An Almost Anonymous Blog

Grading media on a 20-80 scale

Yesterday Carlos reached out to me after reading my post about how I use my bullet journal. It's fun getting feed back a post from 9 months ago! Wild how people discover things. Anyway. I checked out his blog and immediately found this post:

A Brief Primer On The 20-80 Scale

For those unfamiliar, baseball grades players on a scale from 20 to 80, 80 being elite / top of the heap. I was aware of this rating (often hearing from the Productive Outs crew about "80-grade Dad Hat" in jest) but never understood it. Is 80 really good? Why do I never hear about 100?

Carlos explains it rather well, so please read his explanation - I would just be paraphrasing his own words. In his post he suggests that he might start rating everything that way from now on1. I'm here for it.

Similar to the 20-80 scale is the 1-7 scale, but there are fewer points. I like either method, even though they are essentially the same. I think I'm going to start using the 20-80 scale myself. But there comes a problem: most people are not familiar with these systems and are used to 5-star ratings for most media properties. How do you impart meaning in a more digestible manner?

I'm not a math wizard2 so I went to Lumo for help. I started with this prompt:

If I were to start grading things using either the 20-80 scale (like baseball) or 1-7 scale, is there any way to mathematically convert them to a 5-star rating system?

Thus started the solution and explanation of how it would be possible to that something I rate as 55 (slightly above average) might equate to 3 or even 3.5 stars3. At first I didn't quite understand what it was telling me to do, because it was giving me a formula that seemed to be in the format of a programming solution. I asked it for some suggestions based on different applications, and once I got that it seemed a little clearer. This is the original formula it gave me (linear mapping formula):

[ \text{Stars} = 1 + \frac{(X - 20)}{(80 - 20)} \times (5 - 1) ]

Um. OK. Looks confusing. Here's what it gave me for Excel:

=1 + ((A2 - 20) / 60) * 4
(This assumes the original rating is in cell A2)

Ah, now this makes more sense, especially after Lumo broke it down further:

  1. A2-20 → shifts the lower bound to 0
  2. /60 → normalizes the value to a 0-1 range (because 80-20 = 60).
  3. *4 → stretches it to the width of the star range (5-1=4).
  4. +1 → moves the minimum back up to 1 ★

And if I wanted to round to whole stars (or half stars) I'd add a rounding function to the formula. Apparently you need something called "Analysis Toolpak" for Excel to round to half-stars so that's probably not the route I'll go for conversion.

The 1-7 scale is a lot simpler of a formula, but I don't think that is going to influence which one I use - I mean it's only marginally simpler: =1 + ((A2-1) / 6) * 4

I also told Lumo that my eventual goal is to use this in Obsidian, with Bases, using custom formulas. Lumo isn't familiar with Bases yet so it gave me the formula based on DataviewJS. The advantage here is that I can do everything within Obsidian. ChatGPT understands the Bases documentation so I might ask it for a formula4.

I think it might look something like this - assuming I have a property rating:

1+(rating-20)*4)/6.round(0)

I would have to try it out to be sure. And of note, the examples I've seen suggest I could just use rating but all the formulas I've tried give me an error unless I use number(rating) (which outputs the property as a number).

As with everything I do "because I can", there is no practical purpose to this. Honestly, this is to give me personal granularity and convert that to a rating system most people are familiar with.

Edit 1 - had a wrong number in the Obsidian Bases formula.
Edit 2 - missed /60 in the Bases formula.
Edit 3 - realized that if I set the rating field to a number type instead of text I can leave out the number() function.5

Reply by email   Share this post  Mastodon  Bluesky

  1. Side note - he writes for Baseball America! Isn't that cool?

  2. I didn't get into radio for the math.

  3. Personally I prefer whole numbers for stars, but sometimes I like the granularity a half-star provides, most especially for movies. I try to stick to whole stars for books. Who knows why. But converting to stars also gives me a quick view of how much I rated a book, and since there are no half-star emojis available I might just use whole stars...though who knows, maybe there's a way to convert .5 to ½.

  4. The reason I used Lumo first is because it's the only AI that isn't Copilot that my employer hasn't blocked, and I wanted to type on a full keyboard vs. on my phone. I don't know if Copilot knows anything about Bases.

  5. ChatGPT helped me a bit with the formula and tweak it a little. It's one line, and messy with brackets, but here it is: "⭐" + ((((1 + ((rating-20)*4/60))*2).round(0)/2).toString().replace(".5","½")

#echo #obsidian #projects #reviews