Estimations on where I fall on a variety of measures
Traits
Wealth: $1M = 1/100 in the US? 1/10000 in the world?
Height: 5’10” = 1/4 in the US?
IQ: 150 = 1/1000
SAT: 2320 = 1/1000
Math: USA(J)MO qualifier = 1/10000 (top 500 of 5M kids. 4 years = 20M, but not everyone participates)
Games
Mythic in MTG Arena
Subjective
Programming: 1/100
Writing: 1/4
Community building: 1/5
Marketing: 1/3
Sales: 1/3
Product: 1/10
VisD: 1/4
Reading speed
Typing speed
Notes
- Ratios work better than percentiles for expressing
- Peak measure vs current measure
- Peaks make sense in most competitions? But also may mislead where your current skill is at
Different rating systems
- 5-star system (Yelp, Amazon, Uber)
- Rating: 2 stars
- Pros: Simple, widespread
- Cons: Not very granular? Majority of ratings get put in the 4-star range
- But maybe that’s the positive effect of star systems
- Sometimes equivalent to an X/10 rating if you allow half-stars
- X/10 (looks, MyAnimeList, Net Promoter Score)
- Rating: 4.5/10
- Percentiles
- Rating: Top 80%
- Pros: Most expressive of calibration
- Cons: Requires good understanding of the population you’re comparing to.
- E.g. relative to peers? practitioners? Americans? The world?
- Top lists (buzzfeed, top 3 books of all time, best Chinese food in the area)
- Rating: Top 5 rating systems of all time
- Cons: very dependent on personal exposure (maybe you just never went to one particular good place) making it hard to compare lists
- Percent up to down vote (Rotten Tomatoes)
- Rating: Up
- Pros: Very simple
- Cons: inability to express degree of belief. Requires a large-ish audience to be meaningful.
- Letter grades (SSS, SS, A...)
- Rating: B-
- Pros: We all had to do this in school
- Cons: We all had to do this in school
- Likert scale (hiring, psychology tests)
- Rating: Like
- (-3 to +3, or Strongly Dislike to dislike to slightly dislike to neutral...)
Common problems with rating systems:
Not knowing what the median rating should be (should the average rating given be 2.5 stars out of 5?)
Not knowing what the variance should be
Not knowing what the skew should be (lots more 5-stars than 1-stars given out)
Equivalence between rating systems and voting systems?
- Democracy, plutocracy, quadratic voting