I've been looking around different games to see how tournament creation is being handled by other players and I was pleasantly surprised. Over the last 10 months I've been on BGA I see more and more players organizing tournaments for their favorite games and seeing good settings and tournament formats overall. Tournaments are generally being made 0-2 months in advance. There's never a need to create a deeper backlog than that. I've even seen players that create monthly tournaments with a 1 year backlog, but restrict the registration to 30 days (which is default BTW) to minimize the absent player issue.
I don't see very many games that "need" more tournaments added to their lists. Looks like many tournaments are being made across many games.
Special shoutout to Baleae and others that organize many tournaments across many different games.
https://boardgamearena.com/player?id=56625950 https://boardgamearena.com/tournament?id=360251 here's an example of a good Baleae tournament for 2p Odin with Group stage: 5 rounds of Swiss, then second stage is single elimination with an even 32 players.
I'm posting here as reference for tournament creators so we can all create good tournaments.
In terms of "bad" tournaments to delete, I would recommend creators delete tournaments that have broken settings. Any settings that will lead to unintended results due to tournament bugs are "bad" in my opinion.
SWISS TOURNAMENTS
For example many future Swiss tournaments with the "try to prevent same players to be opponents twice" settings will become broken and are therefore "bad" in my opinion due to this pairing bug:
https://boardgamearena.com/bug?id=56504 here's the bug report on pairings.
Under normal Swiss tournament formats this setting works fine (high player count, low number of matches, enough to sort players), but tournaments become broken when there are too many matches compared to the number of players. Once the algorithm cannot make pairings without duplicate pairings it "breaks" (BGA needs to update their coding) and makes broken pairings for the rest of the tournament? How bad is the broken pairings? REALLY BAD, it matches 1st place with the bottom players (likewise with 2nd, 3rd...) over and over, every round.
What does this do to a Swiss tournament? It completely spoils the tournament for that point onwards. Pairings are broken so the top player plays against the weakest players over and over again, maintaining their lead for the whole tournament and it creates poor matchups for everyone the whole time.
How do you know a Swiss tournament with "try to prevent" pairings will have enough plays to function for the number of rounds? Pretty easy, first count the number of opponents each player will have across all matches. For example, 8 matches for a 2 player game will have 8 opponents (8x1) and 20 matches for a 4 player game will have 60 opponents (20x3). Next, look at the player count for the tournament if Swiss tournament or if Group stage, determine the player count per group. In Group Stage, each stage 1 group is considered it's own separate tournament.
There obviously needs to be more players than the number of opponents or else the pairing is guaranteed to break part way through the tournament and therefore the tournament is "bad". However, for games with more than 2p there will likely need to have even more than that many players to function because each table needs to consider everyone paired against everyone, no duplicates which becomes harder as player count goes up. We don't know the details of the BGA Swiss pairing algorithm for 3+ player games, but there have been many examples posted on forums where it struggles to make sensible pairings for 3+ player games when the number of previous opponents gets too high and remaining players are limited.
I would recommend that the number of players is AT LEAST DOUBLE the number of opponents, which should provide some "wiggle room" for pairing with more than 2 players per table. This ensures that the players at the top of the standings will only be paired with players in the top half of the standings. Swiss tournaments are supposed to pair players with the same score or as close as possible each round to create equal matchmaking each round and have the players sort themselves based on their performance in these games. If the 1st place in standing player has already played all players in the top half of the standings and need to be paired against players in the bottom half, then this is creating bad matchmaking and undermining the Swiss tournament system. If this is happening, that means the tournament has too many rounds compared to the number of players. The outcome of the tournament has basically been decided at this point because the leading player(s) will get easier and easier matchups and maintain their lead without much effort). I would consider tournaments that are technically functional, but low on player count to be "poor" settings in my opinion, but at least function
To avoid this happening for small groups compared to round number I would recommend the "don't try to prevent" settings, which will allow some duplicate pairings where needed, but will be doing better matchmaking in the process.
Based on this, any Swiss tournament with an insufficient player count to be in the "bad" category (#players < 1x number of opponents) should be deleted prior to tournament start and tournaments in the "poor" category (1x number of opponents < #players < 2x number of opponents) should be carefully reviewed prior to starting to ensure they are functional. Ideally tournament creators would create sensible minimum player counts for Swiss tournaments to prevent this issue.
As an example, this Federation tournament should have never been allowed to start with those settings:
https://boardgamearena.com/tournament?id=270995 this 4 player swiss tournament has 20 matches with "try to prevent" settings and only 50 players which means it doesn't even meet the BARE MINIMUM number of opponents to function properly. Each player will need to have 60 unique opponents for this tournament to function correctly, but it needs more than 60 players to be able to create all unique pairings at every table (each 4p table is 6 "pairings"). Not sure the exact number, but perhaps 80-100 players would technically be able to function. I would recommend AT LEAST 120 PLAYERS for these settings to function properly.
For 50 players this tournament would have functioned fine with 8 matches (3x8=24 opponents, 50 players). As of writing this the tournament is finishing Round 9 and the pairing are getting a bit wonky. 1st place was paired against 16th, 25th and 28th place out of 50 players because all the stronger players have already been played. This will continue to get worse round after round until the pairing completely break around Round 16 or earlier and behave like in the bug report above. The players have already been sorted after 9 rounds, the tournament should already be over here. Each round will get worse and worse from here.
If this tournament was created with "don't try to prevent" settings everything would be completely fine. Players would be sorted already after 8-9 matches and players would play against similar score opponents over and over to determine a winner. It would have been a fun back and forth battle among the top ~16 or so players with duplicate opponents at top tables to settle an overall winner.
Group Stage Tournaments
Second stage swiss tournament mode is broken as per this bug report
https://boardgamearena.com/bug?id=105389 . Don't use it unless you want exactly 5 matches because you will get 5 matches regardless of what you enter.
Any Group Stage tournament created that will be affected by this bug is therefore "bad" because it is literally broken and should be canceled. The "Long Game" tournaments are examples of this
https://boardgamearena.com/tournament?id=359356 this is a future 2/3p Five Tribes tournament.
This tournament format has many issues. The second stage Swiss will only be 5 matches instead of 20 matches like it says on the page due to the bug so the description is misleading. There's no way of knowing when this bug will be fixed and if a fix will affect this tournament on time.
Other issues with this tournament are the "try to prevent" pairings and player counts. 80 player minimum is actually not enough for Stage 1 because each group (2 groups) may have only 40 players and each player should get 40 opponents. Minimum 100-120 players would be technically functional, but I would recommend 160 minimum for these Stage 1 settings.
The other issue is with the Second Stage again. The intent was to create 20 matches of Swiss with 42 players advancing, ignoring the "always 5 matches" bug. This player count is fixed and determined by the organizer. Each player in Stage 2 will have 2 opponents per match so 40 opponents
over the 20 matches. This means if it works correctly each player will play against everyone expect for 1 player. However, because each table is 3 player the algorithm might not be able to prevent duplicate pairing at all tables (3 pairings at every table) so the pairings might break earlier than that. These settings would work fine with 10 matches or at least 60 players qualified, but would likely break before the end. Regardless, this tournament will only have 5 matches for second stage unless the bug is fixed.
Tournament Creators, we very careful with your Group Stage settings for Swiss. Avoid Swiss for second stage and ensure your minimum player counts are reasonable for the number of matches and player count (2p, 3p, 4p..).