In case anyone is interested ...

Our club runs four games per week, one is an evening game that rarely gets as many as 7 tables, so a Howell is always played. Two of the games run anywhere from 14 to 22 tables, the third usually runs 8-14 tables. We deal the hands once per week for the entire week, so we don't know exactly how many tables there will be on any given day. There is also seasonal variation, more tables in winter (we live in Florida).

Up until this summer, the three largest games always ran Mitchell movements, with two sections if there were more than about 15 tables (it varied depending on which director was running the game). An odd number of tables resulted in different numbers of boards played in the two sections, not ideal. Then someone suggested looking into playing Web movements. I was told by a tournament director that Web movements are too complicated for clubs, but this turned out to be wrong. However, it did get me investigating different types of movements, I had had no idea there were so many choices. There does not seem to be much written about this subject on the internet. This is what we seem to have settled on, at least for now.

Our goals in choosing the movement are :

1. Every pair should play all of the boards.

2. It would be nice if each pair played as many of the other pairs as possible, ideally, all of them.

Lesser goals are :

3. Minimize the number of boards per round if there is a sit-out.

4. Have enough stationary pairs to accommodate those who cannot easily move from one table to another.

For 2-6 tables, Howell movements are easy, no surprise there. In fact, for 2, 3, 4, or 5 tables, Howell movements are perfect movements, in that they satisfy both of our primary goals.

For 7 tables, a Howell and a Mitchell are both perfect movements. The Mitchell is, of course, a two winner movement. If an overall winner is desired for any Mitchell game, well, let's just say that isn't so great for a one session game and leave it at that. ACBLScore is not bothered by it though.

6 and 8 table Howell movements are very good, all pairs play all boards, and each pair plays against all but two of the other pairs. An 8 table Mitchell (with a skip) violates the first goal, and also invites a 4 board sit-out. With 7 1/2 tables, the 8 table Howell is also much better than the 7 1/2 table Scrambled Mitchell (a one winner movement), because it mixes the N/S and E/W pairs more thoroughly.

9 and 13 table Mitchell movements are perfect two winner movements.

Now for the problem table counts.

10, 11, and 12 tables are not so good. A 10 table Mitchell is not too bad, each pair plays 27 out of 30 boards, 11 and 12 table Mitchells are just plain ugly. However, 10, 11, and 12 table Howells work very well, all pairs play 13 2-board rounds. There are also quite a few stationary pairs, all those from #14 on up, most will be arrow switching though. The 12 table Howell does require Tables 11 and 12 to share boards on every round. Most of our club members were surprised to discover that a Howell could be run with that many tables (technically, I believe they're called 3/4 Howell movements). The bridge supply vendors do not carry table mats for the 10-12 table Howells, but guide cards can be printed from ACBLScore and given to each pair. Alternatively, table mats can be created using a set of guide cards (it is a pain, but not terribly difficult, and only needs to be done once, not just before game time, though). It is also possible to run Web movements for 10-12 tables (see below), but 2 sets of 27 boards are required, and we're too lazy to deal them when one set will do.

For 14 or more tables (up to and including 18), a single section Mitchell will leave unplayed boards by all pairs. 14 and 15 tables are not too bad, but single sections of 16 or more tables get progressively worse. 14 tables can split into two 7 table sections, playing either Mitchells or Howells. 18 tables can split into two 9 table sections playing Mitchells. Scores should be combined across the sections, it's easy in ACBLScore. 16, 20, 22, etc. do not play so well even in two sections, in that they violate the first goal. That's where the Web movement comes into it's own. Two sets of boards are needed to play two sections anyway, but the Web movement makes one big section with two sets of 26 boards in play. It's actually easier on the dealer, he only needs to deal two sets of 26 boards rather than two sets of 36 boards. A web movement with an even number of tables is really quite easy to run. Just pick Web movement when setting up the game. We have diagrams showing which boards to start at which tables, ACBLScore's instructions are not too clear, there's a surprise. Two tables will have a stack of boards, and pairs move normally, up one table at a time. There are no skips in Web movements. The section is divided into two subsections, boards stay in the same subsection. Boards move down one table as normally.

What about an odd number of tables, greater than 13? Splitting into two sections of Mitchells leaves differing table counts in the two sections, violating our primary goal. The Web movement will also work here, although there is one negative. Either 3 sets of boards are needed, or some of the tables (#14 and higher) will be sharing boards on some rounds. This is not as bad as it sounds though, we've found that giving each North player (from Table 14 up) a slip telling him which boards will be shared with which table on which rounds eliminates the confusion. We print the slips just before the game starts (after we have the table count). The odd numbered Web movements are located in the "External" movements in ACBLScore, for some reason ACBLScore insists on an even number of tables when selecting "Web" in setting up the game.

With an odd number of pairs, choose either a bump or a sit-out to get to an even number of tables, thus avoiding the board sharing. For example, with 18 1/2 tables, choose an 18 table Web with a bump (again, in "External"). With 17 1/2 tables, choose an 18 table Web with a sit-out.

There is another solution for an odd table count from 15-25, which does not require board sharing. The idea is to play two sections with different movements, combining the scores across sections. For instance, with 23 tables, if it is desired to avoid the board sharing in a Web (which gets worse as the odd table count increases), a 13 table Mitchell could be run in one section, and a 10 table Howell in the other section. Alternatively, an 11 table Howell and a 12 table Howell could be run.

26 tables could run 2 sections of 13 table Mitchells, or a 26 table Web (the difference being that the E/W pairs cross section boundaries in the Web). 27 Tables and up require 3 sets of boards, but the extension is obvious. For example, with 29 tables, run 2 sections, a 13 table Mitchell, and a 16 table Web. With 32 tables, running two 16 table Web sections would require four board sets, but a 13 table Mitchell and a 19 table Web could get by with 3 sets. Of course, if you're getting that many tables, you're getting awfully close to needing to deal four board sets anyway ...

So, to summarize, we play Howell movements with 2 to 12 tables (except 9, for which we play Mitchell), Mitchell with 13 tables, and Web with even table counts of 14-24 tables. We're still deciding which way to go with odd table counts of 15-25, but we have run the Web, and not had any difficulty.

I hope this helps. Actually, I don't really care whether it helps - no, just kidding.

52 Comments

.

OR

Benefits include:

- Create and share convention cards
- Create/answer bridge problems
- Connect with other bridge players
- Write and comment on articles

Plus... it's free!