top of page
  • craigarthurbooks


Updated: Jul 21, 2021

It might surprise you to learn that episode 4 was the hardest in the series for me to write.

About 60% of the time I spent working on book 1 was on chapters 6-8, and it's still the part I've rewritten the most. So what made it so difficult? And what needed to change to reach a version I was happy with?

This post contains spoilers for episode 4: Chasing Shadows (chapters 7 and 8)


Very early on in the writing process I broke the story down into chapters based on major events.

For example, chapter 4 is "the Show", the major event is "Shiva's Resistance kidnaps Emerald". Once I had an overarching plot I'd bullet point any ideas I had for that chapter.

My outlines for most chapters in book 1 were quite substantial, but what eventually became chapters 6-8 was originally a standalone chapter, and all I had noted was this:

"The group cross City C and instead of finding City D they find a rainforest, where Jacob saves them from a Specialist ambush."

I expected to come up with more specific ideas as I got closer, but they never came. So I just made it up as I went along, treating it as an exercise of getting the characters from point A to point B as fast as possible. This first draft of the chapter was approximately 13,000 words from leaving the Resistance Hall to Jacob's arrival.


When I first offered up this chapter to my willing test subjects, the feedback amounted to: "It's boring, nothing happens."

I wasn't worried. I'd had this problem before to a lesser extent in chapter 3, and I'd solved it then by cutting down the length.

Thus I returned to my handy axe and cut down chapter 6 to about 9,000 words. And... it had gotten worse.

I inspected the plot to see if there was any more fat to hack off, but it was already skeletal. The only other option to skip the journey entirely. So I tried that, and the outcome was far too jarring and confusing, while also necessitating flashbacks later in the story that totally ruined the pacing.

I think I initially struggled because this was the first transitional moment in the story, but I prefer to write immediate action. This meant I was approaching the chapter with the wrong attitude, treating it as a speedbump to get to the "good stuff", rather than an important bridge between the second and third acts. If I, the writer, didn't care, why should the reader?

I needed to stop writing a summary, and start writing a story. So I went in the opposite direction, I wasn't going to recount a recipe for a journey, I was going to show the cake being made.


The purpose of a journey is not solely to get your characters from point A to B physically, but mentally too. I looked over the chapter again through this lens, this time focussed on my characters.

In the first draft I'd forced the group to come together on the journey, but why would they? After the destruction of the Resistance this group would be incredibly dysfunctional, their motives are entirely different and they have no reason to trust each other yet:

Wolf wants to find Jacob, Shiva wants to lead Emerald and the Specialists away from the Gutters, Emerald wants to trick them to escape, Tamara wants to prove herself, Owen wants to work out why he was framed.

I leaned into this narrative tension, emphasising their differences to portray a group falling apart at the seams, building to a climax with the confrontation between Wolf and Shiva at the end. With the characters working against each other, the chapter now had a forward momentum that wasn't there before.


Decent characterisation doesn't matter if nothing happens during the journey.

However, I didn't want to add scenes that had no relevance to the story just for the sake of giving the characters more things to do. My solution was to bring forward reveals from the end of the book into this chapter.

This included: Wolf being revealed as Mayfen's murderer, Abacus being in the Gutters and Harlin being alive.

I ordered these moments for maximum effect, to trigger the growing rift between the characters. It also accidentally improved the final chapters, as before these reveals were cramped between others, lessening their overall impact, whereas now they had space to breath.


A few rewrites later, something was still missing. The reveals worked, but the parts in between them sagged. These middling sections lacked a clear external threat, and therefore lacked narrative tension. I decided to change my approach, sacrificing some shock value to build to these moments instead.

Originally the Specialists only appeared at the end of the chapter when they ambushed the group in the rainforest. But considering how chapter 5 ends, the Specialists attacking the group isn't all that shocking. So I rewrote the chapter to show the Specialists slowly closing in: Karth executing Polymer, Karth speaking to Vivian in the Gutters in Chapter 6, Perch telling Geoffrey in chapter 7 that the group have been found, Karth watching Emerald from afar... These were all later additions to make the reader anxiously anticipate the Specialists arrival, giving every scene with the group an added layer of uncertainty and tension as you wonder if this is the moment when they will be attacked. This paranoia also mirrors the experience of the characters, which hopefully makes this section more immersive.

I built towards the Wolf reveal in a similar way (so it followed the rule of 3s). The very last PoV I wrote in book 1 was Geoffrey investigating the Silver Crown in chapter 6, which ends with the cliff-hanger of him learning Wolf's crime. Geoffrey then directly influences Perch into revealing this information to the public. By reminding the reader of a mystery it keeps it at the forefront of their mind, making the reveal more satisfying when it eventually arrives.

I did this with Owen's story too, though you will have to wait to find out the outcome of him hacking the President...


This section kept changing lengths: I'd cut down my previous draft by half while simultaneously adding twice as much new content. I repeated this many times, it's length varying between 9-30 thousand words, going down many garden paths and dead ends.

Bizarrely, when I compared a few different versions to see which I preferred, I found the longest was actually the easiest to read, because the increase in characterisation kept you invested, giving the conflicts more weight rather than them feeling like a travel montage detour.


The middle section of a story can often be the hardest to write, because it has the least structure. While I am generally an "over-writer" who goes into too much detail and fluff, I learned through this experience that I sometimes under-write sections too.

Here are some of my big takeaways for any other aspiring writers who might be stuck themselves:

  • If you don't care about a section of your story, the reader won't either. You either need to cut that section out, or change your approach towards it.

  • There must always be conflict in a story, especially in the middle. Focus on emphasising what threatens your characters and what is at stake if they fail.

  • You don't have to save all your big moments for the end, it's much more satisfying to drip feed them throughout the story.

  • Journeys are a great opportunity for characters to reflect on what has happened and what comes next.

  • A journey should cause characters to change their mindset, not just their location.

  • You can add more tension to the middle of the story by introducing a ticking time bomb into the narrative, literally or metaphorically.

  • By focussing on a characters goals and their progress towards them you can give the middle sections more structure and foreword momentum.

  • Event's have more impact if you build up to them, get the reader to anticipate what comes next.

  • Sometimes during editing the story will get longer instead of shorter, and that's okay! Don't be afraid to try a different direction. You can always trim it back down later.

Most of all, I would say don't give up! Some parts will be harder than others, but enough brute force and you can get through it!


bottom of page