Sponsored By

Featured Blog | This community-written post highlights the best of what the game industry has to offer. Read more like it on the Game Developer Blogs.

Creating quests with branching narratives for open world games is no simple task. By applying a simple iterative method, you can save yourself from many problems later on in the process.

Nick Witsel, Blogger

March 20, 2017

8 Min Read

My name's Nick Witsel, I'm a freelance narrative designer by day, and a mod developer by night. I'm currently working on Lordbound, a DLC-sized modification for Bethesda's The Elder Scrolls V: Skyrim. It's being developed by a team of +10 industry professionals and students to learn more about AAA development.

As quest designer, it's my job it design and implement quests using the Skyrim Creation Kit. I wanted to explain how I recently created the side-quest 'The Consequences of War' and share some insights into my iterative approach to designing quest content with branching narratives.

The First Steps

The Consequences of War began as a brief list of notes written by one of the other quest designers in the team. The premise was simple: an Orc has been imprisoned for his crimes in taking part in a siege on a village. A refugee from that village tasks the player with killing the Orc and avenging is lost wife. The Orc, on the other hand, asks the player to help him escape. Players can decide what to do.

I immediately set out to define the quest flow to a very basic degree. (Seen below) This would give me some boundaries to work within.

In order to confine scope, I limited the quests progress to two distinct branches, which would join together when you spoke to the refugee at the end. It is to my belief, that a side-quest should have only one 'end-objective'. In this case, returning to the refugee.

This makes things much easier to oversee and prevents problems further down the line. If you want a quest to add new end-objectives half way through - make a new quest! Sure, there can be different outcomes to a quest, but in this case that'd just be a matter of writing a different responses for the refugee. Gameplay and scripting-wise, players were undertaking the same actions.

Next up, was research. With the basic confines set, I started to do some digging into the elder scrolls wiki. My aim was to come up with interesting potential backgrounds for my NPCs, and in return , make the quest more grounded in the world of Tamriel.

Quest Design and First Playable Draft

That same day, I sketched out a more detailed flow chart for my quest. Yellow indicates dialogue and blue, player actions. 

This flowchart served two main purposes:

1. Help me get a good overview of how the quest would be structured and foresee possible difficulties. (Like how will players collect a disguise?)

2. Assist in communicating the quest concept to Lordbound's quest team for review.

I also noticed immediately that both the 'Free Orc' and 'Kill Orc' strands could potentially share an objective: find the key. Knowing this made it easier to set up my quest objectives and journal displays later on in the process.

At the same time, I also conducted more research into the Elder Scrolls lore and wrote around 300-word character bios for the Orc and Refugee NPCs.

The quest's main design goal was to introduce players to a major conflict that had take place in the Druadach Valley, the worldspace of Lordbound. A quick quest where two NPCs give you conflicting information and objectives to serve as context would be an ideal way to do this. At the same time it also provided players with a moral dilemma, which helped underline the world introduced in Lordbound had no outright heroes or villains.

When I build multi-linear quests, I initially focus on completing one vital strand from start to finish. The red line you see in the quest flow chart to the right indicates this. To summarize:

1. Get told to kill the Orc.

2. Find the key to his cage.

3. Open the cage and kill him.

4. Report back.

It was simple, and I managed to fully write, script and implement the first strand within the first day. I only wrote vital dialogue. So no additional topics for stuff like: "Where you from?" etc. 

Building something as soon as possible makes easier to foresee scope. Knowing that the first strand had been easy to craft (and finish, as a player) I decided it wasn't a bad idea to add some extra strands. For instance, I added an option for the player to speak to the captain in charge of the prison. As you can see in the flow chart, it's a dead end - but does add some extra insight into what these Imperials think of the Orcs of the valley.

Crafting Interesting Characters

I wanted the refugee to be the character who initially has the moral high ground. The Orc had killed his wife, and he wanted to have revenge. A simple concept for players to understand. Further delving into his background would explain how helpless and innocent he had been when the unfortunate events had occurred. But at the same time, his desire to see someone murdered had to come across as desperate and uncertain - not brash and heroic. Skyrim is full of quests with people asking you to slay people with much conviction. I wanted to subvert this trope. Killing isn't cool, or fun. It's messed up.

And then there was the Orc. At first he would seem like the archetypal representation of Orc society: brash and stupid. But asking him to explain his actions would provide players with a better understanding of their culture, and insight into their point of view: 'Hurt us, and we hurt you back twice as hard.' For him, war and conflict weren't about heroics or dying in battle, but about fighting for a just cause. He also understood the refugee's sorrow, but didn't want to die as a prisoner. Free him, and he would gladly accept to battle the refugee in a duel. This Orc was brash, yet surprisingly thoughtful. 

Subverting player expectations while working with well known archetypes, can keep things fresh and interesting but also accessible and quick to understand.

Further Iterations

When building a quest, it is often you come to the realization that your ideas are not only harder to implement, but can be implemented in many possible ways. The concept of poisoning the orc's food was easy to write on paper - but there were many ways such an action could be implemented in the game.

Would there be a prison chef you could convince to do it for you? Was there a room full of rations somewhere you could break into and replace with poisoned ingredients? To keep things simple and equal with other branches, I simply wrote some dialogue in which you could persuade the Orc to poison his own food - an option most people would fail due to the high speechcraft requirements (already quite a rarity in Skyrim). Failing it would result in some funny dialogue. 

Also, the irony of using the speech skill to talk people to death was one which I felt had a place in Lordbound.

Using my limited acting range could provide the quest with temporary voice assets. It was time to playtest my quest with someone. The results lead to some insightful conclusions. In 2013, Telltale gave a talk at GDC about the narrative design for The Walking Dead: Season 1. One thing they mentioned, was that when players were confronted with a decision between saving one person or the other, players would often try to save the character they simply knew the best. To keep these choices interesting, writers had to ensure both characters would have to have had the same amount of exposition.

When playtesting my quest, I noticed the player instantly decided to save the Orc, simply because he provided the most amount of backstory up until that stage. After hearing of the Orc's escape, the refugee provided the player with more exposition into his character, leading to the player feeling a little uncertain of her actions. So to make things a little bit less straight forward, I added extra dialogue options to the refugee at the start of the quest to give players some more insight into his background.


Crafting side-quests for Skyrim has taught me how to apply iterative design methods to foresee future issues and improve the quality of the content. Working out a single strand first and adding more from there on out allows me to iron out bugs and resolve any usability issues as soon as possible through playtesting (either by myself or by other players). 

Working on a strand-basis also allows me to keep a good eye on how long it'd take to produce further strands - making it easier to decide if I'd prefer to focus on creating more branches - or improving the quality of the existing content.

Lordbound, a Skyrim Mod, is scheduled to release Q4, 2017. It will have around 15-20 hours of gameplay and contain full voice acting. Early playable versions of my content are available on request.

Lordbound's page on ModDB: http://www.moddb.com/mods/lordbound

My portfolio site: http://nickwitsel.weebly.com/about.html

Read more about:

Featured Blogs

About the Author(s)

Daily news, dev blogs, and stories from Game Developer straight to your inbox

You May Also Like