# How should players be seeded at an elimination tournament

0 views

### Elisabeth Scheicher

Dec 2, 1998, 3:00:00 AM12/2/98
to
How should players be seeded at an elimination tournament?

Suppose we have eight players ranked from 1 to 8. In my mind there
are two plausible ways to seed them:

A:
B:

1-+ 1-+
+---+ +---+
5-+ | 8-+ |
+---+ +---+
3-+ | | 4-+ | |
+---+ | +---+ |
7-+ | 5-+ |
+---- +----
2-+ | 2-+ |
+---+ | +---+ |
6-+ | | 7-+ | |
+---+ +---+
4-+ | 3-+ |
+---+ +---+
8-+ 6-+

In (A) the difference of ranks is equal: in the first round
it's 4 = 5-1 = 7-3 = 6-2 = 8-4; if any player defeats the higher ranked
player then in the 2nd round it's 2 = 3-1 = 4-2 and in the final it's
1 = 2-1. The pairing in the first round is the same as in the Swiss
tournament system used in chess.

In (B) the sum of the ranks is equal: in the first round it's 9,
in the 2nd round it's 5 and in the final it's 3.

I think (A) should be preferred since ANY seeding favors the good and
(B)
goes a little bit too far in my mind.

Which of these two possibilities is used at backgammon (or e.g. tennis
tourneys)?

### David desJardins

Dec 2, 1998, 3:00:00 AM12/2/98
to
Elisabeth Scheicher <sche...@xss.co.at> writes:
> How should players be seeded at an elimination tournament?

Generally you want the top player to play the bottom player, and so on.
The reason is that, otherwise, it can be advantageous to underreport
your strength in order to get a lower seed and an easier opponent.

I find it surprising that chess tournaments work any other way. But
they do have the advantage of a reliable way to generate seedings, which
can't easily be manipulated (the rating system).

David desJardins

### Donald Kahn

Dec 2, 1998, 3:00:00 AM12/2/98
to
On 02 Dec 1998 09:24:17 -0800, David desJardins
<de...@math.berkeley.edu> wrote:

In "live" tournaments, there is never anything other than a random
draw. Anything else would be simply a cause for controversy.

deekay

### Chuck Bower

Dec 2, 1998, 3:00:00 AM12/2/98
to
In article <36656E33...@xss.co.at>,
Elisabeth Scheicher <sche...@xss.co.at> wrote:

>How should players be seeded at an elimination tournament?
>

I believe the second one is common in sporting events (like tennis
and basketball). But we're talking BG!

Back in the 70's I remember backgammon touraments being seeded. Of
the dozen or so tourneys I've attended since returning to the game in
1992, NONE has been seeded. All pairings have been RANDOM DRAW. Why?

I believe the answer is simple. At backgammon you don't really
know who the best players are, because of the luck factor. Sure, you
can say "A, B, and C have had the best tournament success. They're
among the best". But which is #1? And you have to seed an entire field!
Also, if byes go according to seeds, the unfairness can be magnified.

Backgammon tournaments tend to attract a range of players. Some
believe in non-random luck; some don't. Almost everyone has a higher opionion
of his/herself than is warranted. Basically that is human nature but
the "dice factor" certainly magnifies the effect. I believe that seeding
in US tournaments fell out of favor because of these nebulosities (and
a lot of hurt feelings, and beliefs that injustices had occurred). I
think the promoters felt that attendance would be better if there was
an attempt to treat players equally rather than giving some an advantage
based at least partially on subjective factors.

I'm not saying you SHOULDN'T seed. "Objective" eeding could be used in
a year-end tournament as a reward for annual attendance at weekly games,
for example. (Rather than rewarding based simply upon attendance, better
might be to seed based on club points which are awarded for matches won
during the year, or for coming in 1st/2nd at the weekly game.) You could
seed based on ratings (e.g. ratings on a server) but this has a downside
as well. But now I'm returning to a oft debated topic in this newsgroup,
so I think I'll just leave it at this.

Chuck
bo...@bigbang.astro.indiana.edu
c_ray on FIBS

### Robert-Jan Veldhuizen

Dec 2, 1998, 3:00:00 AM12/2/98
to
On 02-dec-98 17:43:31, Elisabeth Scheicher wrote:

ES> How should players be seeded at an elimination tournament?

That, of course, depends on your goal(s).

ES> Suppose we have eight players ranked from 1 to 8. In my mind there
ES> are two plausible ways to seed them:

ES> A:
ES> B:

ES> 1-+ 1-+
ES> +---+ +---+
ES> 5-+ | 8-+ |
ES> +---+ +---+
ES> 3-+ | | 4-+ | |
ES> +---+ | +---+ |
ES> 7-+ | 5-+ |
ES> +---- +----
ES> 2-+ | 2-+ |
ES> +---+ | +---+ |
ES> 6-+ | | 7-+ | |
ES> +---+ +---+
ES> 4-+ | 3-+ |
ES> +---+ +---+
ES> 8-+ 6-+

[...]

ES> I think (A) should be preferred since ANY seeding favors the good and
ES> (B)
ES> goes a little bit too far in my mind.

ES> Which of these two possibilities is used at backgammon (or e.g. tennis
ES> tourneys)?

Well, if you organise a big commercial tennis tournament, you usually
prefer to have the best players make it to the final, because these
players often attract publicity and (thus) money. So in tennis, often
the second scheme is used: it's more likely that #1 and #2 will meet and
you'll get a "dream final".

If considerations like these don't bother you at all, the first scheme
seems more appropriate to give the various players a more equal chance
of winning the tourney.

--
Zorba/Robert-Jan