SeanBothar

It doesn't sound pleasant but how do you remove friends on the new BBO version? No problem with the old Adobe Flash version.

On page 193 of Ian McKinnon's book "Duplicate Bridge Schedules, History and Mathematics" he does refer to "Mr Flower". Obviously Ian was unable to discover any further details. Assuming Mr Flower is correct then the original questions can be rephrased: Who was Flower? Where did he live? What was his first name? BBO forum members to the rescue!

Who was Flower? Where did she or he live? What is her or his first name? We know John Mitchell was originally from Glasgow and settled in Chicago working as a bank clerk. He published books on Whist and has been called the "Father of Duplicate Whist". Edwin Howell was born in Nantucket and was a mathematics lecturer in Baltimore. He also published books on his Whist movements. Details about Flower seem hard to find. The Flower Howells have been little used over the past 100 years because of the complex board movement. Because BBO uses the same boards for all tables each round the board movement problem doesn't arise. Flower would be amazed that her or his movements are now used daily on BBO by enormous number of players. Maybe a BBO history cognoscente can give us details about Flower.

This topic was raised under a previous post "BBO Mitchells - switching" In a previous post on BBO Swiss Pairs the question of switching in Mitchells was raised. BBO Mitchells are different from ordinary Mitchells because all the same boards are played at each table for each round. So for a 13T 13R 26B Mitchell on BBO boards 1 & 2 are played at all tables in round 1, boards 3 & 4 at all tables in round 2 and so on. For a 13T 13R 26B Mitchell in club "live" competitions switching is normally done on rounds 12 & 13 to give a one winner result. Many may not be aware that if the same switching was applied on R12 & R13 to a BBO 13T 13R 26B Mitchell this would have no impact at all on the balance. To understand this consider the "travelling" scoresheets for a 13T 13R 26B Mitchell switched on rounds 12 & 13. For the "live" club Mitchell, two EW pairs (#14-#26) are in the "NS" column and two NS pairs (#1-#13) are in the "EW" column on each traveller. This is what gives the comparison between NS pairs and EW pairs and gives a fair one winner result. Now look at boards 23 to 26 on the BBO Mitchell if switching was applied on R12 & R13. All the EWs (#14-#26) are in the "NS" column and all the "NS" (#1-#13) are in the "EW" column so there is no comparison at all between the NS pairs and the EW pairs. That is why switching on R12 & R13 has no impact at all on the balance with BBO Mitchells. If switching in needed on BBO Mitchells there has to be a different approach. Staying with the BBO 13T 13R 26B Mitchell example, switch table 1 on rounds 12 & 13, table 2 on rounds 11 & 12, table 3 on rounds 10 & 11 and so on to table 11 on rounds 2 & 3. Table 12 is best on R13 only and Table 13 on round 2. This gives an excellent balance as good as the "live" club Mitchell. Forum users will undoubtedly all agree that this is far too complex to code into the movement instructions so unswitched Mitchells have to remain the only option!

In a complete or full Howell all pairs play against each other. So for a 6 table complete Howell each pair plays against 11 others and if 2 boards are played a round then 22 boards in total (6T 11R 22B). To balance this movement consider pair A who played 2 boards against pair B. For the other 10 rounds when they didn't play against each other A & B should be on the same side of the travelling scoresheet 10 times and on the opposite side 10 times for 20 boards. If this applies to all the other pairs then the balance is 100%. As mentioned for BBO Flower Howells all tables play the same boards for each round. This is also used in BBO Mitchells (see previous post BBO Mitchells - switching). Balancing a complete or full BBO Flower Howell is much easier than a BBO Mitchell. Usually a whole table is switched for all the rounds. For example if table 3 is switched then the moving NS pair who arrive here from table 2 play as EW and the moving EW pair from table 4 play as NS. Listed below are the balance percentages before and after switching for complete (full) 3 to 8 tables BBO Flower Howells. 3T 5R no improvement with switching 60% unswitched 4T 7R switch table 4 100% compared to 53% unswitched 5T 9R switch table 5 83% compared to 50% unswitched 6T 11R switch table 3 100% compared to 48% unswitched 7T 13R switch table 5 91% compared to 47% unswitched 8T 15R switch 5,6,7 & 8 100% compared to 60% unswitched Whether it is possible to apply this changes to the BBO coding is another matter!

Many Bridge clubs are now having to play their weekly games online and are interested in the type of Howell used by BBO. Most of our Bridge movements are derived from those used in the United States for duplicate Whist in the 1890s onwards. (John) Mitchell, (Edwin) Howell and American Whist are familiar terms. Less familiar is Flower who devised a Howell type movement where pairs move regularly but the board movement is complex. BBO uses Flower Howells and the board movement problem doesn't arise as the same boards are played at all the tables each round. In a BBO Flower movement, Pair 1 stay at table 1 as NS throughout. The other NSs move up a table and the EWs move down a table after each round. When the NSs finish playing at the last table they switch to EW at this table. When the EWs finish playing at table 1 they move to table 2 as NSs. This applies to all BBO Flower Howell movements up to 20 tables (the current maximum permitted).

In a previous post on BBO Swiss Pairs the question of switching in Mitchells was raised. BBO Mitchells are different from ordinary Mitchells because all the same boards are played at each table for each round. So for a 13T 13R 26B Mitchell on BBO boards 1 & 2 are played at all tables in round 1, boards 3 & 4 at all tables in round 2 and so on. For a 13T 13R 26B Mitchell in club "live" competitions switching is normally done on rounds 12 & 13 to give a one winner result. Many may not be aware that if the same switching was applied on R12 & R13 to a BBO 13T 13R 26B Mitchell this would have no impact at all on the balance. To understand this consider the "travelling" scoresheets for a 13T 13R 26B Mitchell switched on rounds 12 & 13. For the "live" club Mitchell, two EW pairs (#14-#26) are in the "NS" column and two NS pairs (#1-#13) are in the "EW" column on each traveller. This is what gives the comparison between NS pairs and EW pairs and gives a fair one winner result. Now look at boards 23 to 26 on the BBO Mitchell if switching was applied on R12 & R13. All the EWs (#14-#26) are in the "NS" column and all the "NS" (#1-#13) are in the "EW" column so there is no comparison at all between the NS pairs and the EW pairs. That is why switching on R12 & R13 has no impact at all on the balance with BBO Mitchells. If switching in needed on BBO Mitchells there has to be a different approach. Staying with the BBO 13T 13R 26B Mitchell example, switch table 1 on rounds 12 & 13, table 2 on rounds 11 & 12, table 3 on rounds 10 & 11 and so on to table 11 on rounds 2 & 3. Table 12 is best on R13 only and Table 13 on round 2. This gives an excellent balance as good as the "live" club Mitchell. Forum users will undoubtedly all agree that this is far too complex to code into the movement instructions so unswitched Mitchells have to remain the only option!

Thank you. A pertinent fact omitted from the original post about this Swiss Mixed Pairs competition was that it was a "paying" one. If it had been "non-paying" apparently many more tables would be allowed in one section which would have reduced the chances of pairs meeting twice or more. Is there a code which can be inserted into a "paying" competition Swiss Pairs format which allows large sections as used in "non-paying" ones? In a "paying" Mitchell what is the new upper limit for table numbers per section and can switching on certain rounds be specified? Finally, is the upper limit for Howells still 20 tables?

A mixed pairs event last night 5th June 2020 (7117-1591382583) was run as Swiss Pairs. There were 57 tables and BBO split this into 3 x 14T plus 15T. Some players complained that they had to play against the same pair for 3 or more rounds. With only 28 pairs in most of the sections this seems inevitable if "Danish" is the default option. Is there any code to insert in the format set-up which prevents splitting (+????+)? Also if the competition had been run as a Mitchell can splitting be prevented? Finally is there any code which allows switching on specified rounds in Mitchells?