Leo X. Robertson

News of my latest publications, events, and episodes of the Losing the Plot podcast!


writer life

Story Systems, Part 5/5

In this series of posts, I’ve invited the reader to see a story as a “meaning-generating system”, and story-writing akin to a simulation that requires good initial conditions to run properly. Now I have some more thoughts on this analogy, on reading other people’s writing, and some final messages.

Sometimes simulations are chains of smaller simulations. Simulations within simulations!

Example: Franzen said “The Corrections” is “five linked novellas.” So then we can break it down into five components, the input of novella 1 sent to novella 2 and so on. Those novellas have chapters, and those chapters have sections, each sub-systems running different sub-routines for some greater whole, generating meaning and passing it from one simulation to the next.

Dostoyevsky’s tomes are usually split into “Book the First”, “Book the Second” etc, and the same principle applies.

Short story writers attempting novels for the first time often refer to them as a series of short stories. And fair enough.

The analogy holds up no matter how big or small the system is! At one end, the story ideas you don’t pursue are the inert components that don’t do anything in the reactor. At the other end, 1000-page novels are factories, with systems and sub-systems that can be scrutinised and re-run to test the effect on the whole machine. Book series are like systems of factories in a production line.

So we’ve covered that most of writing is re-writing. Similarly, reading is re-reading.

Once you can see other people’s writing as a mixture of different components, you can “uncook the ingredients.” My first dumb analogy, but you get what I mean.

Very few stories don’t reveal their mechanisms on repeat readings. And, much like you read through your own story drafts collecting different errors, you can read through other people’s stories and focus on different aspects of storytelling: plot, setting, characterisation, structure etc.

Once you have the components, in your own writing you can say to yourself, “I think what we need here is to run program(Cormac McCarthy).” In William Gibson’s Paris Review interview, he referred to other authors as “pedals”: “revving Ballard” for example.

This may seem cynical—but applying an author’s style is just choosing a sub-system. The overall system will still be unique if the combination is different. And this combination is just one level on which you are making the decisions. It still represents your style.

When I read my stories back, I can feel in different paragraphs which real-life event, thing someone said, person I was thinking of, film I saw, story I read, thing that was happening to me at the time. All these are just a unique set of launching pads.

Okay, one final thought on the re-iteration process of redrafting: remember that whatever flaws are there are going to be most apparent to you, and that it’s impossible to read your own writing with the freshness of its first reader. You may get this way in the future, but it’s not worth considering. I’d advise focusing on the secondary joy of watching the thing get better, better, better—but never perfect.

Even if you’re less aware of the errors than an objective reader, only so much rigor can be expected of you. If you’ve seen the film “Annihilation”, check out this review of it, replete with error messages!

The complaints are valid, but the film suspended my disbelief enough to sweep me away. I loved it! And yet in many ways it is quite imperfect. Could do better. Who couldn’t?!

I think this is interesting: these errors meant that the film “didn’t work for that viewer.” But it still worked for me, either because I’m denser or more forgiving. I prefer the latter.

With more writing rigor, the film could’ve won over a larger audience. Those errors that I didn’t notice would’ve been corrected. That would hardly have impacted my viewing experience, but would’ve satisfied those more nitpicky folk.

There’s too much emphasis on subjectivity. It’s offered to easily appease people who didn’t do as good a job as they could have, and that review is a good example of what it means when a story “doesn’t work” for someone. Failure to resonate may be a question of rigor.

This also reveals how to interpret rejections. It’s rare that a story is a complete outright failure—but the more prestigious the publisher, the more rigor will be expected. And rigor can be approached systematically, using the method of these blog posts.

Writing is a skill you can learn like any other, and improving at it is tangible.


  • You decide what works for you.
  • Rigor is not the same as bullying yourself.
  • Writing is always a light and inviting thing filled with the reward of meaning, requiring trust and curiosity.
  • The easiest way to scare away trust and curiosity is to bully yourself and give up hope.

So believe in yourself, dirtbag!!

Story Systems, Part 4/5

So far, I’ve discussed how I’ve started to look at writing stories, and a way to prepare to write a first draft and how to edit it once it’s in existence.

(As for how to make the first draft? Just type!)

Here’s how I’ve developed this way of writing, and why I think it works.

Last year, I started sending stories to the most prestigious mags I could find. This made me terrified of writing badly. Especially if I combined this with reading the magazines while writing a first draft. “Look at this garbage I’m writing compared to what these writers are capable of!” I thought.

Anyway, my fear of bad writing meant that when I’d written what I thought was a finalised story, I didn’t want to probe it too hard lest it fall apart.

I also came up with the idea that preparing as much as possible before starting a story made me a bad writer.

I was probably thinking of this quote: “[Chekhov’s] friend and fellow writer Vladimir Korolenko wrote in his memoirs that when asked how he wrote his stories, Chekhov laughed, snatched up the nearest object – an ashtray – and said that if Korolenko wanted a story called The Ashtray, he could have it the next morning.”

I’m willing to bet Chekhov spent most of that day taking notes though! Doing a bit of free association, thinking of the role of ashtrays in his life. Did anything curious happen to him involving an ashtray, or when one was in the room? What is his opinion of them, of smokers?

Ray Bradbury did the same thing. He started off writing stories starting from just the titles. So I thought, “This is how you’re supposed to find the good stories.”

I do think the best moments of my writing come straight from the subconscious. When I’m writing a first draft, they’re when the story seems to “take off”, writes itself without me. (I hope you’ve had this feeling! It’s the best.) They’re the parts of my writing that even I can’t explain. But when I look back on when those moments occurred, they were when I’d finally found the time to write a story I was super excited about, because I strongly suspected that it led somewhere interesting. I’d been thinking about the story, or the topic the story is about, for months—not consciously, just in the back of my mind along with all my other thoughts. All those cumulative fragments of thought poured out for those stories. However, conscious effort, in the form of preparation, can be done to increase the chance of a first draft “taking off.”

This emphasis on preparation is present in this blog post, which is one of the best things I read about writing last year.

His advice about the characters didn’t quite click for me, which is why I prefer to think of story preparation as a series of questions to answer, rather than real people to interrogate.

I even have another dry scientific analogy here!

Crystallisation is when solid crystals come out of a solution. This can be done “spontaneously”, just by waiting in hope that crystals will appear in the right conditions; or, by the introduction of “seed crystals.” Seed crystals are small crystals of whatever compound is in the solution, and they’re added so the crystallisation process has something to grab onto. This is a much faster process!

So, if you make a small version of “story stuff” before you begin the story creation process, you will create the story faster.

This isn’t the same as predetermining where the story should go. Having a planned trajectory, so you don’t go nuts in every direction, is placing a seed crystal in a solution. You might get there without it, but you’ll get there faster and with less effort if you have a good seed 😊

Back to these stories I sent out. If anyone was kind enough to give me a personal rejection, they suggested parts that didn’t work, that didn’t make sense, and the ways the story could be improved.

Very important: if those improvements resonate with you, if they were things you had half-suspected yourself or that make sense in retrospect, you can apply them. If they don’t, don’t bother with them.

The more comments you get, the higher chance that some of them won’t apply to what you were trying to do. Because the more comments there are, the likelier it is that you didn’t get as close as you thought you had—or that an enthusiastic editor is shaping your work in their vision.

Even better, if you train yourself to ask the right questions about your manuscript, you can answer them before anyone else has the chance to ask. Whatever you have in-house can be as bad as you like. Who cares?!

Much of the material I wrote in shaky deer-mode is still with me. No one yet wants it! And when I revisited it with my new “curiosity-exhausting” method, a whole string of questions popped up.

Revising is a weird sensation. You’re correcting what you saw the characters first do, and the world they did it in. Maybe even the outcome of their whole adventure. But as you do it, you can feel the meaning of the text expanding, resonating more deeply, becoming more what it was destined to be overall.

I’ve not yet found a weakness in a story of mine that wasn’t also a previously hidden opportunity to make it stronger than I had imagined.

When I started writing something new after developing this idea, I was more willing to change the components as I was writing it. The closer I got to the end, the more I could tell what was and wasn’t needed.

This openness to change has left me with a first draft that appears much closer to the end product. As I wrote, the changes I made got smaller, more specific, fewer. I was iterating towards a correct solution.

One of the stories I repaired this way, which was rejected a handful of times, has now been shortlisted by a fairly reputable mag. That at least means someone will take it as is—or someone else will after a few more passes through. It definitely means I did a good job! (And if you want to keep writing, you’ll have to start celebrating more than just those rare acceptances.)

In the last blog post in this series, I’ll offer some ways that you can expand this analogy and what to take away from reading other people’s writing. Thanks for reading this far!

Story Systems, Part 3/5

So far, I’ve introduced an analogy for how to think about stories and the writing process. I then discussed in more detail how this helps to think about the writing process: initial conditions, running a simulation.

If simulations are set up improperly, the software returns error messages. What are error messages in the context of writing stories?

Here are some from the current draft of the story I’m writing:

  • Rewrite opening scene once we know main character better.
  • Not quite hitting the emotional beats properly overall. Verify on re-type.
  • [Sub-character1] characterization?
  • [Sub-character2] characterization?
  • Send story to dad when done!
  • More characterization of protagonist here. A little more about what she wanted to do.
  • Ending isn’t quite right.
  • Maybe [this idea] doesn’t work. Think more about [that idea] instead.

The reason I’m getting these “characterization” messages is that I’m revising a draft of something I wrote before coming up with my current method. I now have to go back and do work I should’ve done before starting, in order to rewrite what I have. It’s clear to me that I don’t know these characters as well as I have known other characters of mine in other stories (which is, of course, my best gauge of how good a job I can do.)

“Send to dad when done” is not really an error(!) but a great sign—a highly marketable story is clean enough that you would let your own parents read it. (I don’t see this as a compromise, though: I’ve had a great run of violent and disgusting writing—and may have many more!)

All the text in my drafts is now blue, by the way. That’s so I don’t see it as final. I like to type everything because it’s so easy—but it requires extra discipline not to see a story as complete for as long as possible.

Sometimes I’ll wake up and write notes in my phone of something I need to add to the story draft—and when I look at the draft, I see that same note already incorporated. That’s a great sign! The subconscious is almost nagging you to make sure the story is as it should be.

These messages also require me to take out ideas that I liked at the time because they no longer fit with the story’s new direction. I don’t ever feel a “click” when a story is done, but when I make these decisions, I feel the story getting better, better, better… what once was flat becomes alive. The characters feel more real, their decisions gaining weight, the settings more detailed. It’s a real joy.

There’s joy to be found in all parts of the writing process, from first draft to final edit. I think why authors think they dread writing so much is that it’s always tough to start. And you spend most of your day not starting, therefore you’re inclined to let the dread of not-yet-having-started mask the joy of creation.

These stories I’ve repaired recently—I had no faith in them at all. This was a great place to work to test this theory out, because these were the easiest to work on extensively. I didn’t have anything to ruin, or anything to lose. All to gain. And it’s not even like I changed all that much in the end.

When I read through, I typically collect these error messages at the top of the document. If they pertain to specific areas in the story, I’ll copy and paste them there. If, like “characterisation”, they require me to think a bit more about a particular aspect before I revisit the story, I’ll type them out where the error is. This is good enough for me. If you want, you can categorise these errors:

  1. Structure:
    • Need for opening scene?
    • Chapter 3 too long: split into two.
  2. Middle:
    • Scene 3 and 4 can be combined.
  3. Grammar:
    • Is it “lay” or “lie”? I can never remember…
  4. Character:
    • Steve doesn’t serve a purpose significantly discernible from Joe. Combine.
  5. Senses
    • Add smell/colour to beginning section

Maybe you have more diffuse categories related to elements more specific to your story. Up to you 😊

How do you run the simulation again? After incorporating the error messages into the draft, and making as many edits as you can along the way, you read it through another time. I would recommend typing it out from the beginning, quite a literal re-running of the same simulation. If it’s in a worse state than that, this blog post explains how best to prioritise what to fix when.

How do you know when the story’s done? Firstly, you’re not getting any more error messages. And you will likely notice, with satisfaction, that you get less and less each time you go through it! You go through it again and can’t think of a single improvement—or, remaining improvements are so tiny and yet would require so much energy that they can be forgone. It’s then ready to send off when neither you nor a few close writer friends can think of anything worth pursuing.

As far as I can tell, this way of thinking about a story is similar to George Saunders’ method. You can see it at work in most of his stories, but I think “Pastoralia” is the best example. All characters have understandable motivations, which propel the story in the right direction, and the reasons for the setting are clear. And he is a notorious draft iterator!

Todd Solondz, when writing a screenplay, said he sometimes had to add more “gas to the tank” if the resulting screenplay didn’t hit feature length.

There are surely endless white men for whom this process has worked!! 😀

Next time, I write about how I developed this way of looking at writing, and what personal evidence I have that it works.

Story Systems, Part 2/5

In my previous blog post on this topic, I explained how stories are analogous to computer simulations, in a way I thought might help people who think like me. In this post I explain how to set up the “initial conditions” for a new story.

I vividly remember trying to plot a full novel using different post-its for different characters. I wrote up drafts for the first scenes of each character, and they stayed on my wall for a month. I didn’t get any further with it, because I hadn’t written those scenes. I just had my initial conditions, but I hadn’t let the simulation run. Once I wrote up those scenes, the next steps became apparent.

What do good initial conditions look like? To me, it’s notes on plot, character, theme and setting.

(Research is often a big one for me, and a big competitive advantage: few people do the initial-condition due diligence required to start writing something heavily researched.)

What do I know about these characters? Who are they, were they, etc? Where will this story take place and why? What is going to happen? What do I think this story is about?

I answer as many of these questions as I can before I’ll get to typing out anything of the first draft.

I don’t see how someone can write a good story without knowing more about the characters than appears on the page. When it’s done, all a reader’s questions must be implicitly answered by the resulting text. Better to answer as many up front so as to minimise work later on.

And this isn’t a rigid, systematic approach: I don’t know if your story takes place on Earth, if your characters are human or a collection of bitter shoeboxes, or if the plot takes the path of a fractal. There is still infinite scope within this methodology. After all, there are many types of reactors: plug-flow, continuous stirred-tank, semibatch, catalytic. They’re different shapes, and mix different reagents, but they all contain reactions and operate on the same basic principles, which underpin the great machines and chains thereof that lead raw material to product, from beginning to end.

You may want to add “structure” notes, but nothing irks me more these days than cleverly structured stories: lists, FAQs, whatever. I try to avoid anything that reminds me that this thing didn’t really happen, and anything so calculated and transparent just never resonates as much as it could.

(Side note: I think list stories sell so well because readers know when the thing is going to end. I have to imagine they’re written for people who can’t wait to stop reading.)

And if you’re anything like me, you most certainly will not make any notes on genre or audience, what market you will send the story to when it’s done. I mean, you may wish to add certain definite components to the initial conditions if you’re writing for a specifically themed submission—but anything more specific and you can correct it later.

So, when you write a bit about the plot, or the setting, more characters appear. They need characterised. Details of backstory appear. They need added to the plot. They have settings. Those settings have characters. And so on and round and round.

How do you know when you’ve done enough? Because nothing more occurs to you. It sounds, through my outline above, like it will keep going indefinitely—but background characters don’t need as much work. Some scenes take place in the same setting, and so on. There are definite limits. A story itself, no matter how big, is finite.

Next time I’ll write about re-drafting 😊

Story Systems, Part 1/5

I’m getting better at writing short stories. I attributed my increase in success thus far to the notion of “choosing better premises”, coming up with more ideas than necessary and developing those that seemed to have the most energy to me at the time. This is a way of “pre-vetting” a story before you’ve even written it!

What I’m doing now—one of my favourite early-in-the-year exercises, is improving stories I haven’t yet had published. Doing so has taught me what I can do beyond “choosing better premises” to take my stories to the next level in a way that feels intuitive to me. I’ve come up with a method that seems technically rigorous but airy enough to let the subconscious in. In the following series of blog posts, that’s what I want to share.

(And when I say “intuitive”, it means I won’t seem to be saying anything you don’t already know—maybe just articulating it in the way it makes sense to me. An attempted demystifying process.)

They say rejection is the norm, and it’s true that most of it will be out of your control—and is thus not worth worrying about at all. That doesn’t mean we get out of mitigating what of the process is within our control.

So, here’s an intro to how I’m starting to see stories.

When I studied engineering, I did these cool (I thought) reactor simulations. You would have a bunch of fixed “initial conditions” that you would feed to your reactor model—temperature, pressure, concentration of reactants etc—then you would let the simulation run and it would tell you how the mixture of components in the reactor changed over a week, month etc, and what you were left with at the end. If you set up the model properly, and provided enough initial conditions, and those conditions were correct, cool—you’re done, write it up. What was more likely, at least on the first attempt of a new simulation, is that the thing would come up with a bunch of error messages. You didn’t set enough boundaries, so the simulation returned an impossible solution.

Some simulations I had to do were 2D cross-sections of reactors. When they were done, they’d show, for example, how concentrated various particles were in different regions. The images looked like thermal imaging, showing pockets of red where the particles concentrated, the swirls of eddies coming off them diluting into an un-mixed blue. The images would look rough at first, but I’d set the simulation to perform 1000 iterations of the same calculation, over and over, until the image became clearer.

The reason the simulations had to be repeated over and over was because the calculations they used were just approximations of what might happen in real life. Nature by comparison is so complex that it could only be approximated, in these cases, by calculations.

And the calculations had ways of measuring how much error their images contained. Rules of thumb told me what size of error was acceptable in order for me to trust the image I was seeing. I could decide if that was good enough for me and move onto the next simulation, or if it was worth my time setting up the simulation to run for a further 1000 iterations.

In this analogy, the reactor simulation is your subconscious. The initial conditions, the components you feed to it, are the components of story: plot, characterisation, setting, theme. The story is a meaning-generating system.

Any story can be corrected, but those stories that are set up best at the beginning will have the least error messages. Also, it may be the case that you run it once and it comes out perfectly—but whether it does or not is more likely a case of guesswork than a question of your skills as a storyteller/scientist. Sometimes I’ve gotten lucky and guessed it all almost correctly. But very rarely. And is that the point, really? No. The point is good results, not how they are obtained.

Scientists have hypotheses: what they expect to happen. Similarly, I’d advise preparing a rough plot outline. You haven’t killed your creativity if you end up right, or close to right. It just means that you’re a good scientist!

It’s perfectly legitimate to send a short story out if you determine that one fix might take another month’s worth of work, say, but add very little to the overall picture. No matter what anyone claims, no story is perfect. Maybe one definition of perfection would be, “A dead-on match for what would happen in real life.” And there’s no way of knowing that, but there are damn good ways of approximating it. And like my simulations, approximations are good enough for their purposes.

The point of story, I think, is that it’s so close to life that it’s good enough. Maybe even scarily accurate. But never perfect. So close to the truth that it resonates with people you’ve never met.

So this part is just me setting up my analogy. In upcoming blog posts, I’ll outline how the way I write applies this way of looking at story.

You can follow this process without liking my analogy. You can even write something good without following this process, but I bet it would take you longer. It sure took me longer without it! I’ve found it’s best to do this stuff up front because you won’t get out of doing it later.

So I hope you enjoy these posts, which I’ve scheduled to release this week and next!

Website Powered by

Up ↑