5 ms·
Not sure how accurate this, alternatively it might need different assumptions - What about: NYTimes, WSJ, GitHub, BBC, ArsTechnica, Medium etc?
by SeanDav 11y ago
Not sure how accurate this, alternatively it might need different assumptions - What about: NYTimes, WSJ, GitHub, BBC, ArsTechnica, Medium etc?
- anton_tarasenko 11y agoThese websites have the low median score. That is, many submissions, many of them not relevant, so the median is low.
- minimaxir 11y ago"Not relevant" is not the same as "not upvoted." There are a number of reasons why a submission does not receive many upvotes which are unrelated to the quality of the content itself, which is why HN has repost rules. The 10 story minimum is to ensure a reasonable threshold for error and so a single submissions with 1000+ points (e.g. Show HNs) don't skew the results.
- anton_tarasenko 11y agoDo you mean that duplicates from popular sources (NY Times, WSJ, etc) spoil stats for these sources?
- minimaxir 11y agoNot duplicates, but more noise than signal.
- tankenmate 11y agoActually if you look at others' analyses[0] some of those sites have a high mean and/or median. On median score alone samaltman.com should be #1. The highest "mainstream" news source would be newyorker.com, it does have a low median, but the average is ~20 times greater. [0] https://docs.google.com/spreadsheets/d/1-TCo1mxiTkO4ZiXg5acUMProDzxbuWCS0sS9I7HMJAM/edit#gid=3476623 https://docs.google.com/spreadsheets/d/1-TCo1mxiTkO4ZiXg5acU...