[uday]

I have been testing a new movement . Characteristics of this movement:

a) Clocked
Swiss , with playbacks allowed
c) Pairs only (for now)
d) Host can specify a percentage of players to be eliminated at the start of each round (starting with round 3). If specified, lowest scoring pairs are removed at the start of each round
e) Barometer ( pairs receive results at the end of each round, with IMPs/MPs)
f) A pair that contains a missing player when a round changes is eliminated, along with a low-scoring pair, if need be, to avoid half tables
g) If an odd number of pairs registers, the last pair to register is discarded on tourney startup to avoid half tables


I use one board per round with this movement. Generally, feedback has been positive.

Any Tourney host will be able to create this type of tourney once the 3.5.2 client shows up (week or two or three).

I hope to use this format in conjunction with the new client to come closer to a completely automated T.

[inquiry]

Cool changes (this post and the other one)... I look forward to the substitute issue, the survivor game, the barometer games, and most importantly, fewer playbacks.

Thanks Uday,

Ben

[uday]

Just to clarify a small point; we will not see a pure Barometer movement just yet; Swiss and Survivor imply Barometer, though. Maybe i can make all Ts barometer by default tho - have to think about that one.

[Trpltrbl]

Hi Uday, What kinda movements do y use on r tourneys now ? Any chance in future that Directors can decide for themselfs, what kinda movement they wanna use with certain tournaments ?

Thx Mike ;D

[uday]

Not sure what you mean by choice of movement. TDs can currentty choose (for pair events) between simple Clocked/Unclocked /Swiss/ Survivor, each with slightly different characteristics

[Trpltrbl]

Quote

Not sure what you mean by choice of movement. TDs can currentty choose (for pair events) between simple Clocked/Unclocked /Swiss/ Survivor, each with slightly different characteristics

With movements I mean ones like where N/S stay always at same table, or where both pairs move, in a specific way instaed of random.

Mike ;D

[hrothgar]

With movements I mean ones like where N/S stay always at same table, or where both pairs move, in a specific way instead of random.


Given that online bridge is a "virtual" playing environment, it makes very little sense to introduce concepts such as "stationary pairs".

From my perspective, discussions about online bridge movements need to focus on two distinct dimensions.

The first dimension is the “stability” of the tournament. For better or worse, online bridge suffers from an embarrassing number of participants who fail to complete the tournaments that they enter. In some case players suffer from connection difficulties. There are also numerous examples of players who disconnect when they experience poor results. As a result, it is often necessary to design tournament movements that use “graceful” failure models to automatically compensate for disconnections.

The second dimension is the “fairness” of the movement. It is desirable that bridge movements minimize the “luck” in a tournament by ensuring that players are matched against the same set of pairs. For example, playing a Mitchell movement N/S pairs remain stationary, while E/W pairs change tables. Ideally, every E/W pair will compete against every N/S pair. This movement ensures that pairs are not “punished” by missing their expected tops playing against an extremely weak pair.

The challenge of, course, is that these two design goals are mutually antagonistic. Highly stable movements are extremely unfair. Fair movements are extremely fragile.

The most “stable” format is probably to combine all of the participants into a single large pool and then match randomly. Unfortunately, produces a highly unfair movement in which tournament results are very much biased by the luck of the draw. Equivalently, a fully meshed Howell movement is probably the most fair, however, this system will encounter significant problems if more than one pair drops out.

I’m not sure of the “best” way to balance these two conflicting design goals. However, I would expect that tournament organizers would find it useful to be able to use a fully meshed Howell for “special” events when you trust that players won’t deliberately disconnect.

[mghatiya]

Hi,

I was directed to this thread from http://www.bridgebas...897#entry542897

So as I mentioned in that thread, I have a concern on the logic of removing a pair who has been doing well, but due to the flaky connection one of them is unavailable only for the moment when the round is changing. I think it is very unfair.

Survivor movement does have reds and subs when the round is going on. Then why the weird treatment when a person is not present at the time of round change?

We all know there are plenty of frequent disconnections. If we start penalizing a pair like this for something which is not in their control, I am sure it would be a great step towards reducing Bridge popularity.

I understand from the above mentioned point that this happens only when the number of pairs becomes odd, and hence to avoid half table. But I think if you call it survivor, you should remove the pair with second least score. Unless of course you meant survivor in the sense of "the one who survived internet disconnections"

And btw when we are at it, I think the cutting off the bottom pairs should happen after at least two boards. Removing after just one bad board which very often can happen due to bad luck, sounds very cruel.

All in good spirit please

Thanks,
Mukesh

[0 carbon]

I have set the Survivor Cut % to 24%, Logically, this should leave 4 pairs at the end, 2 tables, since with 5+ pairs 1+ will be eliminated.
Instead, I have 7 tables left - 28 pairs for the last 8 rounds of the tourney, with no one cut.

Howcome?

tOM