Assumptions:
1. Rats can and will drink a lot
2. This particular poison is also lethal to rats
3. Rats can be used more than once if needed and still alive
4. The poison is very strong and will not dilute when mixed with non-poisonous liquids
Phase 1: Picking the Best Half of Bottles
Pick 500 bottles to open and pour a tiny amount of each into a small vessel.
Allow the rat to drink from this and wait for it to possibly die (sorry rats :( ).
Phase 2: Narrowing down the Bottles
If the rat died:
-One of the 500 bottles is poisoned so the other 500 can chill
-Take 250 from the original 500 and mix it into a different small vessel
-Allow a different rat to drink from this and wait for it to possibly die
If the rat did not die:
-The original 500 are safe for human consumption and one of the other is poisoned
-Open 250 of the unopened bottles, mix and let another rat taste test
-While waiting for rat to possibly die, allow servants to begin decantering the wine so that it will be ready for when the guests arrive (serve snacks that are thirsty quenching so that guests will not be too thirsty hence buying some time just in case, this wine is the better of the wines so guests can savor it while still sober).
Phase 3: Repeat
Keep this halving process until left with a choice of 2 bottles and thus can determine which is poisoned as shown below.
Phase 4: Safely discard the poison and enjoy the party
Method:
After discussing with others in the class, it was determined that mixing wines was allowed and possibly necessary to figuring out this problem. Being able to mix wines helps to figure out which bottles are safe since you can eliminate many bottles at once. Hopefully the rats metabolize the poison quickly so that it does not take extremely long to test. Testing should not take too long if you can avoid the poison until the end. This process will become problematic if you choose the poisoned bottle in every mixture, which will also mean all the rats will die. If you are lucky and avoid the poison until the end, only 1 rat has to die. Being able to combine quantities and then ruling out entire groups systematically rather than testing one by one is a much more efficient way of sorting and determining something. In this case, it also saves the lives of at least 9990 rats.