Post Mortem: Worth the Candle

As I write this, Worth the Candle isn’t quite done, but it’s got so little left to go that it might as well be finished. It’ll clock in at roughly 1.6 million words, having taken approximately four years to complete. This post mortem will give some broad thoughts on what went right, what went wrong, what I was attempting to achieve, and what I would do different. Because it’s such an incredibly long work, I’ll be leaving a lot out, and not paying too much attention to particular changes to scenes or more minor plot points. Spoilers for Worth the Candle follow. This will not discuss anything in the epilogues, which have not been published at the time of this writing.

Content warning: discussion of rape, discussion of suicide

An Endless World

The world of Aerb was, from the very start, meant to be a gigantic world with lots of magic, an insane number of species, dozens of interesting concepts, and opportunities for fifteen million words of content. The particular feel I was going for was “a tabletop game at the end of its lifecycle with every single splatbook added in”, but another way to put it might be “Skyrim with 500 mods” or “the 53rd season of a lore-heavy television show” or “the aftermath of an extended universe with hundreds of books”. Before I wrote a single word of Worth the Candle, I spent a half day thinking up a bunch of magic systems, since I knew those needed to be the most thought out, and ended up with twenty-two in the “core” set. I didn’t start with a world map, nor a list of species, but I knew that it was going to be a huge world, and a ton of species. Most of the species were given a name and a single line descriptor very early on, but the list was added to every once in a while when I had some particularly compelling idea. Much of the worldbuilding was cribbed from a campaign setting I had built and was DMing for at the time, but a lot of the cribbing went back and forth.

When I think about this approach, I often find myself comparing it to Star Wars, which similarly has an enormous number of planets, new and weird technology, and lots of alien species. Star Wars, and a lot of other space opera, gets away with it by essentially being in another genre or style, one where the overall details of the technology and species can generally be assumed to not matter. If they do matter, they don’t matter for longer than an episode or scene. Worth the Candle is fairly committed to not doing that, or at least to attempting to not give the impression of having done that. I was trying my best to have the worldbuilding be complicated, layered, and cohesive. I wanted every species, magic, etc. to feel like it had its own backstory written that I just wasn’t showing on screen. I wanted every scene of rich variety to feel like if you asked me, I’d be able to tell you the story for each thing, why it exists in that time and place, and what’s going to happen with it. Honestly, I don’t know how much I hit the mark there, but overall I’m pretty happy. Having actually spent time writing out backstories and other bits of unseen worldbuilding probably helped a lot, because there were backstories, rather than me needing to make up some bullshit.

Where this creates problems is that not everything is going to be covered in-depth. There are things that show up for a scene, and you think “Yes, more!” and then they vanish, having never been fully explored, sometimes even with the low-hanging fruit unpicked. This, I think, can be a bit frustrating, because it feels like there are wasted ideas, and when it happens every few chapters, it feels like a rush of wasted ideas. I think I mostly hit the notes I was going for there, but it wasn’t without its problems, and over time, it added a real burden onto the story, because at least some of this stuff needed to be tracked. There’s also some problems in terms of infodumping, because in theory you should be reserving that for the stuff that’s most important, but that wasn’t particularly a problem for Worth the Candle, because it delights in delivering infodumps on worldbuilding, and hopefully scared away anyone who wasn’t into that early on.

The big problems came from using the same approach with quests and/or plotlines. Part of this was important insofar as one of the central conceits and themes was unending narratives, itself a reflection of getting into ruts or patterns of repetition, something I strongly associate with depression. I tried my best to signal that not every plotline would get covered, and certainly not in detail, and that not every mystery would get answered, but it’s very difficult to do that within a story … and sometimes I think if you bring up legitimately interesting hooks, people just want to see them, and they don’t care if there’s some thematic or metatextual meaning. (Plus the kinds of people who read web serials want things to go super long, though I think this is a lesser concern.)

I’m not actually sure that I could have done anything about this, once I started with the base concept of “this is a mature world built from a hundred different campaign settings and then set in the detritus of the Chosen One’s forty-year-long adventuring career”. If that’s the starting point, then you’re already painting yourself into some corners. It’s something I was very aware of while writing, both early and late in the project. I think the “big worldbuilding” as a problem is definitely easier to swallow than “lots of plots that are deliberately unresolved for metaplot reasons”. There’s an “I wrote it bad on purpose” problem.

The Pitfalls of Multimagic

Let me get to the fix first, which will suggest the problem: Juniper should have only been able to have four or five magics (or some lower number).

Combat got to be a lot to track. Again, this is in part because of something I really wanted, something that I felt the story wouldn’t be complete without, which was the “tabletop endgame complexity” feeling. I’m less confident that I actually managed to pull this off, but the idea of someone so overloaded with abilities and powers appealed to me specifically because for me, it was a part of the tabletop experience, in a similar way to people being loaded down with wildly mismatched gear is.

(One of my favorite memories of tabletop gaming in high school was the one time we ran an epic level campaign, without leveling the characters through natural play first. We had demigods, four-armed monstrosities, clouds of Ioun stones thanks to a misreading of the rules, and all kinds of things, characters who had dipped into seven different classes, powerful magic items for every slot … I think that’s common, but I have no idea how common.)

The other problem with having loads of different magic is that some of it just never really gets its time to shine. In the Captain Planet approach to things, there are problems for each magic to solve, and I think that this is generally good … but you reach the point where you have to either figure out interesting combinations of powers, or let some of them fall to the wayside and be kind of useless.

Entads (unique magic items) are another problem, for similar reasons, though mostly in the endgame. Each entad is, by itself, like a fully-contained and limited magic system, so each one adds in an option for characters to use. Worse, at least some of them can be acquired through the use of money, which becomes a big problem later in the story when the characters have access to huge markets and extreme wealth. Further, for the number of magic items that should be available to them, they should be dripping in them, and ideally, they would be described in detail for the reader. Unfortunately, that slows pacing to a crawl, and if you do it in a bulleted list, it’s likely to be glazed over. I think a better idea is to just gloss over what they’re actually using, so long as it doesn’t solve plot problems or become super relevant to any scenes it’s in, but I don’t know, there’s no good solution. Again, some of the problem is just what I wanted from the story, which were some things that aren’t terribly good storytelling best practices.

There’s a thing that Jonathan Blow, game designer of Braid and The Witness, will often say, which is that when a game presents him with a concept, he starts thinking about every permutation and knock-on effect and gimmick you could do with it. Games, in his opinion, should explore the possibility space of their conceit. Now you might be thinking to yourself, ‘If you subscribe to that philosophy, then why are there so many dropped threads?’ Well, the answer is that I apply it selectively. Worth the Candle was trying to fully explore everything I wanted from a litRPG, and everything I ever wanted to say about worldbuilding or tabletop gaming, and probably everything I wanted to say about narrative and metanarrative. I think understanding my process in that regard probably helps you understand why I made some of the choices I did. The group being rich and having lots of magic items is probably something I wouldn’t have done, except that it didn’t feel like the story would be complete without it.

Unfortunately, limiting magics is a fix that would require rewriting the entire thing, as would limiting access to entads.

The Cast

Now we’re moving to stuff that’s a bit more difficult for me to say anything conclusive on. Overall, I think that I like the main cast as it is, and if I had to write it over, I would keep everyone more or less the same, but I think there are some tweaks to make.

Juniper: I don’t think I would change anything, though I think I could write him better, having spent four years writing him. It would be pretty sad if I was still only as good as I was at the start. I think there’s something very specific that I was trying to do in the opening book, which was to have someone who is (largely) accepting all this because he only recently decided not to kill himself, and I kind of don’t think that worked at all, and I never heard anyone pick up on it. At least if it didn’t work, it didn’t work in a way that wasn’t obvious to anyone. I also think some of the depression/grief/suicide stuff should have been moved forward a bit, because it’s important to understanding Juniper, but I’m not so sure about that. The other major change to Juniper that I mull over sometimes is whether it might have been better to have him teleported in at his nadir, before he’s had a chance to reassess and reorient. (The timeline for Juniper is such that he’s about two weeks after having attempted suicide and decided that he’s never actually going to be able to go through with it.) I think it would be a more conventional story then, certainly, but I don’t know whether it would have been to its credit.

Amaryllis: I think Amaryllis mostly worked as a character, though there might be a bit of a broken Aesop in there somewhere, especially with the tuung, who never really ended up coming back to bite her in the ass, like you might expect, except a bit off-screen. There’s also a lot left on the cutting room floor with Amaryllis, which I don’t think I consider to be true with most of the others, and is probably why Amaryllis gets more screentime than the others. Perhaps it would have been good for her to have some comeuppance for her competent dictator stylings, but it wasn’t actually her story. Insofar as her relationship with Juniper went, I generally thought “the ice princess melts herself for strategic reasons but is actually asexual” was … mostly good, if perhaps a bit strained and with a tendency for people to misinterpret? Probably the most aggravating for me was the hells time skip mention of two clones who fell in love, which people immediately interpreted as ‘fucking’, going so far as to question why, if she was asexual, she would start a sexual relationship. The text never states that it’s sexual, just that they’re in love, but because the text doesn’t explicitly say that, people make bad assumptions and then read it as a plot hole or inconsistency. I spent a fair amount of time on ace message boards, tumblrs, subreddits, etc. trying to make sure I got that part of things right, at least from a character perspective. Anyway, more conventional story would have had a climactic moment where they confess their love, but the quiet and subdued nature of the various ways in which they knit each other closer together is both unconventional in ways I think are good for the story, and true to life (as they’re partly autobiographical). The only big strike against Amaryllis is that she’s often very similar to Juniper in approach and voice.

Fenn: I liked Fenn a lot, especially as she related to Juniper. She was always going to die, but there was a time when I thought to myself, “what if she didn’t?” or “who should die instead?” Because she dies fairly early on, relatively speaking, some of her stuff doesn’t get followed up on (the Isle of Eversummer and her son, which I would naturally have happen at the same time/place for the sake of economy and drama), and I’m completely fine with that. I think it helps underline the feeling of loss though, and I don’t think either of those plot threads were so alluring that it’s a real shame we didn’t do them. As stated in the story itself, Fenn was meant to have died even earlier, but I got to that scene, with Fenn hastily put back together after having been cut in half, and I stopped myself to think “Juniper wouldn’t just cry about this, not at this point, he would try his best to save her” and then I took an inventory of what was available to him, and looked at my notes, and tried to figure out whether the solution I thought he had come up with would “fairly” work. Thus, Fenn lived long enough to break up with Juniper. Similarly, it’s true that she came back because I couldn’t think of any particularly good reason for her not to. The best one would be “time ran out on her bottled soul”, but that seemed like a problem that Amaryllis surely would have thought of and had a plan for, especially since forward time travel had already been mentioned and (I knew) would be used later in the book. Thus, Fenn came back, which I thought was nice, even if she’s a bit annoying sometimes.

Grak: It took me a bit to figure out Grak, and some of that figuring out was on the page, which I regret. His core concept was “stoic traditionalist dwarf is actually super depressed and engaging in traditionalism out of a sense of masochism”, and I think that got explored in depth, had good build up, and overall … worked? The biggest problem for Grak was that the climax of his personal story was halfway through, and after that I kind of felt like he didn’t have all that much to do. Maybe there were chances to put his personal growth following the Darili Irid chapter on display, rather than showing it in bits and pieces as secondary to whatever scenes he was in. I enjoyed Grak and thought he earned his place in the story, but I think he might feel the most “off”, at least of the main cast, especially because both the start and end are a little iffy for him.

(Oh, and one thing that I might have changed: Grak goes by exclusively male pronouns within the text, except when Groglir is used. The game interface refers to him using Groglir pronouns, which sets this up like it might be relevant or a source of tension or growth or something. I think this is one of the cases of my ambition getting away from me. I hate using Groglir pronouns, or ‘weird’ pronouns in general, especially if they come in as many forms as they do in Groglir, mostly because I think it screws with readability. Online, where I do most of my talking to people, everyone is ‘they’, just to minimize mental overhead. Maybe I should have just bit the bullet on this one, but if I wasn’t going to, I think it would have been better to remove it as a potential issue. Grak isn’t intentionally coded as trans, though can see where someone would have that reading, and he’s not really nonbinary either, since he’s of his own dwarven gender. I don’t know, I think that aspect of the character was underexplored, and if Grak lacked for things to do, maybe that could have been da’s thing, but I don’t know how it would have interlocked with the other stuff that was going on.)

The Locus/Solace: The real purpose of the locus was to offer a counterpoint to the game system attached to Juniper, and in part to serve as a foil to his largely analytical and detached mode of thinking. Solace is an imperfect avatar of the locus, and serves as its ‘voice’ up until her death, and then at least somewhat after her resurrection. Beyond that, the locus drives a surprising amount of the plot, though not through any agency. I think the locus serves a very valuable role in terms of helping Juniper along, and in facilitating some understand of tabletop games, which was important to me, and I think got explored properly. I think the specific ways in which Solace was wrong about the locus could have used more attention, but I kind of don’t think that it should have been seen through the eyes of Juniper, so … it’s hard to say.

Valencia: My biggest regret with Valencia is that I didn’t do more with her combat abilities. One of my favorite scenes from Firefly was River Tam firing three bullets as fast as the gun could cycle and getting three headshots, and that was some of what I’d hoped to do with Valencia combat scenes. Unfortunately, she got almost nothing in the way of combat scenes, for various reasons, and almost exclusively was used for the social aspects, which I think are the conceptually weaker side (though more novel and interesting). Her powers aside, I think her general growth as a character, and function in relation to Juniper, both mostly worked. That said, she has the least to do of any of the characters, and would definitely be the easiest of the companions to cut. The biggest thing she adds to the story is probably an interrogation of rescuing a weak person to have them fall for you, and naïve broken birds in general, and showing what maturity means. I think there’s a kind of reader that doesn’t necessarily want Juniper to sleep with Valencia, but does want Valencia to remain virginal and lusting after him in perpetuity. I also enjoyed the extended discussions of Harry Potter, especially as a recurring joke, but I know for some it was a little one-note and got old. If I had to rewrite Valencia, I would probably make her less powerful on some axis, but not by all that much. (I should also note that the biggest strike against Valencia is that she accelerates through to normality very fast, which is at least partly because I just didn’t want to write someone horribly abused for all that long. The intake of devils ‘helping’ to develop her have the smell of a fig leaf to me, which maybe means I didn’t do a good enough job there. It’s also a big reason a lot of her development ended up being off-screen.)

Bethel: It’s hard for me to lay out how things went with Bethel without making a confession: I didn’t intend for her to rape Juniper until I was writing the scene where it happened. Originally, she was going to come into his room wearing a new body and they would have consensual sex, which would have created a lot of drama and opportunities for character growth and discussion of how Juniper relates to sex and … I don’t know, it’s so long ago it’s hard to remember those plans. What happened instead was that I was trying to write it honestly and get in Juniper’s head, and when I realized that no, he would probably try to put a stop to it, the whole rest of everything clicked into place. I put a lot of time and research into writing that scene, and also tried to write it in a bunch of different directions that wouldn’t include a rape in the story.

Anyway, it made a lot of things click into place, both with cycles of abuse, dealing with trauma, dealing with having wronged people, lots of stuff that retroactively became foreshadowing, and in terms of getting both Bethel and Valencia off doing their own thing (which I don’t think was necessary, and hadn’t originally planned for, but was helpful). It makes Bethel into a much different character though, one that’s maybe harder to understand and harder to redeem. There are a lot of things I like about Bethel as a character, but that one single act almost completely demolishes the conception of her, at least for me. It also means that a lot of Bethel’s original purpose in the story, the stuff I’d intended, got minimized. Letting go of anger, letting go of the past, staying your hand, forgiveness … Bethel still got some of that, but it had to be done outside of Juniper’s viewpoint, and I’m not sure it really worked, so … meh.

Raven: Raven had a few purposes in the story, and most of them are as mirrors to Juniper in one way or another. Her first purpose was, obviously, as Maddie, whose life story she shares in a few ways, and whose personality she shares in fewer (as a result of being comparatively older). Juniper’s relationship to Maddie reflects Uther’s relationship to Raven, except of course that Uther never thought of her in a sexual way. Similarly, Raven’s reaction to the disappearance of Uther mirrors Juniper’s reaction to the death of Arthur. When I think about Raven, I think about her in the context of several chapters, notably ‘No Sleep Club’ and ‘Parallel Lines’. Her tertiary role is as an exposition fairy. When I think about Raven as a character, I think there are two big problems with her. The first is that in practice, she ended up getting a bit too chummy with Amaryllis, and not offering enough of a counterpoint of her own. It was natural for this to happen, given there’s a relatively short distance between Amaryllis and Uther, but I don’t think it did the story many favors. The second big problem was Juniper’s unwillingness to engage with her, which I think was natural for him given how he viewed things going with Maddie, but unfortunate for the story. The bigger problem was that I didn’t lean hard enough into having a reconciliation between the two of them. Even at the end, it was awkward between them. I think if you understand Worth the Candle as being a personal story for me, you can understand why they were never going to have sex, or develop a relationship. Maybe some of that uneasiness seeped into how I dealt with Raven as a character. (Though I do have to stress that almost everything in Juniper’s life is heightened as much as it could be while still feeling true, and the circumstances for me, in real life, were a lot different. I think one of the hard things about the Maddie stuff was knowing how to calibrate its awfulness, and I erred on the side of ‘more awful’ in almost every way, partly because I didn’t want this to be a big thing for Juniper that the audience would have a ‘so what’ reaction to.)

Pallida: I think Pallida is probably one of the most significant side characters in the story aside from Uther, getting quite a bit of screentime and dialogue. I also think that she never really found her place within the team, and nothing ever really materialized from her being there. Honestly, she could have died at the same time Fenn died, and it would have been totally fine, and that was not long after Pallida was introduced. I think I was trying to fill the hole that Fenn left, and it was just grating, because Pallida was interesting in some ways but also insecure within the group and an outsider and probably a few other things. Really, though I’d have to have toned her down so she wasn’t annoying, twelve-year-old Pallida might have been better right from the start, a kind of brash comic relief. I don’t know for sure though. (There was a brief time, shortly after Fenn’s death, when I thought that Fenn could be playing the character of Pallida … but I didn’t see where that could go, or what kind of hints I’d want to drop or where, and while I think it would be a fun twist, the problems with Pallida not really feeling right made it a non-starter.)

Arthur/Uther: Arthur spends the majority of Worth the Candle off-screen, and he gets almost endlessly discussed by various people, through the worldbuilding, in flashbacks to his teenage self, and a bunch of other ways. Then, all the way at the end, he’s there, in the flesh. The ending is the part that I have the least distance and perspective on, but as of right now, it feels like I was fairly successful at doing something really risky.


Given its long length, there were a lot of antagonists in Worth the Candle, and I’m not going to go through them one by one. I think villains/antagonists might be one of the weakest points of the entire story, with a few notable exceptions, and I think I now know why that is. The big, common problem is that you’re never able to get that far into the heads of the antagonists, and secondarily, too few of them are driven into conflict not through ideology or values differences, but through resource wars or other lame stuff.

I don’t entirely know how best to allow the reader into the heads of the villains. I think a lot of my problem when writing Worth the Candle is that I don’t like monologuing and have a tendency to just have someone hold their cards close to their chest if at all possible. The problem is that this sucks for writing a compelling antagonist. Larkspur in particular could have been so much less flat of an antagonist if we had seen some, I don’t know, desperation, sunk cost fallacy, the dark side to Amaryllis’ decisive ruthlessness, etc. Because you see very, very little of that, he’s just some asshole chasing the party. This change wouldn’t have taken all that much, either a scene or two from his perspective (though this would really have broken convention for when that’s used) or possibly some magical line of communication or surveillance.

Similarly, I think that Perisev, as a villain, really would have benefitted from having her motivations front-loaded. You don’t really get a sense for what she’s doing and why until after the fact, and even then, it’s more opaque than I’d have liked. Far too late to implement it, I had the idea of Amaryllis being a spy in Perisev’s base, which might have helped some, but I never went for it.


The game system used in Worth the Candle was intentionally jank. There are a few things that I like about that, and a lot of it doesn’t matter in the slightest, because I had leeway in what I wrote based on Juniper’s stats, and very few people were under stat considerations. What I liked most was the feel of it, a kind of slap-dash bolted-together thing that you could imagine actually having come from a DM houseruling things or whipping up new rules as the situation demanded. I mentioned earlier in this post mortem that we played an epic level campaign where players were walking around with clouds of Ioun stones circling their heads? Yeah, we’d misread the rule and thought that effects stacked, so we had multiples of each. When I played tabletop games in high school, we had a lot of stuff like that. We thought that you could apply a half-golem template to yourself and become ungodly powerful at the cost of just a few level equivalents. And beyond that, there was a lot of stuff I made that was just broken straight out of the gate.

I think, overall, it worked for this specific piece of writing. It was weird and messy, and didn’t always make that much sense, and you had the feeling that it was improvised on the spot, or shakily constructed.

As far as the game messages go, I think they were another thing that I did as much as I wanted to do with them. They served a lot of jokes, from humor, to punctuating a scene, to horror, to a burst of dopamine, to a surprise reveal. There are less and less of the messages as time goes on, until eventually the whole game system gets thrown in the bin, all of which was by design, and at least partly a reflection of how I view the tides of a tabletop campaign, where the numbers are always important, but the early rush of new powers and abilities fades away over time, especially as the characters are developed, and fades into the background.

Speaking of a burst of dopamine, I don’t think the level up addiction subplot worked, and I don’t think that I would include it if I were writing the story over. I liked the payoff to it, but that was mostly me making the best of a bad situation. I don’t think the subplot enriches the Infinite Library subplot, I don’t think it says much that’s meaningful about addiction, it doesn’t say much about how people or relationships work, and I think it’s quite weak commentary on game design. There’s no easy way to remove it from the text, but it’s one of the few things that I would completely excise if it were easy, and if I had some better resolution to the Infinite Library. Maybe there was something more to say with it about murderhobos or something, but it didn’t actually say much, and is one of the few true, concrete failures of the work.

Planning Problems

I originally planned for this to be a much shorter story. Fel Seed was always going to be the end of it, and the Long Stairs after him, delivering Arthur to the end, but … I don’t know. It grew in the telling, at least some of it because I knew I was writing a web serial, and one of the advantages they have is that they can stretch their legs a bit. There’s a lot of stuff you could strip out, if you were making a more streamlined story, including most of the companions, but that stuff is (largely) good, so why would you? Just as I fervently believe “longer is not better”, I’m also a big believer of “shorter is not better”, and so long as something is good, and good within its context, I’ll usually keep it in.

The sprawl did cause some problems though. In the original conception, Juniper would have gotten to Fel Seed much faster, and you’d have two or three less “books” because of that. But because there’s more sprawl, more adventures, more companions, more subplots, that growing distance between the beginning and the end means that there’s a bit of a sag in the middle, not a place where it’s not entertaining, but a place where you’re not driving forward with purpose. There’s something about moving to the end that feels … artificial, like they were moving toward the end because they needed to move to the end. A lot of this is down to some of the choices I made, particularly in terms of wanting a big world, and wanting Juniper to cut his endless adventures “short”, to make a point of not getting stuck in narrative cycles, in laying bare some of the consequences of, say, Superman saving the world every Tuesday, and the threat escalation inherent in narrative cycles and long-running works.

There’s a particular section of Worth the Candle were Things Are Getting Wrapped Up. It was there for a few reasons, but I also think it’s kind of bad. The first reason is that “it’s what the characters would do” is often something that I try to follow, and something I think is perfectly valid when writing characters, because it makes them feel more real and grounded. The second reason is that “going around and cleaning up low level sidequests before the big boss” is a staple of videogames, and something that I had never replicated. A lot of it is … I don’t know. Not “good”, but also not too bad? The big problem is that there’s so much of it, but I think I was getting pinched by the desire to have a humungous world.

I think that’s probably something that could get hacked out, yes, losing gold magic with it, and when I think about what of value would be lost … I don’t know. I think there was some stuff that was good individually but that it was all worse than the sum of its parts. Note that I’m not including a few things that you could place under “wrap up”, like Blue-in-the-Bottle or Doris Finch, both of which I thought went pretty well and did justify themselves.

There are a few obvious fixes for that section, like having more of a time pressure to get to Fel Seed, which would allow you to cut it. Alternately, maybe some parts of it should have (paradoxically) been longer, like e.g. Juniper’s time with Grak fighting the fleshsmiths. There’s a way of showing lots of stuff happening where you show one or two examples in full and summarize the rest, and I think that section had a bit too much summarizing and rushing of things.

Similarly, perhaps the hells arc should have been longer, and perhaps the Fel Seed arc should have been longer too, and …. I’m not confident that the pacing is correct, and maybe there should have been more time for the novel to breathe, especially in those end sections. The problem with the Fel Seed arc and the hells arc is that they’re back to back, and they’re both quite grim. There’s a limit to how much any reader wants to be subjected to that kind of thing, and there’s definitely a limit to how much I want to write. One solution to that, aside from skipping one or both, is to have a breather, and the best breather would have been “Amaryllis and the others grapple with the loss of the protagonist and have to soldier on without him”. I’d have liked that, I think, but it came at a time I was in a bout of depression and felt a lot of fatigue about plodding along with this same story.


I think that you could make a pretty fair accusation that Worth the Candle was trying to be about too many things. It was trying to be about grief, trauma, depression, wish fulfillment, the ways we hurt others, worldbuilding, tabletop games, stories, and the relation of a creator to his work. It’s 1.6 million words though, and intentionally sprawled out to cover lots of ground, so … I don’t know.

Where I think that sprawl hurts the story most is probably the metafiction angle, because when the metafiction stuff isn’t present, it can kind of slide into the background, and the metafiction stuff takes place on different levels, and … there are lots of words spilled by Juniper about his relationship to his creations, and his process of creation, and what he thinks works in terms of storytelling. There are lots of words spilled by characters about narrative, and what it looks like, and what they expect of it, and the fact that they’re in a story of some kind. But that’s still not really the kind of metafiction that hits you at the end, where you’re told that you’re explicitly reading a book.

I kind of think that either that stuff needed to be in there more or less, and I’m not sure which one. Certainly the metanarrative was wearing a little thin by the end (or the stuff just before the end, because I thought the Pure Narrative of Uther was neat and worthwhile). When I first started writing, I had the notion of hiding things in text tags, or writing an alternate story in the author notes, or really going hard in other ways. Unfortunately, putting stuff in tags where people might miss it isn’t the greatest, and it’s also fairly annoying to do because of how AO3 does things, and once I was multi-platform, anything that’s not strictly text got super annoying. I also didn’t really feel like doing an ARG or whatever, not when so few readers would get anything from it, but maybe it would have been worth doing anyway.

One possible rework, which I’m almost certainly not actually going to do, would be to include more of the “Pantheon” stuff right from the start, with regular cutaways to what the Dungeon Master is doing behind the scenes, and sometimes with the Narrator and Dungeon Master talking together … or something. Done as commentary on the work and teeing up for the ending, maybe it could work, but it might also just be annoying.

I think if I had to do it over, I probably would have also cut the two/three error messages. I think they feel like misdirects, and don’t do enough to spur on simulation/creation/narrative talk, especially not when you have different “meta” stuff like the conversations with the Dungeon Master, certain events in the story, anagrams/puns, narration fourth wall breaks, the game layer … I understand the reason the author (me) put the error messages in, but I don’t think that they do enough to justify themselves, not when there’s so much other stuff that could have been used for roughly the same intended effect, that of .

Dropped Threads and the Cutting Room Floor

Over the course of 1.6 million words, there were a lot of things that didn’t pan out, or got cut, or had to abruptly change. I’ve said above that it was meant to feel like there were lots and lots of plots that potentially could have happened, and lots of places to potentially go, and like a sandbox the characters were playing in. But also, some of the things that were dropped were actually dropped, planned and then abandoned.

There came a point in writing, some time after Fenn’s death (and probably after the Library), when I realized that I needed to stop opening new threads and start closing them, and from that point on, I mostly did. Unfortunately, this came after the introduction of a number of characters and concepts that I really did think that I would get to, for the most part: Thargox, the Other Side, the Outer Reaches and the schlossvolk, Heshnel, Pallida, Gemma, the Void Beast, etc. Some of those I thought would be intentionally left to the wayside (especially the Void Beast), but that’s a hard thing to signal.

So which things suffered the most from being cut? Which things were background and which ones were planned? For the most part, I never introduced a quest without at least thinking about two or three interesting stories I could do with it, but there were a few that I had plans for and in a better world, would have been able to gun for. Burnout was a real looming threat though. Here are the big ones, in brief:

  • Thargox is a hive mind, one that people voluntarily enter into, usually for relatively short periods of time. If you look at Worth the Candle as a metanarrative, primarily as a story about stories, then you probably know where this is going: Thargox represents the fans, and the way that people can opt into sharing a bit of the fictional universe in their own heads. The lesson for Juniper there is why people do this, and that it can be a good, productive thing, but also that there are ways in which it’s unhealthy. Becoming part of Thargox can be an escape without really dealing with your problems. Unfortunately for Thargox, it wasn’t particularly vital to be included, and Thargox as a semi-benevolent force would have needed some good hook to be brought in and have the concept explored. Very easy to cut.
  • The biggest thing that was probably lost was probably the Other Side, which I had lots of plans for, but which didn’t fit into the work in a way that I found satisfying; B-side Juniper was playing the isekai harem role much more straight, and I think might have been able to hold up a mirror to Juniper and his adventures in a way that would have been neat to see, but I also wanted him and his story to be more complex, to be, essentially, in a different genre from either our Juniper or Uther.
  • Honestly, I think the Outer Reaches being an off-screen thing that Amaryllis went through worked, but the original plan was to just kill Juniper off and not do a time skip, slogging through those three years he was gone. Would this have been great? Yeah, maybe, I don’t know. Would it have fucked up pacing and made for a weird break in the story where it becomes a different thing, only to return later with a Juniper we barely remember? Also yes. And remember, this was coming at a time when I was feeling a lot of burnout, and the end that I had envisioned three years prior seemed like it was in reach. Starting the Fel Seed arc meant, for me, that I was finally getting to a part of the plot that I had sketched in my head from the very beginning. I don’t know that you get the full sense of the weight of those three years, and I don’t know that I had the proper sense of it because I didn’t write any of it, but I’m happy I didn’t do it, partly because I’m not sure I could take the strain.
  • Heshnel, Pallida, and Gemma all had specific roles imagined for them, and I unfortunately didn’t think that any of them were working. Part of the problem was that the cast was too big, and I kept feeling like I needed to push someone off to the side. Heshnel was interesting to me as a defender of the Second Empire, which itself is a reflection of the way that Juniper, Amaryllis, and Uther think. Pallida was interesting to me as someone who is essentially stuck in the same kinds of cycles as Uther was, but in a much different way. And Gemma … I don’t think that we even got that much of her ‘deal’ on-screen, because there was so much else going on all the time, and so little room for it, but in the same way that the tuung society is an act of creation, Uther’s Fox Guard is too, one five hundred years down the line with its own seething resentment of its creator (for various reasons). In the end, I think I could have killed off all three of these characters during the fight where Fenn died and lost little, unfortunately.
  • The Void Beast was never meant to be solved, but I think that maybe I should have spent more time on it. I don’t know. The problem, as pointed out in the text, is that there’s no good solution to global warming. People will favor the short term over the long term, they’ll balk at coordination, the rich aren’t really in danger because they can escape, leaving the poor, as always, to suffer … call me a doomer, but this kind of thing weighs heavily on me. I think global warming in the real world will probably be “solved” by large-scale geoengineering to cool the planet, but that scares the shit out of me, as I think it should scare the shit out of anyone. The other major option is some kind of AI hard takeoff, which I think less likely, and also scares the shit out of me. So the big problem with the Void Beast is that it’s a metaphor, and it felt false to have any actual solution. Plus, it provided some time pressure, which I think was a good thing.

Mirrors and Symbols

One of the things that I was fairly deliberate about in Worth the Candle was trying to set up mirrors, reflections, and symbols. I think that mostly this was a success, though the fact that it was metafiction meant that some of this stuff was pointed out in the text itself, which I think is fun, but not necessarily the best way to go about things, and not repeatable unless I want to write metafiction for the rest of my life, which I don’t.

One of my big philosophies with symbolism, mirroring, etc., is that everything has to first and foremost exist on its own terms. If we have a thing, let’s call it μ, then it needs to be μ before being a metaphor for something else, or some insight into a character, or anything like that. μ needs to exist on its own.

There are a few things that get reflected or mirrored a few times over, and those make up the major themes of the work (and in my opinion, this is one of the best ways to develop themes).

  • Loss: Juniper’s loss of Arthur, Grak’s loss of his home, Raven’s loss of Uther, Amaryllis’ fall from grace, Bethel’s trauma at the hands of Uther, Juniper’s loss of Fenn, and probably some others
  • Trauma: Juniper’s abuse of Maddie, Uther’s abuse of Bethel, Valencia’s abuse, the abuse of the locus, and probably some others, Fel Seed
  • Creation: The Dungeon Master’s creation of Aerb, Juniper’s creation of campaigns, Amaryllis’ creation of the tuung society, lots of the ‘what to do as god’ talk, and probably some others
  • Suicide: Juniper’s attempted suicide, the tuung’s doctrine of supremacy of existence, the Harold cult’s doctrine of supremacy of oblivion, Grak’s suicidal gestures, Blue-in-the-Bottle’s zombies, and probably some others
  • The Intolerable: The Hells, Doris Finch, Uther’s narrative cycles, Juniper’s depression, and probably some others

All these things need to exist as themselves first and foremost, so some of the mappings have problems, but that’s a feature, not a bug, and there’s an undercurrent in the work of reading tea leaves and seeing what you wanted to see. Characters are seeing patterns, sometimes where they don’t exist. Often these connection were just things I “found” in the writing while trying to see things from a character’s point of view. This biggest (or at least most recent) was Uther thinking that the entire party coming to the Long Stairs were meant for him; their individual ‘roles’ that Uther assigns them weren’t something that I had planned a million words back, they were just a result of me thinking ‘this is what Uther would think, wouldn’t he?’. And it works well within the story, I think, because when Uther makes the argument, it sounds plausible.

Bonus: Highlights

Looking back, trying to think about my favorite parts of Worth the Candle, and with the caveat that I have not read it, only written and edited it with occasional spot reads:

  • The chapter ‘A Pleasant Interlude in Kansas’
  • Fenn’s letter in ‘Letter 15’
  • The chapter ‘A Cypress Waits’
  • The chapter ‘Darili Irid’
  • The Mome Rath fight
  • The Library fight
  • The Onion fight
  • The Cannibal fight
  • Getting high on unicorn blood
  • Saving Fenn’s life
  • The chapters ‘The Adventures of Valencia the Red’ and ‘The Further Adventures of Valencia the Red’
  • Helping Doris Finch

Hopefully that’s not too many highlights, and hopefully I didn’t miss any obvious ones. These are the ones that, looking back, came out how they were in my head, and there’s a bit of a bias towards those with a bit more action or at least positivity. The chapter ‘The Fel Seed Incident’ came out just about how it was in my head, and it’s a good chapter, but it’s the kind of thing I don’t think I would look forward to on a re-read. Similar, the chapter ‘Maddie’, or the chapter ‘Long Live the King’.

Bonus 2: Stuff to Steal

There are a ton of ideas in Worth the Candle, many as part of having a huge world. As stated, not all of these were explored, nor were meant to be explored, and some of the ones that were explored are, I think, the kind of things that I would see as the sort of things I’d like so see stolen and used somewhere else.

  • Exclusion zones: In some ways, I see these as an evolution or refinement of or take on an idea in qntm’s Fine Structure, which I was reading as it came out in (roughly) 2006. I think exclusion zones are great, and sometimes they’re a thing that fantasy settings do anyway, except that this conceit makes a bit more sense. Why have a weird forest where the rules are different just because when you could have a weird forest which is only contained by the will of the gods, or whatever other implementation you decide on? There are a lots of different takes you could do with exclusions. The other major inspiration for exclusion zones, which I just now remembered about, was a CRPG I was playing in 2017, Tyranny, which has Edicts that are kind of (sort of) the same as exclusion zones, if you squint.
  • Entads: I actually think that the entad system is something that a lot of works of fantasy do anyway, it’s just not being all that mysterious about it, or being mysterious in different ways. In the original tabletop game that entads were taken from, they were created by small gods with vested interests in niche areas, but forge frenzy is another way to do it. Having magic items work like this means that you can have cool stuff that you don’t really have to explain, and there are no problems with just having things that are interesting for their own sake.
  • Pseudomagic: The conceit of pseudomagic is that sometimes people are Just That Good. I think it’s relatively underexplored in the setting, with the exception of the bladebound, but there are references made to other people who have reached a level of mastery over their chosen field which allows them to do impossible feats. This, again, is something that got pulled from other works, where it’s less questioned, so maybe stealing this is a bad idea unless you’re also stealing the execution, and maybe the normal execution is fine.
  • i-level: I can’t recall how much this comes up within the work, but it’s definitely spelled out within the worldbuilding document. i-level is a scale used to determine how powerful things are, and compare them against each other, especially disparate effects. It’s primarily used in entad studies and ink magic, but I think there are loads more things you can do with it. It is, of course, distilled out of tabletop games, but as a meta rule for entads, systems of magic, people, etc., I really like it.


I’m happy to be seeing the end of Worth the Candle. Looking back, I like the work more than I thought I would, warts and all. That said, I am eager to see the backside of it, and to be able to go without thinking about Juniper, and Arthur, and all that. I want to write something less personal, though I think that writing, for me, will always be personal. I definitely want to write something shorter, though I have an affection for web fiction, and do believe that one of its major strengths is that it can run for so long, and become so rich and lived-in, like a stew that’s been simmering for a long time.

I hope that this post mortem doesn’t seem like it’s too glowing. I’m used to getting done with something I’ve been working on and hating it intensely, because I’m able to see all of its flaws, but here, I find my view of the flaws being a bit blunted, and my appreciation of the strengths heightened. Maybe it’s because a lot of it is so old now that I’m mentally past the stage where I hate things and into the headspace where I like my own work as something that was written for someone exactly like me, a guy who gets all the references I do, who writes in a style that I like to read, and who draws on an exactly identical cultural canon, even going so far as to build in little in-jokes that only I can appreciate.

If you read Worth the Candle, I hope that you enjoyed it, and that the bumps and warts didn’t bother you too much. I hope that you got something out of it. I hope that I made it feel real, even if just for a little bit.

If you have found a spelling error, please, notify us by selecting that text and pressing Ctrl+Enter.

Post Mortem: Worth the Candle

24 thoughts on “Post Mortem: Worth the Candle

  1. Personally I think the story was just too long, it dragged on at times with plot lines that weren’t particularly consequential. The whole section of “Dropped Threads”, I think it was excellent that each of those were dropped. Just enough information was given about each thread to make the world feel big and fleshed out and give that world building rush, but not enough time was spent on them that they felt like a distraction or boring. I think several more subplots should’ve been given the dropped thread treatment and just left to exist in the background and left unfinished because Juniper’s able to finish the main Fel Seed quest before they do anything terrible. Especially since you could always return to them as a prequel/sequel/side story and create new characters for whom the dropped thread is extremely significant, instead of having Juniper deal with them, since looking at problems through Juniper’s eyes does get boring and repetitive after a while.

    If you have X amount of action and interesting plot points, it’s a lot more exciting to pack that X amount of action and interesting plot points into 800 000 words than 1.6 million words.

  2. I really enjoyed this, start to finish.

    Speaking of pseudo magic, I was once discussing game design and offered to demonstrate something. The other designer said “just because you can do it, does not mean it is [should be] possible.”

    I took that to heart later, including when I wrote the Expert Set for D&D.

    Anyway. I think you did an excellent job and I really enjoyed your retrospective here.

    Two thumbs up.

Leave a Reply

Your email address will not be published. Required fields are marked *

Scroll to top

Spelling error report

The following text will be sent to our editors: