How to Speed Up Tournaments - Help!

Your average time-per-move is 6 hours! You play pretty quickly, when compared with others.

I wouldn't worry about it. If anyone ever said you were moving too slow, they're just being impatient. It's more likely that everyone else in the tournament is just really fast. 

I agree, your rate here is totally acceptable.

Phew! Thank you both for the reassurance. I did think others were really fast but I still don't want people grumbling about my speed :-)

 1.) From a statistical point of view it is certainly true that the probability that the last unfinished game of a tournament round is decisive for the proceeding of anybody to the next round is far below 50 %. This is basically due to the fact, that only a small fraction (1 of 6, or 1 of 5, e.g.) of the players of a tournament group advances to the next round. So most of the last games of a round (and even of a group) are in this sense "meaningless" and in many cases a tournament round can be shortened by days if there is an effective algorithm implemented to detect these "meaningless" games.

2.) I think this algorithm should not be dependend of the number of unfinished games (5 has been proposed by erik). Consider for example the case, where one player is at vacation and has finished (and lost) only two games against the group leader, who has won all his games in the meantime. This player may have 8 unfinished games (suppose a typical group size of 6) but none of these games can effect the forthcoming of anybody to the next round (suppose that one player per group advances to the next round).

3.) The problem of deciding, whether a tournament can be pushed up to the next round, reduces to the problem, whether the outstanding games of a tournament group can effect the standings in that group that are relevant for advancement. If this is not the case, a group can be marked as "fix". If all groups are marked as fix, the next round can start. At the moment a group is marked as "fix" only, when all games of that group have been finished. This should be improved.

4.) In general it is not necessary (because it's too time consuming) to calculate all possible combinations of points (and tie breaks) for the open games of a group (in the worst case all games are still open). If one is settled for a prediction for most of all possible cases, the following procedure can be applied to decide the question, whether unfinished games of a group have got any influence on group players advancement or not, regardless how many games have already been finished within a group or not:
Calculate for each player (except the group leader) his final score under the condition that each of his outstanding games is won for him. If any player's score now exceeds the current group leader's score or is equal to the current group leader's score, no decision can be made (the group cannot be marked as "fix"). If no (maximum available) score of any player can exceed or reach the current group leader's score, the group can be marked as "fix", regardless of the number of its outstanding games. (Note, that tie break points are not taken into account) In the case where two or more players of a group will regularly advance to the next round, the procedure must be performed for the current group second (or third...), by checking all successive players under the condition, that they win all their games, for reaching or exceeding his score.

this is surely the right way to go. but when we make this calculation? it is certainly a large query to execute for the server, so when do we do it? only when a TD clicks on "advance rounds" or something? we can't just have queries like this running all the time... :)

Hmmm...that's true. Each chess player plays a chess game for one reason: to win. So no matter if the match affects the group's standing, you'll just have to wait and be patient...

Patience is virtue.

Iagree again, but:

A lot of people (maybe even the majority) want tournaments to progress at a much faster rate. At the moment, there are no mechanisms to create such tournaments. Only the players who have no problem with long drawn out tournaments are catered for.

Maybe there should be an (optional) average time per move setting when setting up tournaments, if a players exceeds this for the tournament, he forfeits his games. If he doesn't like that, there are still many other tournaments without this restriction, he has an option. At the moment, fast players have no option.

 Well, my humble self is occupationally active in programming issues, too. I.m.h.o there should be some client procedure, which periodically (once or twice a day, e. g.) queries each current tournament standing (this would mean just one or two standings query per tournament per day in addition to the usual user queries). With the standings data the corresponding calculations can be done (without loading the server), which possibly result in changed "fix"-marks for each group (some UPDATEs). Finally the round-advance-flag for the tournament should be calculated. Alternatively the calculation can be done for one group each time a tournament game has been finished in that group (if that group hasn't been already marked as fixed, of course). The last way, I would prefer, if it is practical from a performance point of view.

How is it checked at the moment, that all group games are finished? I guess this is also a periodic procedure not a manual one, isn't it? (but may be a server stored procedure at the moment...)

There is yet another point: A group should only be marked as fix, if all players (regularly one player) that will definitely advance to the next round have finished all their games in the current round. Otherwise one player could claim disadvantages from too many simulaneous games.

IMO alot of people here just post to complain about people that take the alotted time that they are entitled to take, and complaining about the use of vacation. 

Unless a gun is held to your head, do not enter a tournament that is longer than 1 move a day. 

You mentioned two aspects of the problem, namely the individual player's aim to win and the group-leading player's (or players) aim to advance into the next round as quick as possible. If I interpret your statement right, you said, the latter one has to wait (be patient), until the first one's aim has been achieved (or not achieved in the case of lose). But I think these two aspects (or aims) can be achieved independently of each other, because neither the individuals aim to win nor the final groups standing in the current round (which will not be lost) are touched in any way, if the next round starts as soon as (mathematically) possible. I think we all agree that no "time pressure" should be exerted on anybody in any way, which wouldn't even be necessary for those games that have got no influence on the next tournament round, if that round could start before those games are finished. And I agree with you, if the last game of a tournament's round is a game between two group leaders and it will decide on forthcoming of one or the other, everybody (participating in that tournament) has to wait and be patient and should not complain until this (decisive) game has been finished even if they exhaust their complete vacation times...

I am in a one move per day tournament that started April 1. (guess the date should have been a clue!)

Again, isn't there a way to set up a tournament that would suit all the people who complain? There are a lot of those, atm ONLY the players who don't mind tournaments taking forever are catered for, there is no way to set up a tournament that can guarantee that it will be comleted  by a specific date.

Well, what about this:
No vacation allowed, One move per day, If one game lasts longer than 80 draws it will be valued as a draw. From a mathematical point of view it is guaranteed that this kind of tournament is completed not later than a specific date which depends on the number of participants (for example the first round will not last longer than 80 * 2 = 160 days which is nearly half a year!). Even if you omit the last condition, most of that kind of tournaments will still finish in time... - I personally think that vacation is not the main reason why tournaments last longer than one expects. In the example above which sounds quite satisfactory at first glance, the first round may take 160 days in the worst case and this without vacation allowed...

On 2 Oct I created a 7 player tournament, friends only. I wanted to make it end really quick, but I made a few mistakes in the setup.

I set NO VACATION (good)

2 days per move (bad, should have been 1 day)

Simultaneous games, no (bad, should have been yes)

Even so, there are only 4 games remaining, but two players are still on their 1st game. If I set it up properly, it might have finished within the month.

Next time I'll see if I can tweak it so that it moves even faster. Whether it is possible with many more players, I really don't know.

Well, I'm in a tournament (1st round), and I'm waiting for it to be over. The thing is, it is my group which is not decided yet. Currently, I'm in second place with 5 points, but the third player has 4 points with two games remaining, and I may be eliminated.

This is a decisive group for the next round, so it is well worth the wait. Besides, I only have a very slim chance of going through to the next round, since that third player is much stronger in chess than me.

As I read through this, an idea came across my mind.

Say for example, we have a group in which 2 advances to the next round. That 2 are already confirmed, but another 2 players in the same group are still playing, and do not have any chance of going through, even in the case of a win.

So in cases like this, why not settle the unfinished, unnecessary game as a draw? It would please the players who have been waiting for the next round. If the 2 eliminated players are unhappy, one can send a seek to the other with the SAME settings as specified in the tournament and play a rematch.

For me, this would please everyone...

 I think those "unnecessary" games are still important for the players playing them (not so much for the others of course). And setting them as a draw will not be necessary if one can find a way to start the next round even if not all games are finished. If such solution can be find (and I'm quite sure it is possible), that would clearly be the better way.

the spirit of any tournament is to follow the rules and have fun.  If you wanted a tournament where people move several times a day you should make one with a time control of 4 hours I'm not sure such a thing is an option on this site but you could suggest it to the staff, as long as everyone going in is aware of the time control they can't complain.  

the only way for a person to not respect your "fast" one move a day tournament would be if that person ran out of time.  

Tonymtbird, you claim that anything that anyone can possibly do that is within the rules is therefore a respectful act. Most other people think respectful means something different.

I think a much bigger problem than this is the sheer number of tournaments that are created but never start. The system needs to be rationalized somehow. May a rule stating that if a tournament is not, say, 65% full after a certain period of time, it automatically dies.

Concerning the topic of this post, one way of speeding games up would be to limit the number of turn-based games anyone can play at any time. Some people are playing literally hundreds of games at once. No wonder why they are slow...

No way you can settle that game with a draw. I like the out of the box thinking, but it's not a good idea. People don't only come to the Olympics to try to get on the podium. A much better idea would be to allow the games which don't affect the outcome to continue, but also for the tournament itself to continue to the next round.

What about a no vacation 1 day/move tournament? I wouldn't play it but it could be done. The tournament I created, a 3 day/move no vacation tournament was over in about 4.5 months. It was a 21 person, 2 round tournament.