Algorithm predicts book success based on word usage

[Devor]

Saw this this morning.

Will Your Novel Be a Best Seller? Ask This Super Accurate Algorithm

I don't have a comment, but I'm sure the rest of you do.

 

[Mythopoet]

Just one: lol.

 

[Twook00]

So, what makes a best-seller? There are a few key findings, according to the researchers:

Successful books make heavy use of conjunctions—like "and" and "but"—as well as large numbers of nouns and adjectives.
Unsuccessful works include more verbs and adverbs, explicitly describing actions and emotions—like "wanted," "took" or "promised."
Verbs in successful books more commonly describe thought processes—like"recognized" or "remembered."

So essentially it is distinguishing between well written books and poorly written books?

 

[BWFoster78]

There ya go. Proof that you shouldn't use adverbs.

 

[AnneL]

Wow, I overuse "recognized" and "realized" and "remembered." Guess I'd better start shopping for my mansion.

 

[BWFoster78]

Wow, I overuse "recognized" and "realized" and "remembered." Guess I'd better start shopping for my mansion.

Congratulations on your future success!

I'm editing as we speak to add more of such verbs.

 

[T.Allen.Smith]

There ya go. Proof that you shouldn't use adverbs.

Finally....validation!

 

[Ireth]

There ya go. Proof that you shouldn't use adverbs.

I guess the Harry Potter books are an exception to the "no adverbs" rule, then? (Seriously, they're all over the place. :/)

 

[skip.knox]

I'm so excited. Now I have a guideline for making my book -- pardon me, "writings" -- just like other writings. I'm eager to get to work on my writings so as to incorporate discriminative unigrams, clausal tags, and constituent tags. Especially constituent tags, uh huh. I was also fascinated, as I'm sure everyone was, to learn that academic writing differs from fiction writing. I have always suspected this but now I learn it's true on account of all them charts. Numbers are truth.

What's really needed, folks, is an algorithm that can predict if my writings is gonna get picked up by an agent. Stylometric *that*!

Thanks to @Devor for the reference!

 

[Mindfire]

Hmmmm. Apparently they only used books that have fallen into the public domain. In other words, old and outdated books. How well does this algorithm stack up against modern tastes? Maybe they should refine the algorithm using data from more recently published books.

 

[T.Allen.Smith]

Hmmmm. Apparently they only used books that have fallen into the public domain. In other words, old and outdated books. How well does this algorithm stack up against modern tastes? Maybe they should refine the algorithm using data from more recently published books.

I can't believe someone is actually taking this seriously....

 

[Svrtnsse]

Hmmmm. Apparently they only used books that have fallen into the public domain. In other words, old and outdated books. How well does this algorithm stack up against modern tastes? Maybe they should refine the algorithm using data from more recently published books.

Yes, probably.

There's a note a the end of the article about how unsuccessful books on amazon also were rated as likely to be unsuccessful by the algorithm. However, I would say it's safe to assume that what's considered good/successful writing changes over time (right?).
Would this mean that the definition of "good" changes while the definition of "not good" stays the same?

 

[Svrtnsse]

I can't believe someone is actually taking this seriously....

There's serious and then there's serious.
It's interesting as an experiment. If it's possible to estimate with such high accuracy whether a book was successful or not by simply analysing the words, it probably tells us something.

I think that what I'm taking away from it is that there really is some kind of point to all these "rules" about writing that are being bounced around, both here and elsewhere. There's a point to making sure your writing it easily readable, it's not just grammar snobbery.

 

[T.Allen.Smith]

If it's possible to estimate with such high accuracy whether a book was successful or not by simply analysing the words, it probably tells us something.

Seems to me that having readers enjoy your work is a much better gauge. To me, these programs that try to boil success down to some measurement are beyond ludicrous.

If art success could be reduced to an algorithm, we'd have THE blueprint for success. From there, it'd just be a paint by numbers. What a fabulous museum that would be...

I think that what I'm taking away from it is that there really is some kind of point to all these "rules" about writing that are being bounced around, both here and elsewhere. There's a point to making sure your writing it easily readable, it's not just grammar snobbery.

The point to rules is to use techniques and methods that compliment the style you're trying to achieve. Being easily readable is only one facet of an overwhelming amount of variables. Seems to me this is why people force computations onto art. Outside of amusement alone, I see no practical value.

 

[psychotick]

Hi,

There's only one algorhythym that matters to our selling, and that's Amazon's sales algorhythym which determines how our books are placed on advertising lists, ranked and seen. And it doesn't care about our adverb usage.

Cheers, Greg.

 

[Svrtnsse]

Seems to me that having readers enjoy your work is a much better gauge. To me, these programs that try to boil success down to some measurement are beyond ludicrous.

Yes, clearly. I'm not saying it tells us everything. I'm saying it tells us something.

If art success could be reduced to an algorithm, we'd have THE blueprint for success. From there, it'd just be a paint by numbers. What a fabulous museum that would be...

The point to rules is to use techniques and methods that compliment the style you're trying to achieve. Being easily readable is only one facet of an overwhelming amount of variables. Seems to me this is why people force computations onto art. Outside of amusement alone, I see no practical value.

I didn't mean to give the impression that I thought sticking to the rules was a guaranteed recipe for success. However, just faffing about without knowing what you're doing is unlikely to get you anywhere at all - unless you're extremely talented and/or lucky.
There are always exceptions to all things.

I don't think this algorithm has the recipe for success. It may have some interesting points when it comes to measurable language statistics for successful vs unsuccessful books, but other than that, it's more amusing than anything else.

The question asked though, if the results would change it the system was trained on modern literature instead of on classics from Project Gutenberg, is still interesting (to me). I'm a bit of a maths geek and I find this kind of thing interesting.

 

[T.Allen.Smith]

Yes, clearly. I'm not saying it tells us everything. I'm saying it tells us something....

I didn't mean to give the impression that I thought sticking to the rules was a guaranteed recipe for success. However, just faffing about without knowing what you're doing is unlikely to get you anywhere at all - unless you're extremely talented and/or lucky. There are always exceptions to all things. I don't think this algorithm has the recipe for success. It may have some interesting points when it comes to measurable language statistics for successful vs unsuccessful books, but other than that, it's more amusing than anything else.

I'm just poking fun at the algorithm idea Svrtnsse. I'm not really trying to argue for or against.

 

[Svrtnsse]

I'm just poking fun at the algorithm idea Svrtnsse. I'm not really trying to argue for or against.

Fair enough. I thought you thought I seriously thought... etc...

 

[Devor]

The reactions have been fun to read. It did not disappoint.

 

[danr62]

Great, now I can hire someone to write a program that will create bestellers on autopilot.