[blml] Swiss Teams formats [SEC=UNOFFICIAL]

[Herman De Wael]

richard.hills at immi.gov.au wrote:
> Herman De Wael:
> 
>> Forget it. The players would shout "why do they get 10 VP?". Not
>> needed, I believe, if:
> 
> [snip]
> 
>> So we could make matches A1-A2 B1-C2 C1-B2 D1-D2 immediately,
>> with the constraint that if a B wins over an A, he takes their
>> place?
> 
> Richard Hills:
> 
> Yes, the temporary 10 VP carry-forward is merely an aid to the
> director for the special Round Two pairings - it need not be
> published for the players to shout about.
> 
> Of course, some of the players - those in groups B and C - will
> shout about being paired against a Round Two opponent whose VPs
> differ by ten from their own team's VPs.
> 
> Quick history of chess Accelerated Swisses ->
> 
> Due to chess having a very compressed "victory point" scale, 1
> point for a win, 1/2 point for a draw, 0 point for a loss, in
> some big chess Swisses more than one player gained a perfect
> score, five wins out of five or six wins out of six.
> 
> So the idea behind a chess Accelerated Swiss was to make sure
> that the top players met each other as quickly as possible,
> since once two top players were paired against each other it
> was logically impossible for them both to score 100% in the
> event.
> 
> If one makes the simplifying assumption that the Elo chess
> rating scheme was 100% accurate and that a higher rated chess
> player beats a lower rated chess player 100% of the time, then
> a five round simple Swiss for 64 chess players does not
> guarantee that a player with a perfect score would be the sole
> winner.
> 
> But after two rounds of Accelerated Swiss, if the simplifying
> assumptions hold, only the top eight seeds have a perfect score
> of two wins out of two (instead of the normal expectation of 16
> players with two wins in a non-Accelerated Swiss), so the
> chances of a person finishing with 5 wins out of five being the
> sole winner is somewhat improved.
> 
> 

Thanks for this explanation.

It is now clear however, that the same does not hold true for bridge. 
In a chess swiss, persistant winners will eventually meet, unless the 
number of players is larger than 2^n, which is indeed often the case. 
In a bridge swiss, persistent winners need never meet, because they 
can be overhauled by a one-time loser. So the same principle does not 
hold.

I have once read that studies had shown that the first round seeding 
in a (bridge) swiss has no specific influence on the outcome, so this 
accelerated idea would only have some influence because of the way the 
second round is paired up. I don't think I'll bother, then.

-- 
Herman DE WAEL
Antwerpen Belgium
http://www.hdw.be