We are very near the end of second round in two tournaments I subscribed about one year ago.
Here they are:
As someone told in some other thread, having a “weekly” tournament that lasts more than a year is quite confusing. I didn’t know when I subscribed. Now I do!
To celebrate the (near?) end of second round and start of third (which is the last one, hooray!) I baked some pies.
Only about 63% players manage to pass a round. We already had 126 disqualified players out of 223.
Just 88 players out of 223 survived 'til now.
We also had few players who intentionally left.
I can’t wait to see the start of third round and how many players will resign all of their games immediately (it already happened in previous rounds).
I think I’ll never subscribe one of these tournaments anymore, but I dont’ want to quit. I’ll be there 'til the end!!!
(I prepared a green slice for survivors: it must be filled out)
Neat.
One of the longest-running tourneys sports following distribution:
Signups: 31, Round 6/6
Competing: 7 (22%)
Resigned: 4 (13%)
DQ’d: 20 (65%)
The current Meijin Nines (started on March 1st) looks like this:
Signups: 95, Round 2/3
Competing: 53 (56%)
Resigned: 4 (4%)
DQ’d: 38 (40%)
This leads me to believe that a live tournament could hardly fare worse in terms of finishes.
Compare these numbers to the last big live tournaments, exhibit A and exhibit B. The former was blitz 9x9 double elimination, granted, but 171 participants, 11 rounds, 90 minutes total… that’s pretty promising. Elimination seems to be the way to keep tournament duration limited.
The latter was a large McMahon 9x9 with 40 participants, 19 finishes (48%), 2 resignations (4%) and 19 DQs (48%), lasted 40 minutes.
The longest big (91 participants) live tourney I could find was the summer tournament, which lasted about 21 hours (7 rounds, adds up to main time +komi exactly so most games probably just timed out). That is definitely too long for a live tournament - 2hrs main time + generous 12x5min byo yomi.
Seems pretty normal. Because players get DQ for just one timeout, mcmahon tourneys with large groups tend to have large number of dropouts.
Based on my own corr experience, i have a feeling is that usually ~half the players won’t get past first round.
I think the default correspondence setting of +1 day per move is too high. Maybe +12 hr or less should be experimented with at a larger scale (like in the automatic site-wide tournaments, maybe the ladders, instead of just the fast correspondence group tournaments).
A popular argument against what I am suggesting above is that reducing the increment would prohibit some slower players from being able to participate. However, with the current slow settings (where each game could easily last 6 months or more), tournaments dragging on too long could be discouraging players from participating as well. The attrition rate statistics shown in the original post seem to suggest that this is happening.
I don’t have a horse in this fight, but I want to point out that it is not just “slower” players who would be unable to participate. Although 12-hour increments may be fine for students (who do seem most numerous on OGS), many older players with full schedules might find it impossible to play a move every 12 hours. In my own case, I’m up at 4 a.m., at work by 6:15, and rarely home before 4 p.m., and there is no possibility of playing in that 12-hour period. I imagine that many other working adults here have a similarly demanding job and/or substantial family responsibilities. Not complainin’, just sayin’.
With Fischer +12h, you would still be fine as long as you are able to play 2 moves in 24 hours. Of course, if you are unlucky and draw an opponent with “opposite” schedule of you, so that he doesn’t make any moves while you have time to play, you will still be in trouble (unless you are able to “catch-up” by playing more over the weekend).
I played quite a few chess tournaments before, and usually, 94 out of 100 finished their games in two weeks, then the remaining 6 took 6 months. Annoying as h***. But on chess.com, they kept track of “average time between moves” for all players (tracked per game, so it doesn’t lower just because you have many games), and when setting up a tournament, you could restrict it to only allow players with a certain playing speed (e.g. average time between moves < 6 hours).
Not sure if that is possible here, but it was very useful for setting up fast tournaments as you then only get players that actually plays fast, and it sped up the tournaments immensely. I think rounds of < 2 months could be possible then.
I looked up how long the automatic sitewide and title tournaments last:
Timesettings
Title tournaments: 7d + 1d up to 7d
Automatic sidewide tournaments: 3d + 1d up to 3d (pause on weekend)
Duration of the tournaments
duration
title tournaments
automatic sidewide tournaments
fastest
679 days
318 days
25% ended within
905 days
496 days
50% ended within
946 days
566 days
75% ended within
1112 days
657 days
longest ended after
1936 days
825 days
19 title and 64 automatic tournaments
Duration of games
including timeouts
percent ended within
title tournaments
automatic sitewide tournaments
25%
13 days
7 days
50%
45 days
24 days
75%
93 days
53 days
90%
146 days
90 days
95%
180 days
116 days
99%
251 days
176 days
100%
514 days
346 days
33687 games in title tournaments and 32394 games in automatic sitewide tournaments.
without timeouts
percent ended within
title tournaments
automatic sitewide tournaments
25%
38 days
12 days
50%
72 days
30 days
75%
119 days
62 days
90%
171 days
100 days
95%
203 days
128 days
99%
276 days
187 days
100%
514 days
346 days
17546 games in title tournaments and 18716 games in automatic sitewide tournaments.
median duration per player
including timeouts
percent end their games on average after
title tournaments
automatic sitewide tournaments
25%
14 days
9 days
50%
37 days
21 days
75%
61 days
37 days
90%
87 days
54 days
95%
104 days
70 days
99%
145 days
103 days
100%
210 days
154 days
In title tournaments participated 1538 different players, in automatic sitewide tournaments 1882 players. Of them 268 players participated in both tournament forms.
And just for completeness the difference in speed for the 268 players of both tournaments (someone would ask for it anyway):
In title tournaments participated (at least 1 game ended without timeout) 1244 different players, in automatic sitewide tournaments 1569 players. Of them 230 players participated in both tournament forms.
And just for completeness the difference in speed for the 230 players of both tournaments (someone would ask for it anyway):
Surprising to see that title tournament games take so much longer on average in spite of forcing a higher minimum speed (7 moves per week instead of 5 per week). It seems that the larger buffer is used a lot more than I would have thought. (My assumption basically was that the slowest payers would basically always move with only 1 day or less time left, in which case the max. time of the time buffer would be irrelevant. It clearly isn’t though.
Is anyone honestly surprised that a swiss 19x19 tournament takes that long? If anything, I’d expect longer. Assuming that one game goes to 200 moves, once per round, at a speed of 2 moves/day, the tournament would take over four years to complete…
Btw: This formula seems dangerous for large tournaments. For a tournament with 100 players, one get over 20 rounds. I expected something logarithmic before I found the formula.
]
