Ok, so there are some problems with this kata:

1. In the description you say that only "last cards" are counted if the amount of cards is odd — BUT how about the cases where there are more counters than ordinaries? Also, EACH card which don't have a pair should be counted, not LAST one.
2. In Python tests, there are strings like "{} should be {}" and such. Maybe something should be inserted into the brackets?

Maybe the second issue can be tolerated, but the first one makes the kata description misleading and contradicting with the Wikipedia data + tests, therefore, I do believe this kata should have not a misleading description (like, you need to make pairs of counter+ordinary and ordinary+ordinary and then add the scores of the cards which don't have any pairs or smth like that). =/

Something seems wrong with this Kata in Haskell. A 78 card hand should get a score of 55, but the Random tests are expecting it to have a large negative score.

Example:

-- >> tarotScore "SC15DQ0904D7D41703H3CQ0607H6051612HKH7H2SKHCHQ20D3C9CJS7D5DCS5D6C8H9S41014C7S0SQC4S2H419H001D1C3D0S3D9CKC1DJCCS821S90213C0C5S618S1D8H5H8HJ08C6DKSJ11H100C2D2"
-- (-99) % 2