently been struck by lightning, which fuels their aggression.

03.02 QUAGSQUETON: See Pokémodule B1: In Search of the Unown.

04.04 OLD NEEDLEROOT TOWN: This ruin beside the lake was once a cozy little town from which Pokémon trainers would begin their adventures. After the collapse, it was looted and sacked by wild Pokémon. The forest has since reclaimed the place, with vegetation growing over the old buildings. A variety of Pokémon who shun civilization have made it their home.

Most notable of these is a gang of wild Pokémon outlaws led by Brutus, a male Poliwrath. The gang uses the town's old professor's Lab as their base of operations, extorting any Pokémon who seek to pass through the ruin on the road through the forest.

The ruin is also infested with Rattata who dwell in the mazelike Warrens which wind through the earth directly beneath the site. They are led by their Raticate chief, Gnash.

The outlaws outnumber the Rattata but have been unable to root them out of their Warrens. The Rattata often launch nighttime raids against the bandits, using the Warrens to move about the ruins unseen. Each faction will likely attempt to recruit the player Pokémon to wield them against the other.

Encounters (d4):

1. Metapod: Four Metapod have made pilgrimage to a huge dead pine tree. They cling to its upper reaches, preparing to evolve into Butterfree. They are focused on their ascendance and are indifferent to the local power struggle and to player Pokémon. If convinced to join the player Pokémon in battle against either the outlaws or the Rattata, any surviving Metapod will evolve into Butterfree, at which time they become friendly (+2 to Willpower saves to influence them).
2. Outlaw Spearow: A flock of five Spearow from Brutus's gang loiter amongst the ruins. They are unfriendly (-2 to Willpower saves to influence them), seeking to bully and harass player Pokémon into surrendering any treasure they have.
3. Pokémerchants: A pair of traveling Jigglypuff "merchants" make their way through the village, with fourteen male and female Nidoran serving as guards and beasts of burden, hauling a wagon loaded with goods. A single Mankey serves as captain of the guards. They are traveling along the road between settlements and are helpful (+4 to Willpower saves to influence them), eager to bribe player Pokémon with merchandise in exchange for protection. They are transporting 900 Pokédollars in goods (Antidotes, Awakenings, Paralyze Heals, Potions, and the like) and a pay chest containing 370 Pokédollars.
4. Rattata: Six Rattata emerge from the Warrens looking for allies. They are helpful (+4 to Willpower saves to influence them), offering to accompany the player Pokémon on whatever their current objective is in the hope that they will aid the Rattata against Brutus's outlaws.

Lab: This overgrown professor's laboratory is the hideout for Brutus's gang of outlaw Pokémon. The gang consists of 70 Spearow, nine Meowth, two Grimer, two Lickitung, an Onix, a Haunter, a Magmar, and Brutus himself. The outlaws are unfriendly (-2 to Willpower saves to influence them), bullying, extorting, and potentially imprisoning any player Pokémon who come here, if they think they can get a ransom from them. Alternatively, they may dump prisoner Pokémon into the Warrens, forcing them to clear out the Rattata infestation within. They have seven Pokémon from the surrounding locations currently being held for ransom and another eight kept as indentured laborers and servants.

Brutus (male Poliwrath): STR 13, DEX 11, WIL 9, 17 HP, Armor 1. Water/Fighting type. Resists Bug, Fire, Ice, Rock, and Water. Weak to Electric, Flying, Grass, and Psychic. Items: Full Heal, Max Potion, and Paralyze Heal. Moves (d20):

• 1-4 Body Slam: Deals 1d8 Normal-type damage to the target. On a Critical Hit, the target flinches and misses their next turn.
• 5-9 Hypnosis: Rolls 1d20. On an 8 or higher, the target suffers the Sleep Status.
• 10-14 Bubblebeam: Deals highest 1 of 2d8 Water-type damage to the target. If any damage dice match, the target has Disadvantage to Dexterity saves until they pass a Dexterity save.
• 15-20 Submission: Deals highest 1 of 2d8 Fighting-type damage to the target, but takes 1 damage (not reduced by Armor) in return.

Warrens: These winding, narrow subterranean tunnels are home to a nest of 100 Rattata and their Raticate chief, Gnash. They are helpful (+4 to Willpower saves to influence them), eager to assist the player Pokémon in exchange for allies in their struggle with Brutus's gang. If the player Pokémon accept, the Rattata can sneak them inside the outlaws' hideout.

Gnash (male Raticate): STR 12, DEX 15, WIL 7, 10 HP, Armor 1. Normal type. Immune to Ghost. Weak to Fighting. Items: Hyper Potion. Moves (d20):

• 1-4 Double-Edge: Deals highest 1 of 2d12 Normal-type damage to the target but takes 1d4 damage (not reduced by Armor) in return.
• 5-8 Hyper Fang: Deals highest 1 of 2d8 Normal-type damage to the target. If any damage dice match, the target flinches and misses their next turn.
• 9-17 Focus Energy: His next attack will be a Critical Hit as long as its damage die doesn't roll a 1.
• 18-20 Super Fang: Rolls d20. On a 3 or higher, deals Normal-type damage equal to half the target's current HP (or Strength if the target has 0 HP), rounded down. This attack affects Pokémon immune to Normal-type moves.

04.05 POISON BRIAR DEN: This cavern deep in the forest is overgrown with stinging nettles. Fortunately, its inhabitants, Lady and Tramp, a mated Nidorina/Nidorino pair, and their female Nidoran pup, are immune to the plants' irritating barbs.

The pair live a peaceful existence in their secluded hideaway. They occasionally encounter Rattata or Spearow scouts from Old Needleroot Town (04.04) at the edge of the deep wood, bringing them back to the lair to be consumed by their offspring.

While generally insular, the pair become incensed if attacked in their den. If the Nidoran pup is threatened, Lady will become hostile, and all her attacks against the target will do maximum damage (roll to check for Critical Hits). Likewise, if Lady is attacked, Tramp becomes hostile, and all his attacks against the target will do maximum damage (roll to check for Critical Hits).

Lady (Nidorina): STR 9, DEX 8, WIL 7, 13 HP, Armor 1. Poison type. Resists Fighting, Grass, and Poison. Weak to Bug, Ground, and Psychic. Items: Full Heal. Moves (d20):

• 1-7 Double Kick: Deals 1d6 Fighting-type damage to the target. If the target has a Substitute, the attack breaks the Substitute, and the target takes half damage.
• 8-15 Poison Sting: Deals highest 1 of 2d4 Poison-type damage to the target. If any damage dice match, the target suffers the Poison Status.
• 16-17 Thunder: Deals 1d12 Electric-type damage to the target. If any damage dice match, the target suffers the Paralysis Status.
• 18-20 Fury Swipes: Deals 1d4 Normal-type damage to the target. On a Critical Hit, if the additional damage die rolls its highest possible value, roll an additional damage die. If that additional damage die rolls its highest possible value, roll an additional damage die. Ad infinitum, if the damage dice keep rolling their highest possible values.

Tramp (Nidorino): STR 11, DEX 10, WIL 7, 11 HP, Armor 1. Poison type. Resists Fighting, Grass, and Poison. Weak to Bug, Ground, and Psychic. Moves (d20):

• 1-6 Double Kick: Deals 1d6 Fighting-type damage to the target. If the target has a Substitute, the attack breaks the Substitute, and the target takes half damage.
• 7-13 Poison Sting: Deals highest 1 of 2d4 Poison-type damage to the target. If any damage dice match, the target suffers the Poison Status.
• 14-19 Focus Energy: His next attack will be a Critical Hit as long as its damage die doesn't roll a 1.
• 20 Horn Drill: If he acts before the target, he rolls 1d20. On a 15 or higher, the target automatically drops to 0 HP and 0 Strength.

05.04 LEEKFIELD: This tiny community is home to about 50 Pokémon, who are led by Mallow, a female Farfetch'd. The settlement is a cluster of small shelters, with the surrounding grassland cultivated into fields of berries, grains, and wild vegetables. The settlement gets its name from the abundant leek-like reeds which grow along the lakeshore.

Mallow (female Farfetch'd): STR 10, DEX 9, WIL 8, 10 HP, Armor 1. Normal/Flying type. Immune to Ghost and Ground. Resists Bug and Grass. Weak to Electric, Ice, and Rock. Items: Paralyze Heal. Moves (d20):

• Fly: Soars into the air, evading all attacks except Gust, Swift, and Thunder. On the next turn, she lands and deals highest 1 of 2d8 Flying-type damage to the target. The Move can also transport up to six other Pokémon to the Pokémon Center at Pinecrest Lodge (01.04).
• Swords Dance: Double Enhances her attacks for the remainder of the encounter or until any of her damage dice rolls a 1.
• Agility: Has double Advantage on Dexterity saves until she fails a Dexterity save.
• Slash: Deals highest 1 of 2d8 Normal-type damage to the target. The attack is a Critical Hit if it rolls a 7 or 8.

Mallow and the Pokémon of this community are wary towards outsiders, preferring to stay uninvolved with conflict and danger. Pokémon traveling to Old Needleroot Town (04.04) and the forest south of it are often preyed upon by the outlaw Pokémon, Rattata, and the Nidorina and Nidorino living in Poison Briar Den (04.05). Several Pokémon from the community are held for ransom by the outlaw Pokémon. Mallow and Leekfield's inhabitants will become friendly if any of these problems are resolved for them.

Developing the Sandbox: Strongholds

This post is part of my series on developing the sandbox. You can read other posts in the series here. The series will, when necessary, go into detail on the development of my sandbox for Dungeon Module B1: In Search of the Unknown, but will also discuss sandbox development more generally. You can read play reports from my B1 campaign here.

In this post, I'll discuss the development of those strongholds in your sandbox which are ruled by NPCs. These are your classic fighters' bastions, wizards' towers, clerics' temple fortresses, and thieves' dens and hideouts, but it doesn't stop there. 

Paladins will rule strongholds somewhere between that of a fighter and a cleric. Rangers may rule fortified compounds in a more wilderness locale. Illusionists will disguise their strongholds to appear mightier than they truly are, or as simple earthen mounds or other natural features. Druids will rule secure woodland groves. Monks build monasteries. Assassins rule imposing wilderness fortresses.

Monster lairs can also be strongholds, and many famous "monster lair" modules play very similarly to strongholds. Bandits, berserkers, brigands, and dervishes may come to inhabit them. Giants and many underwater creatures will build strongholds as well. 

Two things that this post will not cover are (1) strongholds built by the player characters and (2) deserted strongholds, which are likely naught but ruins on the surface and which fall more readily into the dungeon category.

Most scholarship (i.e. blogs) on the subject of strongholds in D&D pertains to those built by player characters and the rules for doing so. I'm instead interested in how to design and key an existing NPC stronghold for the players to interact with. 

Strongholds are kind of like dungeons in the sense that the player characters can move from room to room and interact with objects inside. They will contain "empty" rooms, "monsters", traps, and treasure. They are also kind of like monster lairs in that they are home to a type of "monster" - broadly speaking, most often men and/or demi-humans, but more specifically, powerful NPCs that are very much like the player characters themselves, plus their henchmen, men-at-arms, special followers, pets, and the like.

Strongholds are tricky because they are very much real places, and complex ones at that. You'll want to have some knowledge about different types of real-world strongholds and how they functioned. You have to think about how real-world strongholds were constructed and defended, and anticipate ways in which your players might want to infiltrate or lay siege to them. You have to think about who all lives there and how much room is needed to house them.

Dungeons are similarly complex, but they're not exactly real places. It might be hard to imagine them without something historical to reference, but there are often robust procedures for designing them, and lots of advice on blogs. Because they're not really real, you can also let your imagination run wild. 

Monster lairs are "real" in the sense that bears and wolves and the like are a sort of "monster" that lives in our own world, and you can somewhat easily imagine what such a place would be like if instead an owlbear lived there, or what an otherwise normal cottage would be like if a leprechaun were instead its resident. They're also not usually especially complex, so there isn't a huge cognitive load when designing them.

There are also more fantastic elements to consider. What spells are laid upon the stronghold? Are there any monstrous inhabitants kept as pets - perhaps a subdued dragon or some enslaved monsters? Perhaps it's a stronghold like in OD&D, with heroes riding griffons and rocs, or defended by chimeras, trolls, manticores, and the like:

Just as you might steal some dungeons for your sandbox to lighten your workload, you can steal strongholds too. You can pretty easily steal the Keep on the Borderlands, Castle Ravenloft, the Hall of the Fire Giant King, or whatever. But you can't always steal exactly what you need, and you might want to create something original. It's not much fun to just use the Keep on the Borderlands over and over again. Is there a good adventure out there which features a fortified druid's grove? I don't know. I'll probably have to make one.

Because they are complex locations, inaccessible by design, ruled by powerful NPCs, and inhabited by large numbers of potentially hostile defenders, I generally assume that NPC strongholds will be one of the last places which the player characters explore in depth. Thus, they usually wind up at the very end of my list of sandbox locations to develop in more detail.

As with every location in my sandbox, before I go into more detail, I start with a minimalist sketch. For strongholds, this includes the stronghold type, who rules it, and how the ruler and their followers might react upon first meeting the player characters, as well as how they relate to those locations which surround them.

For the type of stronghold, I use this handy table from the AD&D DMG:

(Spoiler: The section on CONSTRUCTION AND SIEGE ENGINES really doesn't provide any details about castle types aside from how long it takes to build them. There's lot of information about excavation and mining - including notes on the efficiency of slave labor - a sort of glossary of castle features and how much each costs, and some rules for sieges. This section is kind of interesting, and the glossary is worth looking at, but it won't help you too much in actually designing such a location for your game.)

A castle can be totally deserted, deserted and inhabited by a monster, inhabited by humans (that is, bandits, brigands, berserkers, or dervishes), or inhabited by character-type NPCs:

Both types of deserted castles fall into the dungeon bucket, so we'll focus on just the other two types. There's another table for determining the type of humans in the castle:

In the case of character-types, it isn't as simple, because in AD&D, the stronghold ruler's class will also have some effect on what type of castle can be built. For example, a thief can only rule a small stronghold in or near a town or city, a monk can only rule a moat house or friary, and a druid's stronghold must be in the wilderness. This means that depending on which class is rolled as the ruler, we might be limited in which type of castle is possible.

I wrote up a handy summary of which type of character can rule a stronghold at what level and what type of stronghold they can rule, which is worth checking out. If you're not using AD&D, it isn't super important, but I think it's at least illuminating, and you can always use AD&D's tables for worldbuilding, even if you're using a different system.

Because the ruler's class affects what type of stronghold can be built and where, I determine that first using this table:

The level column is not quite accurate, which I elaborate on in the aforementioned post on the subject. 

In AD&D, only certain races can attain a high enough level in their class to build a stronghold. For example, only human clerics can reach a high enough level to build a stronghold, so if I roll cleric, I know my ruler is human. 

Similarly, if my stronghold is in the wilderness, I know it can't be ruled by a thief. If I roll monk, I know it must be a moat house or friary, and so on. It's a lot to wrap your head around, which is why I thought it worthwhile to summarize in that earlier post.

Knowing who my ruler is, I can then determine their alignment. This is also dependent upon class, such that paladins will always be Lawful Good, rangers always Good, druids always Neutral, and so on. This then modifies their reaction rolls pertaining to the player characters and their neighbors in the sandbox. 

Using my B1 sandbox as an example, I know that the brigands ruling Fayette's Hold are unfriendly towards Timbershore. They're not hostile, so they aren't attacking the town or riding around the countryside killing people, but they've occupied Ebongrove, the thorp at the foot of the castle and just outside of town, and are demanding tribute from Lambrecht, the wizard who rules his own stronghold in town. I also know that the brigands were originally recruited by Lambrecht to subvert the castle's rightful ruler, and that they're now holding the ruler for ransom. The people of Ebongrove are hoping someone will free Fayette from her cell. Lambrecht is attempting to court the ogres of Okorg's Den to aid him in retaking the castle.

That's pretty much all I need for a first draft for my sandbox. To get this location ready for play, I have to go deeper, starting with who all lives there.

In the case of bandits, berserkers, brigands, and dervishes, I simply use the numbers in the Monster Manual. For character-type NPCs, it will be more complicated:

1d4+1 henchmen, light and heavy cavalry led by fighters, and any additional special followers appropriate for the class. (If I'm interpreting this correctly, a cleric's stronghold is instead garrisoned by their special followers, while the other classes get both their special followers and the garrison listed above? Unclear.) 

The classes of henchmen are determined on a separate table than that used to determine who rules the stronghold:

This table yields slightly fewer clerics, fighters, magic-users, illusionists, assassins, and monks, and more druids, paladins, rangers, and thieves. I will ensure that henchmen are compatible with their masters, so a Lawful Good ruler will not have Chaotic Evil henchmen and a paladin will not have a henchman assassin. The level of these henchmen is determined by the level of the ruler:

If my hypothetical stronghold ruler is a Neutral Good 9th level human cleric, each of their henchmen will be 6th level, and none will be Evil. I find it fun to turn these henchmen into their own little "factions" of a sort, making opposed reaction rules with the ruler to determine whether they're favored, neglected, or on the outs, if they're fanatically loyal, unhappy, or actively seeking to betray their master. Remember that henchmen are, at the beginning, the pitiable wretches of the AD&D world, so it isn't hard to believe they might turn their cloaks if they think they might receive better treatment elsewhere.

A good example of this can be found in my Hexmas sandbox, the Jinglebell Barrens. The Tower of Noel is ruled by a fighter with several henchmen, each with their own personalities and agendas.

A stronghold's garrison will have a mix of from 38 to 64 men-at-arms (average 51). I don't love that this is unaffected by the size of the stronghold, so I like to roll as normal for a medium stronghold, roll twice for a small stronghold and keep the lower roll, and roll twice for a large stronghold and keep the higher roll. So a tower will have 2d4(keep lowest)+8 heavy cavalry, a large shell keep will have d4+8 as normal, and a fortress complex will have 2d4(keep highest)+8.

On to special followers. Clerics get a mix of cavalry and infantry:

Fighters get a medium level leader type and a company of either heavy infantry, crossbowmen, cavalry, or a mix of light cavalry and infantry:

Rangers are awesome and get 2d12 followers of mixed types:

This will mostly be normal stuff like character-type humans and demi-humans, but the lower you roll on 2d12, the more likely you are to get really cool stuff like giant animals, monsters, fey creatures, dragons, giants, treants, and lycanthropes:

This is the only thing in AD&D that really hints at OD&D's table of weird stronghold defenders, and that's too bad. There's no reason that your fighter's stronghold can't be garrisoned by griffon riders or your cleric's temple haunted by undead, but it would be a lot cooler if it was made explicit in some of these tables. Still it is really sick that if you play a ranger you have a chance of a dragon or giant just showing up to your stronghold to hang out.

Anyway, thieves get 4d6 thieves of 1st to 7th level and assassins (upon usurping the reigning Guildmaster/mistress) get 7d4 assassins of 1st to 8th level, 75% of which desert during the overthrow and are replaced by new 1st level recruits. There are more tables for those but they're not really worth sharing following those cool ranger tables.

After all that, you have to do the actual designing of the place, including layout and keying. Since strongholds are real historical places, you can do some research into each kind to figure out what exactly you're creating. You probably know what a tower is, but you might want to look up what exactly is a moat house or a shell keep or a concentric castle. 

Since medieval strongholds are not one of my special interests, I did some research. YouTube has a lot of videos describing different types of medieval strongholds, some with 3D renderings/walkthroughs detailing what might be found inside. I also watched some videos about how various defenses worked and how attackers usually laid siege to such places. I found it all enlightening and super interesting, and it's gotten my gears turning thinking about how I'd handle these scenarios in a D&D game.

The stronghold's character should be informed by the ruler of the stronghold. A thief's stronghold will be filled with nefarious traps. A magic-user's or cleric's will be warded by magic and protected by certain conjured beings. An illusionist's stronghold will be a very frustrating place indeed.

It's good to look over the list of spells available to the ruler and their henchmen to determine what sorts of magical effects might pervade the place, with a special eye towards those with a permanent duration. The stronghold may be guarded by animated dead. It may be lit entirely by continual light spells. Certain areas may be trapped with a glyph of warding. The stronghold may be surrounded by an overgrown thicket affected by plant growth. Locked doors may be affected by Leomund's trap. Other doors may be secured with wizard lock or affected by fire trap. A wall of iron may enclose the stronghold in place of wood or stone walls. Earthworks may be formed by move earth. An antipathy spell may repel attackers from a certain area. Clone vats may be found in one chamber. The windows of the ruler's chambers may be reinforced by glassteel. Permanency can make permanent the effects of enlarge, fear, gust of wind, invisibility, magic mouth, prismatic sphere, stinking cloud, wall of fire, wall of force, and web. 

There are many possibilities which will make the stronghold even more challenging than it otherwise would be, and this can quickly get out of hand. At the same time, this is the place where the NPC lives, so why shouldn't the wizard secure every door in the place with a wizard lock? Some spells will have costly material components or will be too high level to cast flippantly, so these should be treated with special consideration.

For the layout, you might consider using historical castle floorplans. As some of my contemporaries pointed out when I raised this subject, real-life castles are not super gameable. They are impregnable deathtraps which offer very little choice to player characters who might want to infiltrate or assault them, which is by design. They don't make for good dungeons. 

I think that's basically fine, because the types of strongholds in question aren't dungeons and aren't supposed to be. They're what the player characters' own armies of followers and men-at-arms are for - wargaming and sieges. This post isn't about adjudicating mass combat or siege scenarios, so although the topics go hand in hand, we'll leave them for now.

With a good plan the party might be able to stealthily scale the stronghold's walls and neutralize its defenders, or at higher levels they could teleport into the ruler's inner sanctum and confront them directly, but otherwise they'll need an army of their own.

...Unless there was some other way in. I seem to remember a famous D&D module which takes place in a dungeon built under an old NPC stronghold and a passing assumption in the AD&D DMG that player characters (and, presumably, NPCs) will always build dungeons beneath their strongholds. Such a dungeon could not only expand the functional square footage of such a stronghold, but also serve as a secret escape route for the stronghold's ruler in the event of a siege. And where there's a way out...

The dungeon beneath an NPC's stronghold is as important as the stronghold itself. Every NPC stronghold should have one. Like the stronghold above, these should be informed by the character of the NPC who rules it. You might go the B1 route and make this the mad labyrinth of a true dungeon sicko, or perhaps it is a more sensible extension of the stronghold itself - a series of cellars, cisterns, prison chambers, death traps, and treasure vaults. 

Is the dungeon garrisoned with men-at-arms like the stronghold above, or have they captured a menagerie of monsters and locked them down below to deter any would-be besiegers? This might be where the cleric keeps their undead servitors, or where the magic-user has bound their conjured elemental and fiends into service. This might be where most of the thief's traps are laid. 

Through the dungeon, the player characters might gain access to whatever area of the stronghold the ruler calls home. There will be guards to be dealt with inside, but it's still a lot better than storming the stronghold's walls and being peppered with arrows, bludgeoned with falling rocks, and scalded with boiling oil in the mad scramble over the walls. 

The assumption that characters will always build dungeons beneath their strongholds not only provides a more convenient way in for a small band of adventurers, but also a handy justification for defaulting to D&D's default mode of play, where the game is at its strongest. Whether intentional or not, it's a genius bit of design which simplifies a location which poses unique and difficult challenges for the player characters and DM alike.

In a way, strongholds are settlement, dungeon, and monster lair all in one. Preparing one for your sandbox requires you to combine approaches from all three of the former location types along with an understanding of the real-world places in our own history. It is no easy feat, and the resources provided by the game itself are somewhat lacking in this area. If you manage to design a good one, however, you'll be rewarded with a location that can support a variety of modes of play, from social intrigue to mass combat to your usual dungeon fare.