Greatest Science Fiction Film Tournament: Final Results

This is it. We’ve winnowed a total of 392 films down to two, and of those two, we voted on the greatest science fiction film ever. The result: Star Wars (1977) wins by a single vote in the week with the largest voter turnout to date. We’ve got the complete statistics for the entire group, as well as a new poll to decide the format of our upcoming Greatest Science Fiction Television Poll here as well.

The round one results, based on Bayesian statistics, can be found here. Those are the numbers used to “seed” the tournament standings throughout, using IMDB scores to break ties in Bureau 42 reader votes. The voting results (as percentages) for other rounds should be on the GSFFT Brackets file. The brackets were set up for rounds 3-8 starting with round 8 and working backwards, so that (if all went as per round one voting) the top two films from the first round would face off in the last round, the second last round would see movies 1 and 4 in one bracket and 2 and 3 in the other, etc. so that a round with n titles would have pairings whose ranks added up to n+1 in each case. The exception was round 2, where the wildcards came into play. Instead of the top 128 movies from round 1, we had the top 120, with movie #1 going up against #120, movie #2 going up against #119, etc. except for the 8 middle slots (#57-64) which went up against the 8 wildcards such that #57 went up against the wildcard with the lowest IMDB score and #64 went up against the wildcard with the highest IMDB ranking. I was tempted to “rig” the system so that related movies (the two “King Kong” movies, the “TMNT” films, etc.) would end up in the same bracket against each other, but that was unnecessary, as that’s how things naturally fell anyway. Ties in rounds 2-8 were determined by the round 1 rankings of the movies in question. I’d have used IMDB ratings as tie breakers for the wildcard films, but it never came up.

We also have a few options for setting up the television tournament, which you can vote on here. The results won’t just be tabulated, but this time, they’ll also be correlated. (There are options to vote in multiple stages at the series, season and episode levels, as well as for tournament length, so the winner will be the most popular combination, i.e. “these two stages plus this length” style.) We’ll gather information on this poll for the next week, and then announce the final format at the usual day and time.

7 replies on “Greatest Science Fiction Film Tournament: Final Results”

  1. J_W_W says:

    This tournament was great! I’m looking forward to the TV one.

  2. Any chance of a copy of the bracket that shows the seeds, so I don’t have to keep toggling between two spreadsheets?

  3. Fez says:

    I am pleased with the overall winner, even if I wasn’t so happy with the results of previous rounds…

    The TV tournament sounds good. Drilling down to individual episodes seems insane though, seasons even seems excessive. Without detailed synopses or season summaries, there are probably many entries that would be passed over because people didn’t know them by name or remember what happened on season/series X of the show. Most shows I watch I couldn’t pick out episodes by name. Some of them I can, but others, not so much.

    • I can see that. There are two reasons I thought about of drilling down to episodes. First, it means we don’t need as many series to seed the tournament, since 384 good TV series are hard to track down. Secondly, it makes podcasting about each series much easier after the tournament. (Yes, we’ll be doing more movie podcasts, but not for a while. I need to get through the “Doctor Who” podcasts first. Those are about 80% complete.)

      • Fez says:

        Yeah I could see that I suppose at least by season. In my head at first it seemed as though it may give a long-running but mediocre series an advantage over a short-lived but great one. Maybe not an actual statistical advantage but at least a mental one. (Eleventy billion series of Dr. Who vs. one season of Firefly, for example)
        Plus then you’d have people who unfairly weight a show they like vs anything else, regardless of whether or not a particular season was any good.

        Going by episode has other advantages, such as allowing great episodes to shine in an otherwise crappy season, though the sheer number of episodes and the research required to identify and judge them properly just seems … weighty.

        Unless a show’s seasons/episodes only competed with other seasons/episodes of the same show, picking the best episode of each show, and then those best episodes would move on to compete against each other… So you end up with, say, the best Stargate SG-1 episode (Window of Opportunity, clearly) vs the best Fringe episode.

        And then what about TV series’ movies that were just for TV/DVD? (SG1’s The Ark of Truth or Continuum, Farscape’s Peacekeeper wars) – they don’t seem to fit into either tournament. I suppose counting them as an episode or a season may be fair enough. Have to draw a line somewhere, or you end up with Star Trek movies bleeding in. :-)

        • My thoughts on a hybrid series/episode tournament would be to start picking series for the first portion, and then after round one, go through every episode of every surviving series on the IMDB and sort them all by user ratings to populate the next portion of the tournament, again with room for wildcards. Seasons would be similar, but it means less populated tournaments to have multiple rounds.

Comments are closed.