[Cascade]

nickf, on Oct 20 2008, 06:35 PM, said:

I disagree that the cutoff was too low (8 teams out of 74). In fact the whole event was overswissed as it was - I dont know the exact formula but to fairly determine a winner in a Swiss the number of rounds you need is something like:

- the square root of the number of Teams plus 1

With 74 teams I suppose that means about 9 rounds is sufficient to fairly find a winner. I'd have thought that if the top 8 teams cant sort themselves out after twice that many rounds then they probably dont deserve to be in the top 8.

Either that or they were unlucky.

nickf
sydney

I suspect any formula for a swiss is based on the win/loss and maybe draw that is used in chess tournaments and does not take into account the VP scale and the inherent randomness in VP results.

The simulations I have done on swiss suggest to me that there is no such thing as overswissed. The ultimate "overswissing" would be a round-robin and almost everyone thinks that would be a fair format - that is what happened with-in groups in all of the other events at this world championship. My simulations show this in that ever additional round added value to the final rankings (or probability of the best team winning).

Swiss is my least favourite form of event. It is really just a variant of the children's game musical chairs. If you happen to be sitting in the right seat when the music stops you are in. For this reason I would always argue that from a Swiss you need to qualify more teams if you have a choice.

When the Swiss has many fewer rounds than teams then if you win the last match chances are you will finish higher than you have performed in the remainder of the event. This is because after a win you expect to get tougher (higher placed) opponents who may knock you back. Conversely if you lose the last match you will finish below your expectation.

I have encountered both experiences - being in contention throughout an event only to be hammered in the last match and finish nowhere. And to race into contention with big wins in the last match(es) - as the US U28 team nearly did.

If your big win or more often your big loss is in an earlier round you have time to recover. Many times I have noticed in the SWPT in Canberra where we have beaten another team 25-something in VPs that two rounds later after they have enjoyed easier draws and we have endured tougher draws that they have caught or nearly caught us up. Whereas losses in the last round(s) can be fatal.

[nige1]

For a specified number of teams playing a competition over a specified number of boards, I'm sure that the Swiss format produces a ranking closest to "God's ranking".

Suppose, for the sake of argument, that there are about 60 teams and there is time for each team to play about 120 boards. Among your options are:

• Round-robin of 2 board matches: A bit of a lottery.
  
• Knock-out of 20 board matches. Less of a lottery. But the 2nd best team can be eliminated in the 1st round. Stilll not a bad format with good seeding.
  
• Double-elimination. The format of the EBU Spring Fours. Like a knock-out but a team has to lose twice before it is eliminated. Better than a straight knock-out. But not as good as ...
  
• Swiss: 6 rounds of 20 board matches. A "self-seeding" compromise between a knock-out and a round-robin.

Other obvious points:

• The longer the match, the more likely the better team will win; but the luck factor in Bridge means that the better team may still lose even over a long match.
  
• A match against another contender is twice as significant as other matches.
  
• The last 3 formats tend to allow more boards overall because there is less time taken with the movement and system disclosure.
  
• The result of a Swiss competition can depend on when the music stops; but it still seems to produce the most accurate ranking.
  
• What about this case? Team A usually bets B. B usually beats C; C usually beats A. The Swiss format seems to cater well for such rare scenarios.

[the hog]

Cascade, on Oct 20 2008, 02:08 AM, said:

On the other hand even the two hour time difference back from Australia to New Zealand seems to take me days to recover from.

Well there is an obvious reason for that, Wayne!

[Cascade]

The_Hog, on Oct 21 2008, 12:51 PM, said:

Cascade, on Oct 20 2008, 02:08 AM, said:

On the other hand even the two hour time difference back from Australia to New Zealand seems to take me days to recover from.

Well there is an obvious reason for that, Wayne!

Oh really?

what are you referring to?

[cherdano]

I don't believe Nickf's numbers either. 6 rounds to reliably determine a winner in a field of 25 teams?
Give me a break. And the length of the rounds doesn't matter, either?

[matmat]

cherdano, on Oct 20 2008, 08:20 PM, said:

I don't believe Nickf's numbers either. 6 rounds to reliably determine a winner in a field of 25 teams?
Give me a break. And the length of the rounds doesn't matter, either?

takes 6 shuffles to randomize the cards, doesn't it?

[MickyB]

matmat, on Oct 21 2008, 02:26 AM, said:

cherdano, on Oct 20 2008, 08:20 PM, said:

I don't believe Nickf's numbers either. 6 rounds to reliably determine a winner in a field of 25 teams?
Give me a break. And the length of the rounds doesn't matter, either?

takes 6 shuffles to randomize the cards, doesn't it?

Yup - so six rounds would be needed to determine a winner in a field of 52, but only five in a field of half that size.

[the hog]

Cascade, on Oct 21 2008, 08:11 AM, said:

The_Hog, on Oct 21 2008, 12:51 PM, said:

Cascade, on Oct 20 2008, 02:08 AM, said:

On the other hand even the two hour time difference back from Australia to New Zealand seems to take me days to recover from.

Well there is an obvious reason for that, Wayne!

Oh really?

what are you referring to?

Well its the feeling of regret and remorse for leaving Oz and returning to EnZed.