Movements
Pete
1) over three sessions of +- 3 to 4 hours each session
2) with a fresh set of boards for each session
3) where each player plays against each other player once
4) 26 pairs entered (13 tables)
What would you suggest?
Peter Bircher
Chairman
Durban Duplicate Bridge Club
031 266 9422
083 233 1628
Dave Flower
I'm not sure that there is an elegant solution to your problem; I would
add a cautionary tale from the EBU, who often played the following,
deeply flawed movement for 24 pairs.
A) 12 table Mitchell
then
B) Two Howells, one for the North/Souths and one from the East/Wests
from the Mitchell.
This did have the logistic advantage that only one set of boards was
required for the two Howells.
But it had the overwhelming disadvantage that the field was split into
two completely separate sub-fields, with no valid comparison between
them. (If you doubt this, assume that all the N/S players in the
Mitchell were world-class experts, and 11 E/W pairs were palookas, one
a competant club pair.
So add an additional requirement that the movement be balanced
Dave Flower
Gordon Rainsford
> Our club wishes to play the annual championships:
> 1) over three sessions of +- 3 to 4 hours each session
> 2) with a fresh set of boards for each session
> 3) where each player plays against each other player once
> 4) 26 pairs entered (13 tables)
>
> What would you suggest?
The numbers are awkward to do exactly what you want.
You could do a 2-session all-play-all easily enough (or a 5-session!),
or you could do a 3-session for 14 or 15 tables. Do you have any
flexibility with the format?
--
Gordon Rainsford
London UK
Wayne
This will work for 26 pairs and will nearly achieve your three session
requirement with a bit of a fiddle.
Divide the field into five sections of five pairs with one pair left
over.
Play two Mitchells (5 tables) and one Howell (3 tables).
Mitchell 1 Mitchell 2 Howell
1 v 2 3 v 5 4
2 v 3 4 v 1 5
3 v 4 5 v 2 1
4 v 5 1 v 3 2
5 v 1 2 v 4 3
The single pair sit continuously in the Howell.
If you play three boards per round then by combining two of these
'sessions' you can play two thirty board sessions (around 3 hr 45) and
one fifteen board session (around 1 hr 55)
Wayne
Sarah Oliver
"Pete" <jbir...@iafrica.com> wrote in message
news:e9ff9r$32m$1...@ctb-nnrp2.saix.net...
Lets answer the question you didnt ask
14 tables 28 pairs is much simpler
Divide into 3 nines and I stationary pair
Groups A B and C
Week 1 AvB 9 table Mitchell C+ staionary pair play 5 table Howell
(if you have two sets of boards fine otherwise positioning tables carefully
will allow board share between Mitchell and Howell)
rotate the groups ABC in weeks 2 and 3 to achieve all play all
Im no expert but probably 3 round sof arrowswitch is needed in the MItchells
in all play movements
If you wnt to play with your 13 tables then a complete barometer (starting
positions stolen from Movements p224 by our dear friend Hans-Olaf Hellen et
all)
26v1
25v2
24v3
4v23
5v22
21v6
7v20
19v8
9v18
10v17
11v16
12v15
13v14
Pair 26 is stationary pairs 1-25 follow each other round in sequence moving
up or down 1 table each round
You need 25 rounds so 8 8 9 would do it, you probably would settle for 3
board rounds so you need about 7 sets of boards if each table shares 3
boards with its neighbour
This in my opinion is the best solution of your problem if you have the
resources to duplimate sets of boards - perhaps a little more than one hours
work if you have an experienced machine operator
Wayne's solution -fiive subgroups is practical but inelegant because it
really is 2 1/2 sessions and also needs arrowswitches and two sets of boards
at least
If you were really desperate to have 26 pairs you might consider the 28 pair
solution with a couple of missing pairs so everyone sat out twice
Mike Amos
Pete
Not sure about a 12 table Mitchell as I have 26 pairs
Am I missing something?
Peter
"Dave Flower" <DavJF...@AOL.COM> wrote in message
news:1153124543.6...@m79g2000cwm.googlegroups.com...
Pete
Yes we have agreed on some flexability
I have at the moment a 5 table Howell 8 rounds 3 boards a round (2 seated
pairs)
and a 8 table Mitchell (sharing and relay)
The 3 moving groups of eight move into a different position for each session
The only drawback is that there are 2 pairs that don't play each other.
This is not a train smash and acceptable by the committee.
I have had an offer from a Peter Smulders:
"I've got a Howell for 26 pairs in 3 sessions here. The 3 sessions have 8 +
8 + 9 rounds, so with 3 boards per round the required time would be just
right. Each session has its own board sets."
He has offered to email details of the movement and if he does come right it
might be ideal.
Peter
"Gordon Rainsford" <grbridgeR...@btinternet.com> wrote in message
news:1him82w.1po8kpykfgtk8N%grbridgeR...@btinternet.com...
Pete
However, the disparity in session time would prove a major drawback so I
think I will go with an 8-table mitchell (sharing and relay table) and 1x
5-table Howell with 8 rounds and two stationary pairs.
As mentioned in one of the other resonses it has the 'minor' drawback that 2
of the pairs will not meet. This is regarded as not too serious.
Peter
"Wayne" <wjbu...@gmail.com> wrote in message
news:1153132069.4...@75g2000cwc.googlegroups.com...
Pete
Yes, 14 tables would have been a breeze - no sharing, all paly each other in
27 boards a session
And yes, we did consider two 'ghosts' but felt that the lessor of two evils
was to have 2 pairs failing to meet.
This will happen if I play a 5 table Howell 8 rounds 3 boards a round (2
seated
pairs) and a 8 table Mitchell (sharing and relay) with 24 boards in play.
The 3 moving groups of eight move into a different position for each session
This 'solution' appears best at the moment
However I am holding thumbs for something from a Peter Smulders who has 'got
a Howell for 26 pairs in 3 sessions'
This sounds perfect
Peter
"Sarah Oliver" <sara...@onetel.net> wrote in message
news:44bb...@212.67.96.135...
Hank Youngerman
sure exactly who to contact there, to be honest. If you can reach
Solly Weinstein, he might well know.
Your numbers don't seem that bad. 25 pairings could be sessions of
24, 24, and 27 boards.
You could obviously play a 13-table Howell, in either of the two
not-entirely-satisfactory ways:
a) One board a round, repeated three times
b) Much more elegant, play 9 rounds the first session, 8 each of the
next two, with a separate set of boards for each table. This would
require duplicating a total of 975 boards (75 boards for each of 13
tables). If you know anyone with a duplicating machine, it's
obviously much easier.
But someone at ACBL may have an even more elegant solution.
Regarding another post on this thread, I have seen the two-session
movement done just that way, i.e. a Mitchell and a Howell, when there
are about 18-26 pairs in a full-day contest. I think (but am not
sure) that this is how the finals of the North American Pairs is done.
However, (a) by that point the field has been cut several times so
that it is fairly strong (relative to the original entry - it is
flighted, so the non-Life Master flight will be fairly weak, but they
will all be the cream of that group) and (b) in the higher flights,
where the players have some reputation to rely on, the field is
seeded.
On Mon, 17 Jul 2006 09:47:11 +0200, "Pete" <jbir...@iafrica.com>
wrote:
Michael Angelo Ravera
Pete wrote:
> Our club wishes to play the annual championships:
> 1) over three sessions of +- 3 to 4 hours each session
> 2) with a fresh set of boards for each session
> 3) where each player plays against each other player once
> 4) 26 pairs entered (13 tables)
>
> What would you suggest?
ACBL score can generate a 25 round Howell movement. It wouldn't let me
use 75 boards, but it would let me do 50. Extension to 75 boards is
obvious. This would give everyone a chance to play the same number of
boards against everyone else. This has the advantage of only requiring
75 boards in play. The board movement is extremely orderly (boards come
in at table 13 and get put on bye at table 1. When boards 37-39 come
out of table one after the first session, they are retired from play
(and could theoretically be discussed).
Effectively, you can just create a 13-table full Howell and play 75
boards in 25 rounds. You can, for convenience, put session breaks in at
appropriate times. It would seem obvious to use break after 8 or 9 and
16 or 17 rounds. You do have the problem that is is possible to discuss
hands at the breaks (despite its being illegal), however.
If you want to do something where boards are retired after each
session, you could interweave five groups of 5 plus an odd pair and use
two 5-table Mitchell's and a 3-table Howell each with 5-board rounds.
This would give you three sessions each of exactly 25 boards.
Basically, I would take the odd N-S pairs from group 1 in the first
round and put them with the even E-W pairs from group 2 in the first
round in the Howell in the second round. You can then move the other
pairs around in a similar fashion. You obviously can't get everyone to
play everyone in 75 boards with 5-board rounds, but everyone plays the
same boads in each round. For this, you have to either make up three
copies of each board in each round or work out how to share them
properly. It isn't so very hard to share 5 boards amongst 3 tables (and
two sets will be on bye in the Howell, so you don't even really need a
caddy except in the last round where 5 boards need to be shared amongst
5 tables).
Wayne
Another possibility:
One Session : 13 table Mitchell 2 boards per table
One Session: Two Seven Table Howells with two sitouts - one in each
section - 2 boards per table
One Session: 13 table Mitchell 1 board per table followed by two Seven
table Howells 1 board per table
Wayne
One Session
David Stevenson
This does seem to me to be complete baloney.
If I understand your objection correctly, you have the following
scenario:
Each pair plays against every other pair.
Each pair compares with half the other pairs on each board.
The pairs that each pair compares with are different in each set so
provides a fair comparison overall.
That is a balanced movement!
--
David Stevenson Bridge RTFLB Cats Railways /\ /\
Liverpool, England, UK Fax: +44 870 055 7697 @ @
<bri...@blakjak.com> ICQ 20039682 bluejak on OKB =( + )=
Bridgepage: http://blakjak.com/brg_menu.htm ~
Dave Flower
In the Mitchell, pair 1 compares with pairs 2-12.
In the Howell, pair 1 compares with 5 of the pairs 1-12.
At no time does pair 1 compare with any of the pairs 13-24; in fact
none of pairs 1-12 ever compares with any of the pairs 13-24.
That is not a balanced movement
Dave Flower
Peter Smulders
news:e9g4g0$2jt$1...@ctb-nnrp2.saix.net:
> I have had an offer from a Peter Smulders:
> "I've got a Howell for 26 pairs in 3 sessions here. The 3 sessions have
> 8 + 8 + 9 rounds, so with 3 boards per round the required time would be
> just right. Each session has its own board sets."
>
> He has offered to email details of the movement and if he does come
> right it might be ideal.
Yep, that's me. I found this movement together with other Howells in
multiple sessions, for 8 to 28 pairs, in "Groot Schemaboek 2002" by Frans
Schiereck, and optimized it a bit for better balance.
David Stevenson
Codswallop.
Of course you have an arrow-switch in the Mitchell, so in the Mitchell
on every board pair 1 is comparing with some of pairs 13-24.
In the Howell, pair 1 compares with a different five of 2-12 on every
board, and also with the equivalent pairs in 13-24 on every board.
You are not suggesting scoring the two Howells separately, I presume?
Dave Flower
An arrow switch would help, but when the EBU used the movement, there
was no arrow switch.
I seriously doubt whether an arrow-switch would balance the movement.
>
> In the Howell, pair 1 compares with a different five of 2-12 on every
> board, and also with the equivalent pairs in 13-24 on every board.
So if pairs 1-12 are experts, and pairs 13-24 palookas, the average
score of the palookas is the same as the average of the experts in the
Howell.
>
> You are not suggesting scoring the two Howells separately, I presume?
It doesn't matter whether you score the two Howells together or
separately, the average score of the pairs in each Howell is 50%. (If
N/S on any board get 60%, E/W get 40%)
Dave Flower
'An' arrow switch' in the Mitchell is woefully inadequate.
I have done a simulation to test this out. Under the simplistic
assumption that I can assign each pair a number corresponding to their
ability, and match-pointing the differences between abilities.
Assuming 12 tables, 24 boards (with a relay), then switching rounds 3,
5, 9, 10, 11 performs reasonaly well. Numbering N/S 1 to 12 and E/W 13
to 24, and assigning an ability equal to the pair number, pairs 13-24
all outscore pairs 1-12.
However, if only one round is switched, five E/W pairs appear in the
bottom twelve.
Dave Flower
David Stevenson
>David Stevenson wrote:
>> Dave Flower wrote
>> >> This does seem to me to be complete baloney.
Oh, stop making up fatuous objections.
Whatever movement you use if you put all the good players in one line,
and all the poor players in another, it will not work terribly well. In
practice it never happens, and serious events are seeded to make sure it
does not.
Dave Flower
You miss the point. In a properly balanced movement, where one sits
should have no effect.
By illustrating what happens when, as an extreme example, all the good
pairs sit E/W, I have demonstrated the movement to be unbalanced.
In practice, I accept that on the law of averages, the better players
will be seated at random. However, on any given day, this does not mean
that the average E/W ability will be exactly equal to the average N/S
ability, as there are always fluctuations about the average.
There is, in factm a clear bias in playing in the direction of the
weaker players when arrow-switching only one round out of 12.
If you are not prepared to declare one winner in a 12 table Mitchell
with no arrow switch, you cannot justify declaring a single winner by
arrow-switching one round.
Dave Flower
David Stevenson
>If you are not prepared to declare one winner in a 12 table Mitchell
>with no arrow switch, you cannot justify declaring a single winner by
>arrow-switching one round.
Why not? If you arrow-switch no rounds you have no comparison, if you
arrow-switch one round you get reasonable comparisons.
Suppose you have a league where you play two-thirds of the other
teams, then declare a winner. Perfectly fair? No? Better than having
a league where you play none of the other teams? Of course.
Anyway, why not arrow-switch two rounds as recommended?
I know you are trying to prove that some pretty reasonable movement is
anything but, but it does not work, you know. It's a perfectly
reasonable movement, and in fact that is all that is necessary.
Dave Flower
To take you up on a league where you only play two-thirds of the other
teams, and still declare a winner. Surely the analogy is one where you
play a third of the teams once, an third twice, and a third not at all.
It is hardly more difficult to arrow-switch several rounds than to
arrow-switch one round, but I have never seen it happen in the EBU,
where directors seem blissfully unaware of the resulting advantages.
I have obtained the suggested arrow-switches from my copy of the
Encyclopedia og Bridge (1971 edition), under 'Scrambled Mitchell',
where they are attributed to Lawrence Rosler in 1964. Does anyone out
there know what is in more recent editions ?
Dave Flower
Peter Smulders
@i3g2000cwc.googlegroups.com:
> 'An' arrow switch' in the Mitchell is woefully inadequate.
>
> I have done a simulation to test this out. Under the simplistic
> assumption that I can assign each pair a number corresponding to their
> ability, and match-pointing the differences between abilities.
>
> Assuming 12 tables, 24 boards (with a relay), then switching rounds 3,
> 5, 9, 10, 11 performs reasonaly well. Numbering N/S 1 to 12 and E/W 13
> to 24, and assigning an ability equal to the pair number, pairs 13-24
> all outscore pairs 1-12.
>
> However, if only one round is switched, five E/W pairs appear in the
> bottom twelve.
This only proves that if you have 24 pairs with well-defined abilities as
described, and you assign them pair numbers in the order of their ability,
then switching 3 rounds would be better then 1. I think a more useful
simulation would be to assign the pair numbers randomly, and then look at
the outcome. It would be asking to much to check all 24! possibilities but
a couple of thousands random permutations might do. Then I bet you will
obtain results in accordance with the theories of Probst: the optimum
number of arrow switches is 1 in 8. In this case 1 and a half, i.e. one
full switch in one round and half the tables in another.
Sven Pran
"Peter Smulders" <de...@home.nl> wrote in message
news:Xns98067AA3248...@213.51.129.36...
Movements should be checked for balance between each combination of
pairs, that is for 276 combinations with 24 pairs. Some of these
combinations will be "mirrors" of each other.
I have yet to see an arrow-switched Mitchell giving good balance across
the entire field so I agree with Dave that an 'arrow switch' in the Mitchell
is woefully inadequate.
Arrow-switched Mitchell is of course OK if the requirement for "good"
balance across the entire field is reduced to "acceptable" balance.
Sven
David Stevenson
>To take you up on a league where you only play two-thirds of the other
>teams, and still declare a winner. Surely the analogy is one where you
>play a third of the teams once, an third twice, and a third not at all.
Not at all. In football you play against the contestants with whom
you are competing, in bridge you do not. On any individual board you
are competing against roughly half the pairs in the event, but not
playing against the ones against whom you are competing.
>It is hardly more difficult to arrow-switch several rounds than to
>arrow-switch one round, but I have never seen it happen in the EBU,
>where directors seem blissfully unaware of the resulting advantages.
Arrow-switching several rounds gets you a less fair comparison. It is
nice ot refer to "blissfully unaware" but there are no advantages in
doing so, just a major disadvantage.