Tournament Timeframe


#1

I know OGS has a longstanding tradition of drawn out tournaments, but here’s my suggestion to make title tournaments a little more consistent, say.

Require titled tournaments to start and end as opposed to just start in the year that’s mentioned in their title. If it’s the 2018 Meijin whatever, it should conclude in 2018. I’m looking at the 2017 Meijin 9x9 and considering 2018 is almost over I’m wondering why the finals haven’t even started yet. Knowing it’s a Bo5, if it goes to 5 matches with that time limit, we’re easily into 2019 and probably closer to 2020 by the time it concludes.

Edit: Meijin Nines concluded 4-1, but the table doesn’t show it yet (5 months after). My bad.

It’s like compound interest, at some point you might know the title challenger for 2 years ago but you’re still waiting for the decision on who’s the title holder from 3 years ago, and moreover how do you know whether to allow one of the finalists in “next year’s” edition, because it’s nonsensical to have the title holder compete in the challenger’s tournament but at some point they could reasonably enroll - and mark5000, you’ve in fact set the precedent.

Edit: This seems to be provisioned by the current rules, even: “If the title holder wins the tournament again, the title match will be held between the title holder and the 2nd place.” I strongly suggest to throw title-holder enrolment out the window.

All it does is lead to an unfair advantage (you get multiple opportunities to eliminate the stronger challengers by beating top contenders and losing to people who aren’t top contenders).

It’s a mess.

How do we make sure the tournaments start and end in the same year? There are many different time settings we can explore to find the best solution, but Fischer time is not among them.


The neverending story uhuhu uhuhu uhuhu
#2

Super absolute time: all games that aren’t finished by a certain date will be annuled.
No vacations, no pauses, nothing. :smiley:


#3

it would be nice to have (title) tournaments actually finish. of course with absolute timesettings of eg a couple of months the games wouldnt really be correspondence games anymore, but some hybrid. id assume for (most) games to finish in that time the players would be required to bump into each other online at least once and play about half their moves, if the game progresses at normal correspondence pace otherwise.

alternatively the paired players could be required to set a date within the alotted timeframe for the round and play the whole thing live. (doing that can be hard, considering timezones and language barriers. a tool would help.)
its hard to think such a format would not end in mass timeout/dq, but seeing the disqualification by timeout-ratio in long tournaments now… maybe its worth a try.

i dont think vacation can be disabled atm, although it has been discussed in the past to create the option for tournaments.


#4

Lets estimate what time per move we need to reach that goal for the Main title tournaments (19x19):
Let the tournament start at January the 1st. Than we have 360 days for the whole tournament.
The tournament lasts for 3 rounds, therefore each game has to end within 120 days.
Each player has to play at least 1 move per day! (For a 240 move game)
Or from a game perspective an average of a move per 12 hours per game.

At the moment the time increment of this tournament is 1 day (the above calculated max value).
If you lower the increment to less than 1 day, than you exclude all players which are only willing/able to login on OGS once per day. (If the opponent played just after they logged out, they need a 1-day increment to not time out.)

Faster time settings will exclude these players from the title tournaments. And since I think to play once per day is legit for a correspondence player, the time settings should allow them to take part in the official tournaments.


My opinion for the different time controls in correspondence games is as follows:

  • Simple: 1 day / move leaves no room for short delays. A player only playing once per day has to login every day at the same time to avoid timeout. [And simple is buggy.]
  • Absolute: If a player stops playing on OGS, you can watch a clock slowly ticking down for a month or 2 before the player times out. And you don’t know if the player comes back in time, leafing you with a game which can continue at any time withing that time frame.
  • Byo-Yomi: Same as simple or absolute, depending on main time.
  • Fisher: It gives you some freedom at what time of day to play and also to don’t play for some days at all. But at least fisher enforce an average game progression equal to the increment time.
  • Canadian: Either like absolute or fisher depending on main time.

Therefor I prefer Fisher time control for correspondence games.


That topic could be related to this one


#5

Well, excluding slow players is as much or as little a problem as it is to exclude those players who aren’t willing to spend multiple years on a single tournament. I’m also not saying all title tourneys should be thusly amended, but at least let there be some variety (for people who like extremely slow title tourneys, there’s the Tianyuan I believe, whose league system makes the whole thing last forever and a day).

As far as I’m concerned, the added legitimacy (no special treatment for previous winners) is a welcome bonus.


#6

I prefer tournaments with a faster progression as well.

I think a large slowdown of the tournaments is caused by player playing exactly 1 move per 24 hours.
There are players participating in all title tournaments who take more than 24 hours on average to respond to an opponents move (calculated over all their active games, not only the tournament ones). You can easily imagine what happens if 2 of them get matched together.

I would like to see an option to restrict players from tournaments, if their average response time is bigger than … (for title tournaments maybe something like 20h?)

To lower the vacation time to 1 month per year (at least for tournament games and ladders) could help as well. 3 month per year are just too much.


It’s the Kuksu Main Title Tournament and the league system makes it actually the fastest (it ends within a year). It’s 1 round Round Robin per league and cycle. The winner in League A gets the title.


#7

What do you know! Sometimes all you need to do is to make a mistake. :stuck_out_tongue:

Thanks for the info. And I agree, in the worst case scenario (two slow site supporters) both players’ weekend pauses + vacation add up to a ridiculous 8+ months a year (2*52 + 2*73 = 250 days), leaving 115 days which makes for about 58 moves per player; the average 250 move game would take 4 years. ^^


#8

Simple solution: live title tournaments!


#9

That’s a great idea, someone should write a proposal! :partying_face:


#10

Another case in point:

The 2014 edition just ended this year. C’mon folks, this is silly.


#11

I can’t argue with the point that the naming of these tournaments is a bit silly. (It just doesn’t bother me much.)

But I do want to point out that I like the format of these title tournaments as such a lot. The time settings, the frequency and the fact that many seem to view them as ‘main’ tournaments, so the number of participants is fairly large in comparison to others. So if you find something else which better fits on the name ‘title tournament 20xx’, fine. But please let’s not give up on a comparatively popular format.


#12

Well the least we could do is to create live (versions of?) title tourneys.

The silly naming could be taken care of by calling it the 1st, 2nd, … Honinbo or whatever, exactly how Japan did it.