May 2007 - 7

Letters to the Editor

Ten-point scale
To the Editor:
It seems clear from Bobby Wolff 's
article in the January Bulletin (page
36), "The Thrill of Victory (Points)
..." that neither the 20-point nor the
30-point VP scale addresses the issue
that a team can be undefeated in a
Swiss teams yet fail to win.
To solve this problem, I propose
a simple adjustment to the 20-point
scale. I refer to it as the 10-point VP
scale. In this approach, the winning
(or tying) team gets 10 points, no
matter how badly it beats the other
team. The losing team gets the same
number of points it would have
received under the 20-point VP scale.
Here's the scale:
IMPs
VPs
0
10-10
1-2
10-9
3-4
10-8
5-7
10-7
8-10
10-6
11-13
10-5
14-16
10-4
17-19
10-3
20-23
10-2
24-27
10-1
28+
10-0
In this way, the highest score a
team can get in a seven-match Swiss
is 70, and the only way it can get
70 VPs is by winning or tying each
match. Every team that has lost one
or more matches will have fewer
than 70 VPs. Thus a team with a loss
cannot finish ahead of an undefeated
team.
HAROLD C. MATHEWS
Bala Cynwyd PA

Swiss gambit
To the Editor:
When the mathematics of Swiss
teams scoring favors winning your
last match more than winning your
first, the integrity of the rankings
is compromised. Players become
frustrated.
As a case in point, at a recent Sunday Swiss teams, an expert asked how
my first match had gone. I said, "We
won by 15. You?" "We got blitzed,"
he said, "Swiss gambit is on!" His
team faced the soft underbelly of the
May 2007

field for the next four matches and
was soon in second place with 78 out
of 100 victory points.
In a Swiss, an equalizing effect
always seems to occur: The stronger
your team's results, the tougher the
competition you'll face. Using VPs,
there is little ranking advantage to
being at the front of the pack (unless,
of course, it is the last match). You
might play solid bridge against a top
team to squeeze out a 3-IMP victory,
but other teams with lesser scores
(and therefore lesser opponents)
will blitz and pass you by leaps and
bounds.
Bobby Wolff created the 30-point
VP scale as an improvement to the
20-point scale to stress whether the
match was won or lost. I suggest
going one step further: Score each
match by win/loss, but also by victory points. Teams would be ranked
by their win/loss record first. Ties
would be broken by VPs.
For example, a team with a 4-0
record with one blitz and three close
wins might have a score of four wins
and 62 VPs, or 4/62 for short. They
would face one of the other teams
who also had four wins and would get
matched according to their rank by
VPs. A team that got blitzed in their
first match but scored blitzes the next
three would have a record of 3/60.
If, in match five, the leading team
tied and the trailing team blitzed, the
standings would become 4.5/72 vs.
4/80.
With straight VPs, it's almost
impossible to hold on to a lead, and
there's the Swiss gambit where you
can be "rewarded" for an early loss.
The downside with straight win/loss
is that teams are not rewarded for
their big wins.
With a combined score, players
get the best of both worlds: Win/loss
to protect the leaders, VPs to give
(secondary) credit for big wins and
marginal losses.
I hope the ACBL will consider

using it in tournaments.

BRAD BART
Burnaby BC

Better pairings?
To the Editor:
In a recent sectional stratified
Swiss teams, in which ACBLscore
presumably was used to determine
round-by-round pairings, the following unfortunate results occurred.
After the first match, the strongest
A team in the field was paired against
a C team when two other B teams
with identical scores were candidates.
Later, in the pivotal, penultimate
round, the pairings resulted in this
same A team playing another C team,
when again an alternate pairing of
the leaders (not this strong A team)
against one of two candidate teams
with identical scores, would have resulted in a more equitable pairing of
the contending teams (top 16 of 24).
The actual pairings resulted in four
A vs. A matches, two A vs. C, one
A vs. B, and one B vs. C matches.
The alternate pairings would have
been five A vs. A, one A vs. C, and
two B vs. C matches.
I understand the program matches
the top teams that haven't played
each other on down the line, but I am
unclear as to how matches are made
when there are multiple possibilities
with identical scores. Is it random,
by team number, or by some other
optimized means? If by some other
optimized means, I suggest it should
be optimized to select the most stratcompliant alternative result.
For example, when you have three
or more teams with identical scores
to pair against each other, you never
pair an A with a C when it can be
avoided. For another example, when
pairing a higher score against two or
more same-score candidates, you examine each result and select the one
that minimizes A vs. C pairings.
With computerized pairings, this is 

Letters to the editor are welcome by regular mail, e-mail editor@acbl.org or fax
901-398-7754. Brevity is considered a virtue, as is subject matter with relevance
to a majority of ACBL members. Unsigned letters are not considered for publication, so please include your name and hometown. Letters may be edited.
All letters will receive a response.

Table of Contents for the Digital Edition of May 2007