Hi Paper Tiger, You are a grrrrrreat inspiration and a very good source of info! Thank you. As per our private email conversation I’m going to critique the math in this video even though I don’t like giving bad news, especially to someone who works as hard as you. Regarding how to choose books it’s not appropriate to use that stopping strategy in that way. It was meant for situation when you can’t go back and pick something you already passed over. In without getting all mathy, shouldn’t the person selecting a book pick one of the first books, if it is the best 2-page read? Why pick a lesser just because it came later. Also, it doesn’t guarantee a 37% chance of liking the book. Wiki explains stopping strategies, but the writing is so technical that most readers would get lost in a forest of math jargon. I have found all your other videos to have wonderfully useful information. Thank you.
Hey Jim, thanks for pulling me up on the maths! I am very much not a maths person so am always keen to learn more and happy to be corrected when I've understood something incorrectly. I 100% agree that the logical choice is to go back and pick whichever one you liked best based on the first 2 pages...which means your next read should always be the best one in the batch that you're sampling. And now that you've made me think through it, if we're slicing up the total population of all potential reads that fit our criteria into randomly selected batches, that really does change the probabilities quite a lot. I suppose it means 37% of the time you would end up selecting the best book in the randomly selected batch if you ignored the books you sampled at the beginning, but 100% of the time if you're allowed to go back and pick the best of the batch. But then because the batch itself is randomly selected, there is a chance that the books in there are all 1- or 2-star reads, or equally 4- or 5-star reads. Suppose I pick a 2-star book out of that low star batch-I can still choose to DNF beyond 2 pages since picking a book out of a batch doesn't mean I have to commit to finishing it, it just means I'm committing to reading beyond two pages. If I DNF, I would go back to another batch and repeat the process. Sooo...hmmmm. I think you're right. There's no guarantee that there's a 37% probability of selecting the book you'd like best in the entire population this way. Or is there? So does the 37% only hold true when we consider this over consecutive batches? Or does the scenario breaks far too many of constraints that it's no longer really an optimal stopping problem, as those are mathematically defined? Since in terms of practical implementation, I could just decide to not pick any of the 10 books in the batch if I hate all of them. Perhaps it wouldn't actually outperform the current system because you'd have to go through batches and batches and batches before you found anything that appealed to you enough to get you to read beyond two pages. Though, I don't know. I feel like that also makes certain assumptions about the natural distribution of books in terms of their quality along some scale, which would be different for everybody. But then again, if I'm sampling batches at a time, the books are ordered randomly, and no book gets pulled up twice once you've noped out of it, isn't this a question of user interface? It still wouldn't change the fact I'm considering books one at a time. Though I guess what's changed is, if I keep extending the number of batches I'm sifting through, then it becomes a question of whether my initial sampling was sufficiently large. And then yep okay, we're into sampling theory territory now, which I also understand is very technical, but I feel safe to say based on the way auditors do sampling, it wouldn't be unreasonable to rely on attribute sampling? Since really we're testing for a single attribute, i.e. "is this book my jam?", with the underlying assumption that most books are not my jam. In which case, my auditor brain goes, okay then, an initial sample size of somewhere between 25 to 60 would be appropriate. But then usually when you're using that method, you're trying to draw conclusions about the population based on the sample, not to find the optimal selections, which means this isn't right either! 🤣 My conclusion is, if I ever happen to end up actually trying to build this, I would hire someone who is good at maths to run some calculations. 😂 And then I supposed you'd have to run a bunch of trials to figure out if this method actually outperforms the current methods. But I think the randomness factor this method relies on means it goes a long way towards levelling the playing field! Book reviewers, paid traffic, etc those would all still be options and for authors and publishers who can afford that, they'll still be able to find readers that way. This would just give those who can't afford to compete on those levels a chance to hook those first readers...and help readers find books that would never have come across their radar.
@@PaperTigerProductions Additional thoughts. If we limited to two pages, and I think a reader would be able to pick up on the writing style of the author. I think this would be good for readers who are unfamiliar with an author. Procedurally, maybe the following would be a decent strategy for selecting a book given that historically only 1 in 8 books that the reader picks out are exceptional. (If 100% were exceptional, previously, keep doing the same selection process. So, yes, your a priori probabilities matter.) I suspect that the first two pages of a book are the worst indicator of whether it is a good book to read because so many authors polish the beginning of their book. So, pick out 16 books based on artwork, blurb, and other available data (‘best seller’ lists, ‘if you like that you might like this’ list, friends, and, heck, enemies). 2 might be good. Decide what drives you to read. If it is character interactions that are believable and dramatic, without being overly dramatic, find a passage 1/3 of the way into each book. Discard the weaker half-ish of the books. Repeat for passages halfway into each book. You’re down to about 4 books. Check them out of the library and read each till it becomes irksome. Why those locations in the book? The middle of the book is always hardest to write. If the author can keep your attention there, it’s probably a good book. And, reading there won’t spoil the ending. For a person, like me, who likes MacGyver solutions, even magical ones, the author, blurbs that mention creative climaxes, friends, and lists are the only way. Ditto for internal consistency. Thank you for bringing up a very interesting problem.
@@jimf2525 it really begs the question of how much of a book do you have to read to know whether or not it's not for you. Prose, character interactions, etc are easy to judge from a couple of pages, but it's harder to judge something structurally without having read the whole. I have definitely had the experience where I've really enjoyed the opening of a book, even the entirety of a book one in a series, and then gotten progressively more disillusioned with it until I was thoroughly disgusted by the end and absolutely loathed it. But I also know there have been readers who had the opposite reaction to the same book/series. Interesting things to ponder!
Book tinder is a great idea, and I'm surprised it doesn't already exist. Look at what Tiktok has done for books. I think you're onto a winner. Also, like you, I've been thinking about just how short life is versus how many 5-star reads for each of us are out there that we'll miss because instead we're reading books we don't like because of the size of our individual slush piles. I'd been thinking of starting a booktube channel where I judge a book on opening line alone, and if it hooks me, I go on to read the first paragraph, then page, then chapter. If I'm still hooked by then, I add it to my TBR. If at any point I'm no longer hooked, I drop it and move onto the next... because, again... life is too short.
Since the solution I propose doesn't exist yet (to my knowledge), I've got some fantasy-specific alternatives for you over in this community post → ru-vid.comUgkxlt5D7j_MaqNncp2zm6aGVp2IWFq5CE60 And if you're looking for some advice on how to succeed at self-publishing, here's some advice from New York Times best-selling self-published author, Will Wight → ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-0Xoqx_GUxoE.html
An update! Nikky Lee mentioned via the Indie Accords discord that there's a little NZ-based start-up called Narrative Muse that is trying to do something similar-you can check that out here: matchmaker.narrativemuse.co/
I feel like the Storygraph app does a great job with high-level filters, but since I only search and sort there, and don't pay attention to recs, I must be guilty of bias...but I also get satisfaction more than 37% of the time 😏🤣
I've got to give Storygraph a shot one of these days! I'm so bad, I've realized that I pretty much never **search** for books, I always either browse or go with recs. It's terrible!
@@PaperTigerProductions I can really highly recommend it! I especially love filtering my TBR, and I've been reading more books I already own because they were suggested by TSG after I set some filters.
@@bumblewyn Oh nice! I reckon I'll give it a shot once I get around to inventorying all of the books I have, digital and physical. 😅It's all a bit of a mess atm
Tinder for books sounds awesome! I think Narrative Muse has done something a little like that - have you heard of them? From memory, you spend a little time going through what you do and don't want, and then marking books as either read, don't want to read, definitely want to read, or might read (I think! it's been a little while).
Yes, someone else mentioned Narrative Muse to me in the Indie Accords discord, which I've put in one of the other comments and in the description! I hope more people give their app a shot.
Fascinating! I'd really like to see more discussion about discoverability. In the past few years, I've read a wider range of authors, with many of them being self-published. Some of that is due to people I've followed on social media. In regard to your mathematical method, I hope you develop it. It seems adjacent to some concepts of The StoryGraph, which is one of the places I track my reading. They've done a great job of breaking down habits with statistics. That's helped me shift my reading to more authors that might not have caught my eye.
Almost all of my newly found books came from the "other readers who liked this" recommendations on Amazon. But usually A) I've already read them , or B) (and this is speculation) those are promoted somehow, because it's really hit or miss. I would help with this project! Not good with database, but I am good with WordPress and websites, and, hell, just entering data.... Maybe get a Kickstarter going? This sounds really cool!!! 👍
Yeah I can't remember if David Gaughran has looked into some of those other recommendations on Amazon and how they're derived. I've read Amazon Decoded but I can't remember anything concrete about them, I don't think he delved into too much detail about those periphery algorithms because it mainly focused on also boughts, bestseller lists, and the popularity lists. Re: Book Tinder, I won't lie, my brain has been going overtime with thoughts on this 😅 but I have a sequel to write first, so we'll see how things go!
@@PaperTigerProductions Books first for sure! But brain in over drive never hurts. Let me know if you decide to work on it. I have some friends who are very good at Kickstarters, and would be willing to set that part up, or whatever would help. O:-)
