Collaborative research by scientists in two countries, two hemispheres, and two daylight saving time policies, has created a massive breakthrough!
We open the scene with some more thrilling scenes of heap-sorting chocolates. Ramping up the statistical analysis with not two, not three, but FOUR parallel tests, the Australian researchers have discovered a remarkable stability of data points. The four boxes (rated 450g each) each contain the exact same distribution of chocolate types - specifically:
* 4 Dairy Milk
* 2 Hazelnut Praline Crisp
* 4 Strawberry Cream
* 4 Classic Fudge
* 3 Chocolate Supreme
* 4 Caramello Deluxe
* 3 Orange Chocolate Delight
* 3 Hazelnut Whirl
* 4 Turkish Delight
* 4 Vanilla Butter Caramel
* 4 Peppermint Cream
* 3 Cherry Heaven
Just before intermission, a thrilling theory is posited. The distinction between stable and unstable data is a geographic one! All of the data from Ireland, so far, has been somewhat randomized, while all of the data from Australia has been stable. Also, the set of keys is geographically aligned - all the Australian boxes, and all the Irish boxes, match. But the final words, before the screen goes dark, point out that one of the Australian data sets had a key set matching the Irish pattern...
The second half opens, naturally, with a recap of the preceding information. Once that's finished wasting a good quarter of the available time, the answer to the preceding puzzle is found. The one Australian package with distribution matching the Irish ones has proven to be the production of Cadbury UK, as evidenced by the under-tin information. It is also an 850g tin, as opposed to being a 450g or 1000g box. This tin would appear to be... a foreign import.
An opportunity will arise in a few months for some members of the Australian research team to take their research to England. If, as it appears, Cadbury UK differs from Cadbury AU/NZ, then on-the-spot research will be vital in finally answering this question.
The four 850g tins of Cadbury UK roses (one in Australia, three in Ireland) carry unique data signatures. This suggests either a lack of stability, or a deliberate snowflake policy. To determine which is the more likely, far, FAR more research will be needed. Can you ever find two 850g tins exactly alike?
The producers wish to thank Peter Flynn for his material assistance.