The new testing framework should be used

A ton of missing specs:

• What is "invalid rating"? The scope of this is not defined by the kata
• Are ratings like 6.9/10 or 1.23/10 valid? This is not defined or tested either
• Empty lists should be treated as "0/10": The actual condition isn't "empty list", but "no valid ratings"
• It is not mentioned that the rating should be rounded to 1 decimal place

Random tests have a very high chance to contain no valid ratings.