Hacker News new | ask | show | jobs
by usui 687 days ago
You're correct on this. Tabelog users center hard around 3 stars which is considered the real average. Users on sites like Yelp and Google Maps skew toward an average of 4.5 or 5 for what is considered a standard experience. This makes it extremely difficult to sort for places that are extraordinary because everyone keeps rating the max amount of stars by default. Same problem on Amazon. It doesn't really help to only filter for 1-3 star reviews for each place and read them all because my time is finite and I don't have all day.
2 comments

I would disagree that the issue in Yelp and Google is that folks skewed to the right in their ratings - my experience has been that if a restaurant has 4.5+, it’s a safe bet. There’s literally no more information you can glean from google. Perhaps the number of votes but even that is not absolute. You have to depend on other sources for any more indication that it’s a good place. My go to source is Reddit. If I go to that city’s subreddit and I can’t find that restaurant mentioned it’s likely not mind blowing.
My point is we're different types of users, or our expectations are different. "A safe bet" is not 4.5 stars to me, it's 3 stars because average restaurants should be safe bets. I believe anything 4.5+ should be home runs because I'm putting faith in the score of the restaurant actually having more meaningful information embedded in it. Unfortunately this is not the case because American sites think anything 3.5 and below is automatically shit or something will go horribly wrong. I keep going to 4.4+ restaurants on Google Maps which are actually ~3.2 on Tabelog, but when I go to a 4.2+ on Tabelog I know for sure this is a top-tier restaurant and this has never failed.
For google maps I honestly have better experiences with 4 out of 5s than ones closer to 5 out of 5. No idea why
4/5 might mean they're being more adventurous with their food. This won't necessarily suit everybody.