flesh out readme

README.md

@@ -1,6 +1,37 @@
 # Elo Anything!
 
-todo: readme text
+I love using [idea fight](https://idea-fight.hoelz.ro) to rank things.
+It's pretty nice!
+I can get an idea of the top item in a list in `O(n)` time, and finding all values is `O(n log n)` (see [Rob Hoelz's blog post about idea fight](https://hoelz.ro/blog/idea-fight).)
+I've used idea fight for years, but recently I've been running into some problems when using it:
+
+1. I second-guess my choices. I'd like them to sort out over time instead of having to commit once.
+2. Some things are pretty equal, but the selection method means that choices below the "losing" choice are automatically lowered when I'd kind of like them to be higher.
+3. I can't see what's going on in the process when it's running.
+4. There's no undo.
+
+I think most of these are caused by the fact that I often do not put a series of items with a strict objective ordering into the tool!
+So I could give up on the nice `O(n)` sorting time if I could get some of those benefits.
+
+This is an exploration of sorting things with an [Elo rating system](https://en.wikipedia.org/wiki/Elo_rating_system).
+Essentially, it solves my problems in the following ways:
+
+1. If I second-guess a choice, it's fine. The match may come up again, and this style of ranking is only a rough order--it's fine if it's not strictly ordered!
+2. The rating system models ties, so if two things really are equal they will become closer in rank over time.
+3. I can see the rankings as I'm making choices.
+4. There's still no undo, but I can add it if I find that it's really a problem! (So far I haven't.)
+
+## Building
+
+You don't have to build anything.
+Go play with a live version at [elo.bytes.zone](https://elo.bytes.zone).
+
+But if you want to hack locally:
+
+1. have direnv and nix
+2. clone and `cd` here 
+3. run `modd` to start the development server (it will tell you where to go to view the result)
+3. run `nix-build` to build a deployable version
 
 ## License