It seems like once you favor one choice ("Star Wars" in movies, for example) a couple of times, it only lets you compare other movies to Star Wars.
I agree with latexr that for almost every use-case you want this to behave like a bracket that presents a rank of your choices at the end.
Before my boys were born I wrote something similar to help us decide baby names. I put all of my wife's and my choices in a head-to-head arena in a bunch of rounds. They all started out with a base 50 out of 100 score and the winner of each round got bumped up to the loser's cumulative score (if higher) plus some. The loser got bumped down to the winner's score (if lower) minus some.
The game would stop when all names had competed at least 3 times and scores stopped changing much.
I then took my wife's top 10 and my top 10 ran the game again on the smaller combined list. It worked really well, especially helping to resolve the "My favorite is X but yours is Y. You suck" logjam. We could see our favorites overlapped a lot which made it easier to agree on a mutually acceptable choice.
I agree with latexr that for almost every use-case you want this to behave like a bracket that presents a rank of your choices at the end.
Before my boys were born I wrote something similar to help us decide baby names. I put all of my wife's and my choices in a head-to-head arena in a bunch of rounds. They all started out with a base 50 out of 100 score and the winner of each round got bumped up to the loser's cumulative score (if higher) plus some. The loser got bumped down to the winner's score (if lower) minus some.
The game would stop when all names had competed at least 3 times and scores stopped changing much.
I then took my wife's top 10 and my top 10 ran the game again on the smaller combined list. It worked really well, especially helping to resolve the "My favorite is X but yours is Y. You suck" logjam. We could see our favorites overlapped a lot which made it easier to agree on a mutually acceptable choice.