L19 Tournament Alternative Structure Idea

[ez4u]

I have been playing with an idea for an alternative structure for (at least some of) the L19 tournaments. What do you think of the idea described below? Does anyone think it sounds like fun? What are the potential problems (and potential solutions please)?

Why?
I find even the double elimination format to be slower and less active than I would like. IMHO the need to schedule a match with exactly one other person seems to slow things down considerably. Yet I did not particularly want to go to something full-blown like a Swiss. I hoped for something in between that would stimulate activity and lead to games that are interesting to the participants and the fans.

What?
What if we structured the/a tournament as a series of rank-driven mini-leagues, starting with the lower ranked players and producing a winner at each level who gets promoted into the following league?

For Example:
The current 2012 tournament has the following members. What might this mean with such a set of participants?
HermanHiddema (herminator) 4d
crux (venkman) 3d
ez4u 3d
Uberdude 3d
redundant 2d
stalkor 2d
Loons (Practise) 2d
lobotommy (tommyray) 1d
topazg 1d
quantumf 1d
Laman 1d
lindentree 1d
Jedo 1d
Celebrir 1k
Marcus (Marcus361) 2k
LucianoS 2k
danielxr 2k
Tryss 2k
jts (jtscarry) 3k
Rafa 4k
speedchase 4k
oren 5k
rpchuang 5k
Nathan 6k
NinG 6k
phillip1882 6k
hailthorn011 (hailthorn) 7k
schultz 8k
Jordus 9k
lesenv 9k
TwitchyGo 9k
Lincarte 10k

As one example of what I am proposing, we might start with a 9-person league of the members ranked 6k and below...

League I
Nathan 6k
NinG 6k
phillip1882 6k
hailthorn011 (hailthorn) 7k
schultz 8k
Jordus 9k
lesenv 9k
TwitchyGo 9k
Lincarte 10k
They would play games against each of their 8 rivals and then we would celebrate the success of the winner of League I. We would then follow with a 10-person league of the players ranked 2k-5k plus the winner of league I...

League II
Marcus (Marcus361) 2k
LucianoS 2k
danielxr 2k
Tryss 2k
jts (jtscarry) 3k
Rafa 4k
speedchase 4k
oren 5k
rpchuang 5k
Winner of League I
They would each play 9 games against their rivals and we would have a second celebration in honor of the winner of League II. We would then follow with an 8-person league of the players ranked 1d-1k plus the winner of league II...

League III
lobotommy (tommyray) 1d
topazg 1d
quantumf 1d
Laman 1d
lindentree 1d
Jedo 1d
Celebrir 1k
Winner of League II
They would each play 7 games against their rivals and again we would celebrate a winner. We would then follow with the final 8-person league of the players randed 4d-2d plus the winner of league III...

League IV
HermanHiddema (herminator) 4d
crux (venkman) 3d
ez4u 3d
Uberdude 3d
redundant 2d
stalkor 2d
Loons (Practise) 2d
Winner of league III
They would each play 7 games against their rivals and we would celebrate the winner of the tournament... Or we could add something else at the end like a match with the final league winner and last year's champ.

Notice several things here:
- The league size is not particularly important, so we can fit the leagues to the available participants. The example above has four leagues, ranging from 8 to 10 players, but we could also produce a schedule with five smaller leagues or three larger leagues (or four leagues ranging from 6 to 12 players for that matter).
- Having someone sign up but then not play is not particularly damaging since we can simply drop them off the list without disrupting things significantly. We do have to decide how to handle people who drop out mid-way however.
- Having two people who are supposed to play but find it difficult to arrange an agreeable time is less disruptive since within each league the order of play is not important. Such pairs can work on finding the right time over the course of the overall league schedule (one month? six weeks? TBD).
- Since each player has a range of opponents at the same time, it is possible to just cruise the L19 room on kgs, looking for any of your opponents who might be logged in at the time. We may see bursts of activity in the L19 room, particularly at the start of league play. This may be of more interest to our fan base.

[jts]

How about a slightly more baroque idea - a title-holder, a league 8 or 10 players who play swiss to determine the challenger, and a preliminary filed that plays double-elimination (or even single elimination) to replace the lowest-ranked player(s) in the league?

I think this fits with the intention better. We all would love to see more hard-fought games between the strongest members of the forum. We also don't necessarily want to wait for the other 30 players to play 100 games before we get to the final stages of the tournament. (At the rate last year's tournament went/is going/will have gone, that could take years.)

[Joaz Banbeck]

I too am not a fan of the double elimination. ( I only ran the last tournament that way because a quick poll showed that the majority preferred it. ) I suggest a total points winner. Not winner of the greatest number of games, but winner of the greatest net number of stones.

A large number of the games played in our double elimination tournaments are lopsided, and both players know it at the beginning of the game. The motivations of both players make for less than the best go. The weaker player gives it his best try, but about 30 or 40 moves in, he knows that he is behind, and often stays in just looking for a place to resign. The stronger player has an incentive to play solid but unenterprising moves, waiting for the inevitable mistake, upon which he will capitalize to turn the game into an easy win.

Converesely, a total points tournament would bring out the best. A 5D could be playing a 10K, and he knows that he has to win by at least 150 points, because the other 5D beat a 9K yesterday by 140. The weaker player is in the numerically reverse situation, hoping to fininsh the tournament with no more that a net negative 1200 points, which may be enough to win his class. So it becomes an exciting game for both, right down to the last one-third yose point.

A total points tournament could be run as a Swiss, a McMahon, maybe even a ladder.

[HKA]

I too am not a fan of the double elimination. ( I only ran the last tournament that way because a quick poll showed that the majority preferred it. ) I suggest a total points winner. Not winner of the greatest number of games, but winner of the greatest net number of stones.

A large number of the games played in our double elimination tournaments are lopsided, and both players know it at the beginning of the game. The motivations of both players make for less than the best go. The weaker player gives it his best try, but about 30 or 40 moves in, he knows that he is behind, and often stays in just looking for a place to resign. The stronger player has an incentive to play solid but unenterprising moves, waiting for the inevitable mistake, upon which he will capitalize to turn the game into an easy win.

Converesely, a total points tournament would bring out the best. A 5D could be playing a 10K, and he knows that he has to win by at least 150 points, because the other 5D beat a 9K yesterday by 140. The weaker player is in the numerically reverse situation, hoping to fininsh the tournament with no more that a net negative 1200 points, which may be enough to win his class. So it becomes an exciting game for both, right down to the last one-third yose point.

A total points tournament could be run as a Swiss, a McMahon, maybe even a ladder.

The comparison between the current system in which - "The motivations of both players make for less than the best go" and Joaz's preferred method "A 5D could be playing a 10K, and he knows that he has to win by at least 150 pts" is an interesting one.

Having just run the Maryland Open, and been overwhelmed by the number of dan level players who attended, only to be disapointed by the turn out weaker than 5kyu, I am concerned that we are increasingly damming the river of go players and choking off many of them. We all want to be dan level and if we cannot squeeze through and make it, somehow we do not belong - this is in my view a disaster.

While I do not endorse the current system as perfect - I think it is a rare chance of of weaker players getting AND stronger players giving a serious even game. Lopsided, perhaps, but I think it is a wonderful moment of inclusiveness. And I do not think the go it produces it so bad.

Stronger players trying to humiliate weaker players by winning by over 100 pts, on the other hand, produces some really, really nasty go and reinforces the divide and humiliates the weaker player. Go is about winning by one point - when you set an artificial numerical komi the game is truly distorted and, in my view, not useful for either player.

[jts]

I agree that it would be fun to have a tournament where stronger and weaker players were both doing their best and could be satisfied with "the best they could do", but also that go is about W+0.5 or B+0.5, not running up the score.

This is part of why I like ez4u's suggestion, or my hybrid suggestion - a beginner can tally up advancing to the big boy division as a concrete accomplishment, and be proud of it even if he has the worst record in that division.

Perhaps an alternative version of Joaz' idea would be to let players negotiate a handicap, and award 10 points for a victory, modified by one for each handicap stone. In an even game if a player is up fifty points, trying to double his lead with overplays is a fairly cheap tactic. On the other hand, if he decides in advance that he wants a 9H game rather than a 5H game, he is running a strong risk that the weaker player will skunk him, and he'll get no points for the game at all.

[ez4u]

A couple of things:

- For advocates of a point-based tournament, please see the Hahn Tournament threads in this forum. That version at least simply died out in mid-tournament. Why was never discussed. Personally I think the reality was less interesting than the expectations (particularly because the anticipated excitement depends on a counting ability that most of the participants, like me, do not possess) but others may have different opinions.

- For people who do not like the idea of waiting for the lower level leagues to complete their games, we are probably imagining different outcomes since I did not describe full details in my example. Personally I imagine running the tournament leagues something like the ASR league at present, i.e. all games have to be finished by 23:59 GMT on the last day of the month (or other time limit). Experience in the L19 tournaments has shown that some people have more trouble than others at finding a mutually agreeable time to play (hint, hint herminator and crux!). I would propose that we run on a tighter schedule and possibly penalize unplayed games.

- Again on the waiting time, I think it is quite reasonable to run multiple tournaments at once. For example, we could run two tournaments a year and start the lower level league while the higher level league(s) of the previous tournament were still running. For example, if we had a four-league structure with two-month league playing schedules, each tournament would require eight months to run. League I of the latest tournament would overlap with League IV of the earlier tournament.

[jts]

In academia they talk about "methods-driven research" - here I worry we're starting to meander into methods-driven tournaments. Lets try to switch this around by figuring out what problems we're trying to solve.

First cut:
1. Were there sufficiently serious problems with the last tournament that we want to thrash them out and choose a different tournament structure before this years' tournament really gets going?
2. Were there some problems with the tournament, and so therefore we want to start an additional tournament as a way to experiment with a different organizational form?
3. Do we want to start an additional tournament because additional tournaments are fun, and so long as we're doing that, we might as well improve the structure?

These three possible reasons to design a new tournament format point to various solutions.

Second cut:
Are these all the different goals people have brought up?
1. We want to play more games
2. We want the tournament to progress faster
3. We want a higher level of activity among participants
4. We want more games that attract a lot of observer interest
5. We want games that are more interesting to the players
6. We want to easily discard people who sign up and drop out
7. We want to avoid delays caused by two people not being able to schedule their game
8. We want the games to be more competitive
9. We want weaker players to feel included in the tournament
10. We want lots of people to participate

I'm sure I'm missing some, but just to keep in mind how multifaceted the demands that we're putting on the tournament structure are. To a first approximation, some of these goals are actually incompatible - for example, anything that increases the number of games will make the tournament slower.

[Joaz Banbeck]

...Were there some problems with the tournament, and so therefore we want to start an additional tournament as a way to experiment with a different organizational form?...

That's my assumption.


The biggest problems with the current scheme:
1) Everyone has to wait on one game to be played.
2) Many of the games are lopsided.
3) The strongest players - who presumably would play the most entertaining games - don't show up until late in the tournament.

[jts]

Since ez4u, Joaz and I were the only ones intrigued by the prospect of building a better mousetrap, I'm going to assume that most people on the forum didn't think there were problems with the last tournament serious enough to redesign it.

[oren]

Since ez4u, Joaz and I were the only ones intrigued by the prospect of building a better mousetrap, I'm going to assume that most people on the forum didn't think there were problems with the last tournament serious enough to redesign it.

I think most players will just play in whatever is proposed. The format could change all the time.

[HermanHiddema]

One of the advantages of the current system is that people know they are not signing up for a lot of games. If you play 8-10 person leagues, with a deadline of one month to six weeks, you're signing up for more than one game per week. Some people may find that too much.

Perhaps, to accommodate those that do wish to play more games, we could organize a second event, with a different structure?

[karaklis]

Since ez4u, Joaz and I were the only ones intrigued by the prospect of building a better mousetrap, I'm going to assume that most people on the forum didn't think there were problems with the last tournament serious enough to redesign it.

Most, maybe. I didn't sign up this time because the games were spread over such a long time. It would be cool if we had a tournament format that would allow a tournament to be finished in one or two weekends.

[oren]

Most, maybe. I didn't sign up this time because the games were spread over such a long time. It would be cool if we had a tournament format that would allow a tournament to be finished in one or two weekends.

The problem is scheduling with vastly different time zones. Ning and I had to schedule our game over two weeks out.

[bakekoq]

wowh..
timezone is really making u mad, right? lol..

[ez4u]

Since ez4u, Joaz and I were the only ones intrigued by the prospect of building a better mousetrap, I'm going to assume that most people on the forum didn't think there were problems with the last tournament serious enough to redesign it.

I did not make my original post as a proposal to redesign the 2012 tournament. Personally I am in since I am lurking around L19 all the time anyway, but I know there will be nothing to do for a considerable period of time.

IMHO the only way to really test the viability of an alternative design is to attempt to run such a tournament. It might be distracting to do it right now as the regular tournament is just getting underway. However, perhaps around the end of the summer we could give it a try. We could fine tune the proposed structure in the mean time.