Movement

[mink]

Uday told us that in the first implementation the movement is simply that NS pairs stay at their table and EW pairs go up 1 table after each round. This has a drawback: an NS pair never gets compared to an EW pair, and one could say that the NS and EW pairs play different tournaments. Often, if this kind of movement is chosen, there are 2 rankings given, and there are 2 winners.

 

As I dislike the idea of having 2 winners, there should be some scrambling in a way that a pair can be compared to most ohter pairs of the tourney with more or less boards. The usual method to do this is to switch the direction NS/EW orientation of the pairs after some predefined round, but the moving pairs keep moving. But this does not work as desired in an online tourney, as the boards do not move from table to table but each board is played at all tables simultaneously.

 

Here the following movement could be applied to achieve the scrambling. Assume we have 5 rounds and 5 tables. Pair A sits always at table 1 NS. Pair B sits at table 2 NS in round 1, C at 3 NS, D at 4 NS, E at 5 NS , F at 5 EW, G at 4 EW, ... and J at 1 EW. After each round NS pairs at tables 1-4 move 1 table up, EW pairs at tables 2-5 move 1 table down, the NS pair at table 5 changes direction and becomes EW at 5, and the EW pair at table 1 moves to 2 NS.

 

This movement achieves that all pairs play NS and EW for a different number of times. In that way any pair compares to all other pairs but one, but for a different number of boards.

 

If there more rounds than tables, this movement works for e.g. up to 9 rounds in case of 5 tables - in this case each pair plays against all others. If there are much more tables than rounds, sections of tables could be created, and the movement operates in each section independently of the other sections.

 

Of course in the latter case a swiss movement would be more desireable, but it is also much harder to implement. The movement described here should not be only a little change to the current movement.

 

Karl

[McBruce]

Arrow switches are a response to the oft-seen club situation where most of the strong pairs and few of the weak ones sit NS, resulting in a skewed game when there are two winners, especially when some of the EW pairs miss the strongest NS pairs in an incomplete Mitchell. BBO tournaments seat players at random, so any inequity is going to be by chance.

 

However, I think there are two better solutions than Mitchell movements. One is a pure Howell. Table one's NS pair is given the highest pair number, table one EW is given the lowest pair number. Table two's NS pair is given the second

highest numbner, 2 EW the second lowest pair number, and so on. At the end of round one, you find the pair you are following (the pair with the next lower number) and proceed to that table next. Every location "points" to another location where you will play the next round. The only exception is that the highest numbered pair always remains stationary and the lowest numbered pair follows the second highest numbered pair.

 

The other solution would be to make a random movement after round one; then, for round three through to the end, match pairs based on their score following the round before last, as in Swiss teams. So in round three, the rankings after round one would be the guide to pairings, with #1 vs #2, #3 vs #4, etc. This is usually called a BAROMETER movement. It seems to me that it's more trouble than it is worth to fiddle with this to avoid having pairs play one another twice or more, especially as most BBO tournaments are going to have quite a few pairs, making repeating opponents unlikely and three rounds against the same pair extremely remote. But software could likely be set up to accomplish this as well. The reason to rank by the standings from two rounds ago is to avoid having to wait for slow pairs. There used to be a weekly four boards, four rounds tourney on the Zone which matched pairs by score, and there were incredibly long waits when players were slow.

 

Both options are easy to program. Some may argue that the baromoeter has the same weaknesses as Swiss teams: a weak pair can get back into contention by playing weak pairs. I think this is overrated: even at the top table somebody is going to get at least 50%, and if a weak pair gets two or three tops they'll find themselves in a higher class right away. The 4 x 4 Zone tourney seldom had a winner outside of the top three tables in the last round. More often the case is that a decent pair gets a chance to place after being blitzed in the first round by the best pair in the field: they start with an unlucky draw but get compensated for this disadvantage later.

 

-McBruce

[Erkson]

Mink wrote :

Here the following movement could be applied to achieve the scrambling. Assume we have 5 rounds and 5 tables.

 

Hi, Mink.

A good idea, but why assuming a situation which never exist in this BOL reality !

 

Assume you have 4 rounds of 2 boards and 70 tables !

That's the problem : you meet a short sample of the field.

 

Erkson.