[HN Gopher] The Shape of Information
       ___________________________________________________________________
        
       The Shape of Information
        
       Author : effect
       Score  : 52 points
       Date   : 2024-06-18 23:05 UTC (1 days ago)
        
 (HTM) web link (kucharski.substack.com)
 (TXT) w3m dump (kucharski.substack.com)
        
       | yu3zhou4 wrote:
       | We can also open first 50 bottles, mix drops of each, use test,
       | if positive then we know that one of them is poisoned, if not
       | then we know that one of the other 50 is poisoned. So we take the
       | poisoned 50, split on two brackets and repeat. With 7 tests, we
       | can narrow down to a single poisoned bottle (50, 25, 12/13, 6/7,
       | 3/4, 2, 1)
        
         | The28thDuck wrote:
         | Binary search strikes again!
        
           | yu3zhou4 wrote:
           | Nothing clever, I know :p
        
             | selimthegrim wrote:
             | Maybe so, but pooled testing as they refer to in the
             | article involves using Lambert W function to solve for
             | optimal size of individuals tested.
        
         | photochemsyn wrote:
         | The limit of detection of the test comes into play; detecting
         | poison at a 1:50 dilution might be impossible with that test.
        
           | TeMPOraL wrote:
           | But then, if that test _can_ detect poison at 1:50 dilution,
           | it means it probably has enough reagent to detect poison at
           | 1:25 _twice_ , if you're clever about it. Done well, you may
           | end up with a single bottle identified and a bunch of unused
           | tests to sell.
        
         | ubittibu wrote:
         | I hate when riddles are told badly and ruined. I remember the
         | original one which was not solvable this way and really puzzled
         | me. When resolving the one in the article I asked myself why I
         | had found this to be much more complicated, and it was beacuse
         | the number of bottles was 128 in the original.
        
           | Chinjut wrote:
           | What do you mean? It works the exact same way for 128
           | bottles. 64, 32, 16, 8, 4, 2, 1.
        
         | Chinjut wrote:
         | This is essentially the same approach as the solution described
         | in the article. You are observing the bits of the poisoned
         | bottle number in binary, bit by bit. The only difference is
         | that you avoid retesting any bottles once they have been ruled
         | out by previous tests, whereas the fixed design in the article
         | does not bother with this.
        
       | ReleaseCandidat wrote:
       | The answer is easy: by his way of testing exactly one bottle
       | won't be opened and of course that's the one that contains the
       | poison. Drinking 99 bottles of wine before they go stale isn't
       | too healthy either, I'd guess.
        
         | mirekrusin wrote:
         | Exactly better to keep 7 with scotch taped test and sell the
         | rest on amazon, call it a day.
        
       | photochemsyn wrote:
       | When faced with incomplete data, it's really always better to go
       | back and get a more complete and reliable dataset than to apply a
       | host of statistical tools to that poor data in the hopes of
       | extracting some trend or meaning from it.
       | 
       | > "Fortunately, you have a test that can detect poison very
       | accurately, even among a large volume of liquid. Unfortunately,
       | you have only 7 of these tests available. What should you do?"
       | 
       | You should estimate the cost of getting another 93 tests,
       | consider if a cheaper method is plausible and could be developed
       | at lower cost, then compare that to the value of the wine
       | bottles.
       | 
       | With infectious disease, this is problematic as putting a cost on
       | every individual human life is tricky, even if that's the life
       | insurance business model. As a practical example, the cost of
       | running the CDC's malaria detection program (captures some
       | 2,000-3,000 cases a year from incoming travellers) in the USA is
       | considered worth the benefit of not allowing malaria to become
       | endemic again across the Southeastern US.
        
       | Nevermark wrote:
       | The puzzle fails spectacularly.
       | 
       | Without additional action all that wine will go bad.
       | 
       | Important detail is one must recork each bottle immediately after
       | sampling & refrigerate. Now the wine will be preserved, and the
       | very slight pre-drinking oxygen bump is likely to improve it over
       | first corking.
       | 
       | A little overlooked & underrated trick.
       | 
       | (Remember to decant reds and allow them to reach room temp before
       | drinking.)
       | 
       | Also left out is value/opportunity maximization, which is a
       | crucial implicit factor in any gamed, especially adversarial,
       | situation.
       | 
       | Math & algorithm problems without complete context value
       | optimization are just ivory tower toys!
       | 
       | The real world, Machiavelli, your ancestors, and your
       | shareholders demand better!
       | 
       | The poisoned bottle, once identified, is worth a great deal.
       | Minimum value realization is obtained by rebottling, and re-
       | gifting it to its original source. Anonymously. Of course!
       | 
       | Or you can save it for a special occasion! Sometimes the best
       | move is no move. Accumulate optionality.
        
       | esafak wrote:
       | To paraphrase Crocodile Dundee; that's not the shape of
       | information, this is:
       | https://en.wikipedia.org/wiki/Information_geometry
        
       ___________________________________________________________________
       (page generated 2024-06-20 23:00 UTC)