Open Bartleby2718 opened 6 years ago
Actually, it's just logic, not a sudoku-solving algorithm, but anyway, let me give you a concrete example. If the rightmost decks, none of which is open, are:
Deck 8
contains at least one 11
, and Deck 9
contains at least one 13
besides its delegate
.
In fact, the possibilities are:
Case | Deck 8 | Deck 9 |
---|---|---|
1 | 12, 11, 11 | 13, 13, 12 |
2 | 12, 11, 11 | 13, 12, 13 |
3 | 12, 11, 12 | 13, 13, 11 |
4 | 12, 12, 11 | 13, 13, 11 |
5 | 12, 11, 12 | 13, 11, 13 |
6 | 12, 12, 11 | 13, 11, 13 |
For done
, both Deck 8
and Deck 9
must be opened, so maybe you should look for other numbers first. On the other hand, the big sums can be used to dare
some of the biggest decks the opponent has. Finally, there is a guarantee that joker
is in neither of Deck 8
and Deck 9
.
Some links on sudoku-solving algorithms:
The problem is that I have no idea how I should implement this... (because I haven't looked into it yet)
Another example:
In this case, Deck 1 has to be {2, 1, 1} and Deck 2 has to be {4, 4, 3} because it is the only possible case, although it may be evident at first glance.
What concerns me is that this is clearly a logical but may not worth the computing time.
If a value
x
must be in one ofm
decks, sayDeck 3
, and if there aren
numbers that can be inDeck 3
, what is the probability thatx
is inDeck 3
? It's complicated. Neither1/m
nor1/n
might be the answer. Since a value in acard
affects the values in the otherdeck
s, you actually have to count the number of possible cases. It's like solving a sudoku. This issue will add on to #94, although not anytime soon.