How about the quality of a puzzles collection?

There are a lot of puzzles collections available on OGS.

So if you want to improve your reading ability, that is the place to be.
The quality of the various collections varies from excellent to absolutely worthless.

In searching for puzzles for my puzzles thread

I discovered a lot of worthless collections: without solutions, or incorrect solutions, no captions or nonsensical captions, etc.

How to determine the quality of a puzzles collection?
Well, the puzzle section supplies you with some data that can help you in choosing good quality selections.

The columns Difficulty (of puzzle as perceived by the author) and Puzzles (number of puzzles in collection) don’t help you much in this context.

But the other columns Rating, Views, Solved and Created do provide you with hints about the collection’s quality.

Rating: vistors can award a collection with 1 to 5 stars (not good to excellent or easy to complicated). If only a few visitors awarded stars, that doesn’t mean much.
If you see a collection with :star::star::star::star::star: (1) that usually means that the author award his own collection with the maximum score.

Views: the more views a collection has, the better it is.

Solved: if a collection has a low score compared to the number of views, the problems are difficult. And vice versa. Most collections have a conversion score between 40-60%.

Created: the date of creation also reveals something. An old collection with relatively few views might suggest that it is not a good collection.

Combining these variables will usually provide you with an accurate quality score.
Looking into the collection is of course the best way of ascertaining the quality of a puzzles collection.

Have fun solving puzzles.


Sounds like some culling of puzzle collections would be good… Is such a thing possible? (Also groups but that’s another story)

1 Like

There are some highly rated puzzle collections (the highest on the site, in fact) that fail the criteria you give, e.g. no solutions, and IME large jumps in difficulty. I don’t expect to see a 15k, or worse, SDK problem in a series labelled “for beginners”. I don’t know how they got such high ratings, but as a very new beginner, they don’t serve me at all.

True, maybe more often than not the rating of an individual puzzle is questionable.

Not at all.
Popularity is no proof of quality for example.
The title of the puzzle and the author is the only “insurance” of quality if any.