I review a lot of books (100 last year, according to Goodreads; although that includes some I abandoned and excludes some I beta-read, it makes a nice round number, so let's use it).
Here are the figures by the number of stars I awarded:
5 star: 9
4 star: 54
3 star: 25
2 star: 8
1 star: 0
0 star: 4
The 0-star reviews were ones I decided against by the end of the sample, but had already put on Goodreads as books I was reading. Three of them I received when I was a reviewer for the Kindle Book Review, and one from Netgalley. I didn't feel I could give them a star rating because I hadn't read enough for a proper review, though I'd read enough to know that I didn't want to keep reading.
So more than half of the books I reviewed got four stars. Why is this? Is it a problem? If so, what can I do about it?
First of all, I'm getting pretty good at filtering out the bad books. I begin with the cover, though I really only glance at it, because I've seen good books in bad covers (and vice versa). Then I read the blurb. If the blurb has a typo in it, or is poorly expressed or confusing, or sounds like yet another rehash of tired tropes, or just doesn't sound like a book I'd be into, I don't bother to download the sample.
I check the reviews, particularly the ones with low numbers of stars. If anyone complains about a lot of typos or other language errors (which I find more distracting than most people), I rarely if ever download the sample.
Much of my filtering is by the sample. I sample any book that sounds like I might enjoy it (I have well over 100 samples on my Kindle right now), and if the sample doesn't grab my attention, or is full of typos and homonym errors or excessively dark for my taste, I delete it and move on. If it's borderline, or if it sounds interesting but I think it's a bit overpriced, I put it into a "possibles" collection on my Kindle. I have to say that I haven't, so far, gone on to buy any of the "possibles", because there are always plenty of other books.
No doubt this means that I've missed out on some books that improve dramatically after the (approximately) first 10% that's in the sample. That's unfortunate, but I have no way of telling those books from the ones that don't improve, which are, in the nature of things, more common.
It also means that I do read some books that get worse after the first 10%, which is where the three-star and two-star reviews mostly come from. Some of those low-star reviews were books I committed to read when I was reviewing for the Kindle Book Review, though, and I didn't feel I could just abandon them because I wasn't enjoying them. That no longer applies, so if anything, my proportion of 4-star reviews is likely to grow.
My Rating System
Here's the rating system I use. I've seen a lot of other reviewers describe their rating system, and it's pretty consistently similar, though everyone uses their own words to express it.
1 star: Awful. Dire. Pretty much all bad, with no redeeming features.
2 star: Bad, but with at least a hint that the author is capable of better; or very, very much not to my taste (despite being reasonably well done for what it is). I usually don't persevere with the second type these days.
3 star: The good and bad balance out; either there's a lot of both or not much of either.
4 star: I liked it, it was definitely more good than bad, but it wasn't without flaws (or, even if it had no significant flaws, it wasn't amazing enough for 5 stars).
5 star: One of the best books I've read, I will recommend this to all my friends and think about it long after I've finished reading it and buy whatever the author writes next and hug it and squeeze it and call it George.
So, between the filtering and the definitions and my increasingly-developed critical eye, most books end up in the 4-star bucket.
The trouble with most books being 4-star is that 4 stars starts to have such a wide range of meaning that its significance is diluted. There are 4-star books that are barely 4-star, that are only a little more good than bad, that have significant flaws which weren't quite enough to prevent me from mostly enjoying them; and there are 4-star books that are almost 5-star, that only miss out on 5-star because of one or two issues, or because, while they are really competently done, they're not amazing.
In the past I've tried having sub-ratings for language, plot, characters and setting, rating each of those out of five stars and then giving a combined rating. In most cases, though, the sub-ratings end up pretty close to the overall rating, and they're often all four stars.
I've tried starting with 100 points, knocking off points for each issue, and then converting to stars at the end by dividing by 20. It gives much the same result, though.
I think what I need to do is create a scale within the 4-star space to indicate where in that relatively wide space a book falls. A 10-point scale would give me the most flexibility. I'll probably need to explain, each time I use a value above 5, that the scale doesn't round, that 4.5 means that it's in the middle of the 4-star range and 4.9 means it's almost, but not quite, 5-star material. Maybe I should use some punctuation other than a decimal point to make that clearer. A dash can be read as "to" (4-5), a slash as "out of" (4/5), and a comma just makes it look as if I'm German (4,5). A tilde looks like "approximately" (4~5) and a caret as if I'm raising it to a power (4^5). How about "4#5"?
No, I'm going to have to explain it each time anyway. I might as well just say, "On my 10-point subscale within the 4-star space, ranging from 0 (just above mediocre) to 9 (just short of amazing), this book rates a..."