Home | About | Tournament Winning Decklists | Forums

How long should tournaments be?


#21

What’s the mathematical equation for number of players in the cut? That’s a great way to save on time. I don’t think 5 rounds for a 25-32 person Regional is too much. And if the cut is Top 4, that seems pretty reasonable; 12%-15% of the attendance making the top cut seems just fine in determining the best players.

If the day isn’t going great and by round 4 or 5 you know can’t make the cut and are sick of losing, it’s not the end of the world to just drop from the lower tables. Or, if you’re there to play Netrunner because you enjoy the game and play in tournaments to get good competitive games (or more practice to get better), having more rounds of Swiss means you have more rounds where the structure pairs you against someone of equal skill, since it takes a few rounds to get enough past the random factor.


#22

I think discussion about IDs is fine here if it relates back to tournament length.

I should mention, though, that since it was a question about Excalibur that kicked this off, I am particularly interested in the question of how long tournaments with single-game matches should be, and not only in what the best length is with the current match structure.


#23

Would you just start with doubling the current Swiss round numbers to convert to single-game matches? So, with my suggestion or 25-32 players it would be 10 games + Top 4. So about 6:40 hours for Swiss (about an hour longer than regular structure), but you save more than an hour with a smaller cut.


#24

Maybe? There’s three issues with that approach, I think.

  1. It’s overkill. Going to eight or nine rounds would still be plenty, and wouldn’t add time regardless of whether the cut was changed.

  2. I think it might be possible to have too many rounds of Swiss? More rounds won’t increase the number of people who can possibly have a perfect record, but it might mean more people tied for first or second place, and it might lead to people being paired further up/down the standings than would be ideal.

  3. With a large number of rounds relative to the player count, it can become impossible to avoid rematches later in the tournament. For instance, with six players, it can be impossible to avoid rematches in round four. I suspect this only affects rather small player counts in practice, but I want to get a better handle on what the limit is before I recommend large numbers of rounds. This one will probably involve me writing a Python script to figure it out soonish.


#25

Right, I would start with doubling, then pare it down once you have data for single-matches. For the two-game structure, there definitely is a need for more rounds than the recommended structure.

What about a way to let them face the same opponent, but opposite match? Netrunner is tricky because of the asymmetrical nature. You kind of have two players in one: one Corp and one Runner. And maybe the true best player is an average of those two records.


#26

Maybe.

I do intend to improve how the software chooses rematches when it can’t avoid them, and forcing the opposite match will definitely be part of it. But I think avoiding rematches is still the better option, and I’d like to know when rematches can become unavoidable because it may affect what tournament structure should be chosen. For instance, for the specific case of a six player tournament, some kind of single-game-match round robin might be better than four rounds of Swiss.


#27

Let X be number of single elimination rounds needed for a given number of players (so e.g. X is 5 for 17 to 32 players). Then I would do X+2 or X+3, depending on how which side is closer to changing value of X. So 7 rounds for 12-23 players, 8 rounds for 24-47 players, 9 rounds for 48-96 players, either with no cut or (if you really want the cut) only the final match, played between leader before last round (who is removed from final swiss round) and leader after last round.

This is more or less the most that can be reasonably played in a single day, so any event with 100+ people should be played over two days - most likely playing 7-9 swiss rounds day 1, then 3 swiss rounds + final match day 2 (usually this is Sunday, so you want to avoid ending late in the evening so people have time to return home and not be very sleepy next day at work).

Finally, for biggest events (big enough that most players come by plane instead of by car, something like Worlds or Euros) I would do two full days of swiss (7-8 rounds each day) with day 3 top cut - 8 people, best-of-three for quarterfinals and semifinals, best-of-five for final match.

This is interesting idea. I wouldn’t completely rule this out, but I would first hard-code that such rematch should happen only after 2-3 rounds from the first game, and even then I’d give the pairing algorithm at least some incentive to avoid such pairings.


#28

I was kind of thinking it might make sense to have more than X rounds for smaller player counts, and then reduce to X rounds - or even X-1 if there’s a cut - for larger numbers, so that fairly large tournaments could still be reasonably played in one day.

For instance, either nine rounds or eight rounds and a cut would theoretically be sufficient for a tournament with up to 1024 players, and would be easy enough to fit in one day.

Presumably there will still need to be different guidelines for different levels of competition, so more competitive events might still have more than X rounds and go to two days.

I think this would be a bit too long. It’s likely more than 2X Swiss rounds. And that cut would be at least five or six hours, which is a long time for a part of the event with only eight players. I think I’d rather see something more like you suggest for regular 100+ people events, but perhaps with a top four cut with best-of-three matches throughout. The third day can be used for draft, Cache Refresh, etc.


#29

Dredging this topic up to add a note that I messed up the time for Netrunner finals for the comparisons in the third post - I have the final round as 60 minutes, but it’s 60 minutes per game, so if it goes to a second game then the cut can be an hour longer than the charts show.

Note that even with this error, Netrunner has the longest tournaments of the games compared.

On the other hand, most Netrunner finals don’t actually take two hours.