January 2020

[axn]

Hmmm.... Looks like they're trying to obfuscate some patterns by introducing some redundancy in the example. If you've tested a barrel individually against a plant on a given day, you don't need to test it again ever on any later days.

Since B & C were individually tested on day 1, they can be removed from consideration during day 2, giving us a simpler

1 	B 	DG
2 	DF 	A
3 	AH 	F
4 	C 	GH

Does this reveal any patterns?

[henryzz]

Originally Posted by axn

Hmmm.... Looks like they're trying to obfuscate some patterns by introducing some redundancy in the example. If you've tested a barrel individually against a plant on a given day, you don't need to test it again ever on any later days.

Since B & C were individually tested on day 1, they can be removed from consideration during day 2, giving us a simpler

1 	B 	DG
2 	DF 	A
3 	AH 	F
4 	C 	GH

Does this reveal any patterns?

Doesn't that logic lead to something like:

1 	A 	E
2 	B 	F
3 	C 	G
4 	D 	H

revealing the poisonous barrels.

Also if you know that only 2 of the barrels are poisonous couldn't you do:

1 	A
2 	B
3 	C
4 	D
5 	E
6	F
7 	G
8 	H

You know that only 2 would be sacrificed.
I assume I am missing something.

[axn]

Originally Posted by henryzz

Doesn't that logic lead to something like:

1 	A 	E
2 	B 	F
3 	C 	G
4 	D 	H

revealing the poisonous barrels.

No. *If* a valid solution exist with a single barrel test on day n and same barrel being tested on n+k day, then deleting all _future_ occurrences of the barrel will still keep it as a valid solution. I don't think you can extend this backwards in general (EDIT:- because the plant responsible for that test might have perished in the previous test, and hence the test didn't happen at all -- so I guess only day 1 is safe for this optimization).

Originally Posted by henryzz

Also if you know that only 2 of the barrels are poisonous couldn't you do:

1 	A
2 	B
3 	C
4 	D
5 	E
6	F
7 	G
8 	H

You know that only 2 would be sacrificed.
I assume I am missing something.

Haha! I think you just pwned the question (not the puzzle, per se, but the backstory).

EDIT:- We can retcon the puzzle to say that these are the 4 plants that are expendable and the rest all are super rare one-of-a-kind ones.

[henryzz]

Originally Posted by axn

No. *If* a valid solution exist with a single barrel test on day n and same barrel being tested on n+k day, then deleting all _future_ occurrences of the barrel will still keep it as a valid solution. I don't think you can extend this backwards in general (EDIT:- because the plant responsible for that test might have perished in the previous test, and hence the test didn't happen at all -- so I guess only day 1 is safe for this optimization).



Haha! I think you just pwned the question (not the puzzle, per se, but the backstory).

EDIT:- We can retcon the puzzle to say that these are the 4 plants that are expendable and the rest all are super rare one-of-a-kind ones.

I think we can assume that we can only use 4 plants full stop as you suggested.

I believe that my first set of logic still works just about in that case but isn't extendable to 3 days, more than 2 poisoned barrels or an unknown number of poisoned barrels.
If 0 die on the first day we are fine. If 2 die on the first day we are fine. If 1 dies on the first day then we won't test one barrel but we only need to know the status of 7/8 barrels as we know the distribution of statuses.

[Dr Sardonicus]

The backstory is fraught with mystery.

How is it known that there are two bad barrels?

If it is known that any of the barrels are bad, why not just replace the whole lot of them?

[Nick]

Originally Posted by Dr Sardonicus

The backstory is fraught with mystery.

How is it known that there are two bad barrels?

If it is known that any of the barrels are bad, why not just replace the whole lot of them?

Shhhh!

[what]

Originally Posted by Dr Sardonicus

If it is known that any of the barrels are bad, why not just replace the whole lot of them?

Shipment will take 4 days

[axn]

Originally Posted by Dr Sardonicus

How is it known that there are two bad barrels?

Many Bothans died to bring us this information.

.

[Dr Sardonicus]

Originally Posted by what

Shipment will take 4 days

Originally Posted by axn

Of course, the whole premise of the backstory is nonsense. If you want an orchid to die within a day, either put it in the freezer or water it with a flamethrower.

The American Orchid Society page on Basic Orchid Culture indicates that orchids need to be thoroughly watered when they begin to dry out (maybe around once a month), but that overwatering will kill them. They also grow better with the judicious use of fertilizer.

As to "special fluid," the American Orchid Society also has a page on Rainwater Collection.

[Xyzzy]

http://www.research.ibm.com/haifa/po...nuary2020.html

[LaurV]

Originally Posted by axn

%56 = 1

Do you mean you had to do 56 tests till you got it without error?