No truth in show ratings? Yes, there is error, but the rest of your claim is absolutely 100% completely and utterly wrong.

I've gathered show ratings for all 524 shows, between 2009 and 2022, and used regression analysis to link the musical elements of a show to its ratings. Guess what?

1. As the number of ">" segues increases (indicating pace of a show) the rating goes up
2. As the number of "->" segues increases (indicating flow) the rating goes up
3. As the average length of songs in a show increases (more jamming) the rating goes up
4. As the average song gap increases (a measure of song rarity) the rating goes up
5. As the number of narration songs increases, the rating goes up

You suggest that ratings are random and arbitrary, and this simply isn't so.

Rather, your primary complaint is that people do not use the full range of the scale, which is absolutely true. Fully 50% of the shows in my database are rated between the median value (4.025) and the maximum value (4.648), which is a narrow interval relative to the range of 2.261 to 4.025 for the bottom 50%. But just because people use only 3, 4, or 5 when rating doesn't mean the show ratings have no informational content.

Finally, show ratings are subject to at least two different selection biases and any number of other biases as well, including recency bias. My analysis of recency bias suggests that it's less than 5% of a show rating, and the tendency of people to listen to and rate only those shows that are already highly rated is about 3%.

And anyone who wants to see a paper that supports my claims, just send me a PM.