Very nice kata, really interesting overall.

this consistenly fails the random tests now

The reference solution implemented the tiebreak rule ("If there is a draw, players with the same score must be presented in alphabetic order") incorrectly. The upgrade to Node 18 exposed it, as the implementation of Array.prototype.sort() changed (it is stable from Node 12 on), which made the refsol fail the fixed test "whoWon when there's a draw". I tentatively fixed it by replacing:

return a[0] > b[0];

with:

return a[0].localeCompare(b[0]);

It is very possible that something else is broken, as there were only 6 random tests, so buggy solutions could pass the tests.

Currently, the number of random tests is limited to 6, this is clearly not enough.

Nice One!!!

A lovely kata. It was fun. Is it matter where 'totalLoosers' appears in the winnersList, I mean 'Ben: 23' should go first or 'Jack: 27'. Ideally Ben in this case should go before Jack since its score closer to 21. But at this moment either implementation is allowed ([Jack, Ben] will pass tests).

Approved

Thanks for pointing that out. I've fixed these issues now

Issue: Occasionally with the random tests, a player's card will be given as 'undefined'.

Here's an actual test case that occurred:
{ Ben: '7, 6', Amy: '7, undefined', Sam: '5, 2' } 'Q' [ 'Sam' ]

Suggestion: (unrelated to the issue above) In the Description, for the first rule:

The person whose name is in the second argument must have the value of the extra card added to their score, unless they have a score of 21.

The last bit is unnecessary since the maximum hand a player can start with is 20 (since Aces cound as 10).

Thanks for the feedback. I think I have resolved these issues now.

• Initial code says whoWins
• Random tests are not generated individually