Bookworm textual experimentation

Know Your Markets - Part 2

This is the second in a series of posts on using text analytics to understand the short story market in SF/F/H.

  • Part 1 covers the market as a whole.
  • Part 2 breaks down the market by the magazines I was able to sample.
  • Part 3 looks at what's missing, what's likely wrong, and what comes next for this experiment.
  • Raw Data in CSV form for those who want to run their own analysis.

Read Part 1 if you haven't already. It gives an overview of the magazines analyzed as a whole.

This post provides a breakdown of the sampled SF/F/H stories by magazine to see if any discernable differences appear. Knowing differences between the magazines could help a writer hone in on where they're most likely to sell a particular story.

Note that in this post I've removed Apex Magazine, Buzzy Mag, and Shimmer from the charts because they each had less than 100 stories. You can add them back in using the checkboxes, but I felt there just weren't enough samples in each to see anything meaningful.

UPDATE: 2015-01-31 - I recently reanalyzed all of the sites to get some other details fixed (stuff not in this analysis). Apex Magazine now has over 200 stories on their site, so I've added them back in to the charts for this post. Some of the numbers in Part 1 changed a little, but overall the pictures are pretty much the same.

Readability

Learn About Tableau

I picked the Coleman Liau Index for the breakdown because it most closely represents a nice bell curve, but I believe all of the readability measures show the same trend. What I'm looking at are peaks, specifically a central peak, and the range of scores around that peak to get an idea of readability level preference. Apex Magazine, Beneath Ceaseless Skies, Clarkesworld, Lightspeed, Nightmare, and Tor all have similar peaks, around 7.75. Daily Science Fiction and Strange Horizons are noticeably lower at around 6.5. Those values in and of themselves aren’t a measure of good or bad, but to me this says Daily Science Fiction and Strange Horizons prefer slightly less complex writing than the others.

I also noticed that Beneath Ceaseless Skies, Lightspeed, and Nightmare are higher and tighter (peaks reaching around 11% of stories) while Clarkesworld, Daily Science Fiction, Strange Horizons, and Tor are a little lower and wider (peaks around 7%), which makes me think the preference is stronger in Beneath Ceaseless Skies, Lightspeed, and Nightmare.

How could I use this? Say I had written a science fiction short story that met all the submission criteria for Daily Science Fiction and Lightspeed. Which one is more likely to accept it? If the Coleman Liau Index was between 5 and 6.5, I'd submit to Daily Science Fiction first. If the score was between 7.5 and 9, I'd submit to Lightspeed first. If the score was between 6.5 and 7.5, it probably doesn't matter who I submit to first. And if the score was outside the range of 5 to 9, submitting to either is probably a long shot.

Point of View

Learn About Tableau

This was pretty consistent across all of the magazines. There's a clear preference for third person, covering nearly 65% of all stories, and a clear preference against second person at less than 5% of all stories. Beneath Ceaseless Skies has the highest percentage of stories in first person, so if I were to write a story in first person that fit its submission criteria, I would consider sending there first. But overall, I think really it's just important to note that third person sells best, first person sells okay, and second person is going to be a tough sell. Probably not news to anyone who's spent any time trying to sell short stories in SF/F/H.

Dialogue

Learn About Tableau

While all of these magazines seem to prefer 25% or less dialogue, their preferences still vary.

Beneath Ceaseless Skies has a clearly defined peak in the 15-20% range (nearly 25% of thier stories), and Daily Science Fiction has a clear peak in the 0-5% range (nearly 30%). These clearly defined peaks likely mean that the amount of dialogue in a story is more important to these markets than the other. Interestingly enough, Daily Science Fiction seems to have a clear preference for little or no dialogue, but is also the only magazines that has published stories with 100% dialogue.

Nightmare's preference seems to lie between 0-15%, with nearly 75% of all stories falling in this range. Lightspeed and Strange Horizons are spread out a little more evenly between 0-20%, and Tor and Clarkesworld are fairly evenly spread between 0-25%.

If I had a story that was more than 25% dialogue, I probably shouldn't submit anywhere but Daily Science Fiction, and even that would be a long shot. If the dialogue was somewhere between 15-20%, I would consider submitting to Beneath Ceaseless Skies first, and if the dialogue was between 5-15%, I would put Lightspeed at the top of the list.

Words per Sentence

Learn About Tableau

Apex Magazine, Beneath Ceaseless Skies, Lightspeed, Nightmare, and Tor all peak roughly around 16, and range between 10-24. Daily Science Fiction and Strange Horizons peak around 10 and fall between 6-18. Clarkesworld doesn't really have a clear peak, but appears evenly spread between 8-18. Daily Science Fiction and Strange Horizons seem to prefer shorter sentences overall.

As with dialogue, Daily Science Fiction's preference falls toward the lower end but has also published the story with the longest average words per sentence, a single sentence story that clocks in at 534 words. Check it out:

These preferences are close to what I saw with the Coleman Liau Index above, and that stands to reason since number of words per sentence is a component of all of the readability metrics. So I think comparing the average sentence length of a story you're submitting to these results would show pretty much the same picture as the readability metrics.

Syllables per Word

Learn About Tableau

This is not a component of the Coleman Liau Index, but is found in some of the other readability scores. Unlike Words per Sentence, this doesn't seem to show any significant variation between the magazines, with the possible exception of Lightspeed, who seems to have a clear peak at in the 1.3-1.35 range.

Regardless, all magazines fall in the 1.2 to 1.5 range. I'm still mulling this one over, as I'm not sure there are many writers out there who could manipulate the number of syllables per word in their writing. I think this would have to come through refining vocabulary, but I really just don't know.

Conclusions

All of the measures discussed here, with the possible exception of Syllables per Word, show variation in the preferences between the magazines, and could potentially be used by a writer to hone in on magazines most likely to accept a particular submission. But there could be some problems, too. In the next and final post for this initial experiment, I'll share a little of what might be wrong with some of the underlying assumption as well as some thoughts on what could be added to help get a clearer picture.

And like I said at the end of the first post, I don't think adjusting to meet these preferences guarantees a sale. If you can't produce the art, the craft won't mean shit, but the same goes the other way around.

comments powered by Disqus