[criptik]

I have directed 40 or so small free tournaments (3-5 tables), all have been unclocked. I always set the number of rounds to 2*T-1 to allow each pair to play every other pair (so for example 7 rounds with 4 tables). As far as I can tell, no one has ever played the same pair twice in a tournament which is great. Often pairs have to wait for a table that is still playing to finish, which is fine.

My question is what are the requirements to allow a pair move to the next round? For example assume 3 tables with round 1 having A playing B, C playing D, E playing F. In the next round assume the system wants the pairing A-C, B-E, D-F. Now let's assume A-B and C-D finish their first round but E-F is still playing. Will A-C be allowed to start their second round without waiting for E-F to finish? Or do both A and B have to be able to move before the move occurs, so A in effect ends up waiting for E-F to finish? Asking because whenever I've been waiting for a table to finish, my opponents also seem to have to wait, have not seen a case where one pair gets moved and the other keeps waiting. And thus when there are only 3 tables it appears that everyone moves at the same time.

With 4 or more tables I definitely see cases where some tables move and others are still in the last round which makes sense, but just trying to understand what are the requirements for the move to happen? A related 4-table question would be: one table is finishing round N, everyone else is finished. Obviously 2 pairs have to wait for the last round N table to finish before moving to round N+1, but will the other 4 pairs always be allowed to start round N+1 earlier?

Also given that the system (thankfully) avoids having pairs play twice if at all possible, does it start with a predetermined movement schedule or is the movement "computed on the fly" as pairs finish the early rounds?

[m00036]

criptik, on 2020-August-08, 11:26, said:

I have directed 40 or so small free tournaments (3-5 tables), all have been unclocked. I always set the number of rounds to 2*T-1 to allow each pair to play every other pair (so for example 7 rounds with 4 tables). As far as I can tell, no one has ever played the same pair twice in a tournament which is great. Often pairs have to wait for a table that is still playing to finish, which is fine.

My question is what are the requirements to allow a pair move to the next round? For example assume 3 tables with round 1 having A playing B, C playing D, E playing F. In the next round assume the system wants the pairing A-C, B-E, D-F. Now let's assume A-B and C-D finish their first round but E-F is still playing. Will A-C be allowed to start their second round without waiting for E-F to finish? Or do both A and B have to be able to move before the move occurs, so A in effect ends up waiting for E-F to finish? Asking because whenever I've been waiting for a table to finish, my opponents also seem to have to wait, have not seen a case where one pair gets moved and the other keeps waiting. And thus when there are only 3 tables it appears that everyone moves at the same time.

With 4 or more tables I definitely see cases where some tables move and others are still in the last round which makes sense, but just trying to understand what are the requirements for the move to happen? A related 4-table question would be: one table is finishing round N, everyone else is finished. Obviously 2 pairs have to wait for the last round N table to finish before moving to round N+1, but will the other 4 pairs always be allowed to start round N+1 earlier?

Also given that the system (thankfully) avoids having pairs play twice if at all possible, does it start with a predetermined movement schedule or is the movement "computed on the fly" as pairs finish the early rounds?

I don't know what the algorithm is exactly (BBO normally tries to wait until 4 tables have finished and then start the movement for those tables, then the next 4 etc., with 3 tables for the last few moves as required, but it's possible it occasionally does 2). For 3 tables they will therefore always move at the same time.

Anyway if you want a full Howell (i.e. everyone plays everyone else) then it's best to email BBO, get howell movement rights added and then run a clocked tournament with the Howell movement. There is no point running an unclocked movement if you want every pair to play every other pair since you will always have to wait for the slowest pair!

[criptik]

m00036, on 2020-August-08, 17:18, said:

Anyway if you want a full Howell (i.e. everyone plays everyone else) then it's best to email BBO, get howell movement rights added and then run a clocked tournament with the Howell movement. There is no point running an unclocked movement if you want every pair to play every other pair since you will always have to wait for the slowest pair!

Our group prefers unclocked because they don't want to have a board terminated with an average in the middle of a board or even with a very few tricks left to play. (And the director does not want to have to go in and assign fair results to such boards). As I mentioned we have had no problems with the movements over the many tournaments we've had, i.e. assuming the number of rounds is set up properly, everyone always plays everyone else and no one plays the same pair twice. So I'm not sure what the advantage would be of an official clocked howell movement. I also believe the total tournament time can be less with unclocked (but not by much).

[criptik]

In my original post on this topic I said in my experience with unclocked movements with Rounds = TotalPairs - 1, I had not seen anyone play the same pair twice. But that experience was with 3 or 4 tables. I have since seen that with 5 or more tables, you can indeed play the same pair more than once. Is there a reason it works well with 4 tables but not with 5? Does the system generally try to avoid playing the same pair more than once, but revert to replaying under some conditions? If so, what are those conditions?

[peterb001]

criptik, on 2020-August-22, 16:56, said:

Our group prefers unclocked because they don't want to have a board terminated with an average in the middle of a board or even with a very few tricks left to play. (And the director does not want to have to go in and assign fair results to such boards). As I mentioned we have had no problems with the movements over the many tournaments we've had, i.e. assuming the number of rounds is set up properly, everyone always plays everyone else and no one plays the same pair twice. So I'm not sure what the advantage would be of an official clocked howell movement. I also believe the total tournament time can be less with unclocked (but not by much).

You can still run a clocked tournament with long rounds (10+ minutes a board), and you can even extend it before the round ends if you need to! The system moves everyone once all tables are finished, even though there is time left on the round.

[mycroft]

The other issue with unclocked (and the reason people like it) is that the faster pairs can move. With the 10 minute rounds, we're all stuck twiddling our fingers for another 5, 6 minutes over 7 board every round, but we can't do what we do IRL and walk away for 5 minutes because the round *could be called* at any time. If your club doesn't have a large proportion of those faster players, then fine (my church basement didn't, either. I don't play there any more).

The downside of unclocked is that eventually that catches up with you - if there really is a "fast field" and a "slow field", then eventually the system doesn't have a new "fast pair" to put you with, so you have to either replay or wait (and now the wait is 15 minutes or more!) for the slow field to catch up and give you new opponents. If you have a field that is basically the same, just "a little slower than normal", then you don't have that issue. But I certainly remember some unclocked tournaments I played in where I ended up playing a pair three times, because there were 4 tables out of 15 playing at my speed.

The other downside, of course, is that for the results, you have to wait until everyone's done, so my tournament started at 1900, finished about 2100, and I got to wait for 30-50 minutes for everyone else.

[pescetom]

mycroft, on 2021-February-24, 13:42, said:

The other issue with unclocked (and the reason people like it) is that the faster pairs can move. With the 10 minute rounds, we're all stuck twiddling our fingers for another 5, 6 minutes over 7 board every round, but we can't do what we do IRL and walk away for 5 minutes because the round *could be called* at any time. If your club doesn't have a large proportion of those faster players, then fine (my church basement didn't, either. I don't play there any more).

The downside of unclocked is that eventually that catches up with you - if there really is a "fast field" and a "slow field", then eventually the system doesn't have a new "fast pair" to put you with, so you have to either replay or wait (and now the wait is 15 minutes or more!) for the slow field to catch up and give you new opponents. If you have a field that is basically the same, just "a little slower than normal", then you don't have that issue. But I certainly remember some unclocked tournaments I played in where I ended up playing a pair three times, because there were 4 tables out of 15 playing at my speed.

The other downside, of course, is that for the results, you have to wait until everyone's done, so my tournament started at 1900, finished about 2100, and I got to wait for 30-50 minutes for everyone else.

Fully agree. With my club (probably not much above your church basement, but with several faster pairs) I found that unclocked easily brought howls of protests due to replays, unless I was very careful/lucky about the number of rounds. Plus immense delay for TD and the slower players themselves. I went back to sorting out occasional interrupted boards.

[criptik]

OK next time we have 5 or more tables (which sadly is not very often) I will give clocked a try.

[criptik]

Would still like to understand how the system decides who to schedule next in an unclocked tournament. As mentioned previously, I have not seen any repeats when we have 4 tables, and there are times when people have to wait 15 minutes. For example 3 rounds in, the 4 pairs A,B,C,D have all played each other and E,F,G,H are still playing, the system will always wait (no matter how long) for one of E,F,G or H to finish rather than have A,B,C, or D repeat. This is good, as we would rather wait and avoid playing the same pair twice. The problem is that the above logic does not seem to hold with 5 or more tables.

[mycroft]

I bet they have a "block" of 4 tables that triggers seating. But it's just a guess.

If you're allowed to use the howell command, that will give you N-1 rounds (in a movement that is not at all Howell, but with duplicated boards gets the job done) with N-1 pairs in a clocked event.