These days I am reading the book Thinking Fast and Slow by Daniel Kahneman. It is definitely not an easy lecture, for a non-english speaker. You need time and dedication to fully understand it. I am sure I will need more lectures.
In the edition I have, on chapter 15 called “Linda: Less is More” there is a paragraph where the author reports an experiment run at the University of Chicago. Professor Christopher Hsee asked people to name a price for sets of dinnerware, using a method called “joint evaluation”. The method consists of compare two different sets of things and propose a value. The Set B had a list of items, all in good conditions. Set A, instead, contained the same items of Set B plus more items partly broken.
Since Set A contained exactly the same things that B, but with more things (partly broken), logically participants valued Set A more than B: $32 versus $30 (average).
The professor run the same experiment but with single evaluation. The result is quite interesting: Set B was priced way more than A. $33 versus $23 average.
What does that mean to us game designers? Sometimes the best you can do for a game is to remove some part that is broken. Having a feature that does not work can be a problem. It is better to remove features, and test if the things improve without them. Players can genuinely value our game better without that synchronized broken multiplayer mode, believe me!