In part 3, I described using a list of genre influences - an Appendix N - as a guide for creating encounters and non-player characters for a swashbuckling sandbox. Genre tropes in a sandbox may provide a measure of familiarity to players which in turn increases their confidence in their decisions pertaining to their characters, encouraging them to drive the action, which is the goal in a sandbox setting.
This may be perhaps the most obvious and least useful advice ever, but from reading blogs and intreweb forums, sometimes that which seems so obvious to me is a complete mystery to others, so I'm not taking anything for granted.
One of the fundamental differences between the sandbox settings most familiar to fantasy or science fiction gamers, is that the cape-and-sword sandbox less likely to be organized around places - the keep, the Caves of Chaos, and the Mound of the Lizard Men, for example, or Hefry, Ruie, and Forboldn in Regina subsector. Rather, the swashbuckler's sandbox is organized around people.
The non-player characters in the swashbuckler's sandbox form a web of relationships, a web which is in some ways analagous to a traditional dungeon. Think of each person as a 'room,' and the relationships between them as the 'corridors' by which they are connected.
To take that social-relationships-as-dungeon analogy a little further, the social world of the swashbuckler contains some of the hazards associated with old school dungeon crawls. Many of the relationships/'corridors' between characters are hidden behind 'secret doors,' for example, and must be discovered by the adventurers - in our Flashing Blades campaign, for example, one of the player characters' close allies is secretly connected to one of the princes of the blood, but the adventurers are unaware of this relationship at the moment. There are also 'chutes' which in the traditional dungeon drop players from upper levels to the more dangerous lower levels, and in the context of the swashbuckler's sandbox connect a less powerful npc to a more powerful one - in our campaign, the player characters found such a chute when one of them challenged a romantic rival to a duel and belatedly discovered that the rival's uncle is a lieutenant in the Cardinal's Guard.
A swashbuckler explores this web of relationships like a dungeoneer crawls through a lich's tomb, and for the same reasons - as the dungeoneer finds risks and rewards in the crypts of that dusty necropolis, it's the non-player characters populating a cape-and-sword game-world who provide danger, riches, and a chance at glory for the adventurers.
That said, I must allow for 'the piracy exception' - the sandbox for a pirates campaign may closely resemble that of more traditional roleplaying sandboxes, with pirates seeking out galleons and ports and lost temples to plunder in the style of their dungeoneering peers in a fantasy campaign.
The piracy exception aside, in the swashbuckler's sandbox non-player characters are the gateway to rewards and hazards - a jeweled ring is more likely to make its way into a swashbuckler's hand as a token of thanks from a grateful queen than it is from a locked chest guarded by a black pudding. Relationships with npcs are often rewards and hazards in and of themselves, particularly in games and campaigns where increased social status is one of the rewards offered to player characters. Non-player characters may offer an adventurer wealth, influence, sex, membership in an organization, and access to new skills - Lagardère seeking an introduction to the duc de Nevers in order to learn the botte de Nevers is a classic example - as well as danger, in the forms of duels, intrigue, manipulation, and loss of reputation.
To this end, the referee running a swashbuckler's sandbox should devote the same imagination and effort to creating engaging, challenging, mysterious non-player characters as other referees do a megadungeon or an alien world, and those non-player characters should be woven into a complex web of relationships, known and secret, for the adventurers to explore.
Once again, this is probably boilerplate to many - it gets covered in some manner in most referees' sections of game-books and in numerous articles in gaming rags and blog posts. Tips like, "Give each npc two secrets!" are as ubiquitous as rules for initiative or experience. So how is this different with respect to a swashbuckler's sandbox? With the possible exception of the some of the really extensive published city settings, the number and reach of the non-player characters in a cape-and-sword campaign are greater than that of more familiar settings. Consider the extent of the social network of a duke and peer who is a courtier to the king, a soldier in time of war and an extraordinary ambassador in peace, and the lord of a large estate. This one non-player character could have a web of immediate contacts, within a single degree of separation - family, servants, courtiers, and clients - comparable to all of the named npcs is the village of Hommlet and the keep on the Borderlands and then some, and many of those npcs may have considerable networks of their own . . .
Now, very few referees prepare anything other than a relatively small sandbox by attempting to detail every single person anymore than they describe every structure in a large city or every hovel in a marcher barony. What a sandbox referee does prepare are a few key locations in varying depth with the expection of improvising additional details as the need arises. The non-player characters in a swashbuckler's sandbox are handled the same way; to use an example from my campaign, I created a number of families in different parts of France who are given considerable detail - to continue to use the dungeon analogy offered earlier, these families can be thought of as 'levels' or 'sublevels' in the dungeon, and they in turn may be connected to other 'levels,' such as a gentleman's club, the King's Musketeers, or a bishop's curia.
In a traditional sandbox, the adventurers are expected to utilize their wits and resources to learn more about the game-world and set goals for themselves to pursue. The players and their characters may hear rumors or legends, or find an old map or a strange atrifact, offering clues to where rewards may be found. The same is true in a swashbuckler's sandbox, but rather than the legend of the old moathouse overheard in the tavern, the rumors and such concern the non-player characters in the setting. For example, one of the families I rendered in some detail for my Flashing Blades sandbox is that of the vicomte de Praz-de-Lys, a Savoyard nobleman who relocated his household to Paris in order to pursue closer ties to the French noblity. The player characters heard a rumor that the vicomtesse needed to fill a number of positions in her household - a tutor for her children, a fencing instructor for her sons, a banker to manage her finances, and so on - giving the adventurers an entrée into the lives of the Praz-de-Lys', their network of relationships in Paris, and the schemes in which various family members engage.
That last point bears noting - while specific cape-and-sword tropes may vary in different campaigns, one should be present in every swashbuckler's sandbox, and that is that the whole thing is awash in intrigue. "Give each npc two secrets!" should really only scratch the surface of the potential rivalries, dalliances, and alliances connecting the non-player characters. Political rivalries, ancient vendettas, and forbidden romance are the stuff of adventure in this game-world, and these agendas should permeate the relationships between npcs.
In addition to rumors for drawing the adventurers into the web of non-player characters who make up my sandbox, I also lean heavily on random encounters, many featuring named npcs, an idea borrowed from Chaosium's Thieves World box set and TSR's Boot Hill adventures. To borrow the dungeon analogy one last time, using named non-player characters in random encounters produces something like falling through the roof into a chamber of a lost shrine, the 'dungeon' presenting itself to the adventurers! The player characters in my campaign, for example, rescued the duchesse de Chevreuse and her lover, Lord Holland, from bravos in a Paris alley, then snuck the pair back into the Louvre, earning the powerful and dangerous noblewoman's gratitude - this was a random encounter generated as a disguised nobleman and a gang of thieves using the generic tables in the Flashing Blades core rules. Another random encounter - the aftermath of a murderous duel - earned one of the adventurers an enemy whom the players believe has tried twice now to assassinate their characters.
And that last touches on my final point, that the sandbox must react to the actions of the adventurers. The player characters' actions - and sometimes their failures to act - carry consequences. My simple rule is this: if the adventurers are winning, someone else is losing, and that someone may choose to do something about it. This is where the dungeon analogy breaks down a bit for me, for in these instances, the 'rooms' of the 'dungeon' may actually seek out the adventurers! Earning a friend may mean making that friend's enemies the adventurers' own as well, and if that holds true, then the enemies of their enemies may offer alliances as well. A swashbuckler's sandbox is marked by a dynamic tension as the adventurers tug, os simply brush against, the web of relationships they encounter in the game-world.
Swashbucklers may travel the world, but what defines their adventures are the people they meet and engage in their travels, and a swashbuckler's sandbox should emphasize this.