I didn't find any other bug.
Thanks again.

Now that seems to be fixed, thanks.
I'm still reviewing other test cases.

deleted

I'm not passing the kata because of this issue. Apparently I cannot contribute further before passing it. And I'm not going to code a wrong solution in order to pass the wrong tests.

The same is happening to me in the C version. Not passing the tests because of that.

In the C version, random tests produce invalid inputs like "3C KH 5D 5S KH".
The King of Hearts is twice in the hand, so that hand is invalid.

I totally agree with Matteo. The only reason why I wasted my time with this problem is to be able to properly complain about such a loss of time.
It is not about English not being your mother tongue though. English is not my mother tongue neither and I can give complete and non contradictory instructions.
The instructions doesn't specify why the value of the old car matters (you have to understand from the examples you can sell it in order to buy the new one). The explanation about the float rounding is contradictory. It tells you to truncate and to aproximate it to the nearest int in the same phrase. It is a disaster.

I honestly and respectfully think it is hard to explain the problem that bad.
Don't waste your time with this kata, I did so I can advise you and complain about it.

But the mother sees the ball once.
The instructions are at least ambiguous.
I'm downvoting this.

I totally agree with this comment. The test is testing a number out of the scope of the function.
It does not matter how many people has passed the tests. The tests are wrong.
You should change either the description of the funcion or the tests.