How to Choose 2 teams in 1 Trials

[JanM]

For a change of pace, and to get any good ideas you might have :P, how would you choose 2 teams from one Trials. Assume you start with the decision that both teams should be chosen as teams (iow, even if you think a pairs trials would be better, it's been rejected), and that the basic selection method for one team will look like this year's USBC, how would you go about selecting the second team from the "losers bracket" ?

[cherdano]

A double knockout? Since I am sure you want to finish with a knockout phase this seems like the normal idea.

[Guest Jlall]

I would have the losers of the semi final match play while the finals were going on. Winner of finals=USA 1. The loser of the finals plays the winner of the the match between the losing semi finalists. The winner of this match is USA2.

[mike777]

I would have the losers of the semi final match play while the finals were going on. Winner of finals=USA 1. The loser of the finals plays the winner of the the match between the losing semi finalists. The winner of this match is USA2.

ditto

[uday]

I'm sure I don't know what I'm talking about but if we assume that

 

any team beaten by the winner is a candidate for 2nd strongest team (in this event)

any team beaten by any other team is not.

 

Then perhaps the teams that lost to the winners in head to head play, and no one else, should compete amongst themselves for 2nd place.

 

This is perhaps complicated/simplified by byes granted to highly seeded teams.

[gwnn]

I'm sure I don't know what I'm talking about but if we assume that

 

any team beaten by the winner is a candidate for 2nd strongest team (in this event)

any team beaten by any other team is not.

 

Then perhaps the teams that lost to the winners in head to head play, and no one else, should compete amongst themselves for 2nd place.

 

This is perhaps complicated/simplified by byes granted to highly seeded teams.

I think that's a great suggestion.

[mike777]

Since only one trial, how can you do this? You do not know the winning team until the end?

 

This sounds like two trials :P

[awm]

Run a double-elimination event. The team that goes undefeated is USA1. The team that ends with one loss is USA2. Constructing this can be a bit complex, but should be possible. Allowing byes actually makes this easier to construct, and it removes the luck factor of playing the best (winning) team in an early round.

[jdonn]

I'm sure I  don't know what I'm talking about but if we assume that

 

any team beaten by the winner is a candidate for 2nd strongest team (in this event)

any team beaten by any other team is not.

 

Then perhaps the teams that lost to the winners in head to head play, and no one else, should compete amongst themselves for 2nd place. 

 

This is perhaps complicated/simplified by byes granted to highly seeded teams.

I have never heard of this being done but it actually makes a lot of sense. Simply have the team that lost to the eventual winners in the semifinal play the team that lost in the final. The scenario that this would avoid is a team winning the semifinal, losing in the final, then having a rematch with the team they already beat in the semifinal and losing this time around. If they won the first time by 100 and lost the second time by 3 somehow it would seem unfair that they don't get to go. So the logic Uday presented appeals to me.

[Gerben42]

The best I can do is 2 rounds longer than single trial.

 

16 qualifiers from RR or so, then:

 

R1: 8 winners, 8 losers

 

R2: 4 semifinalists, 8 with 1 loss, 4 out

 

R3: 2 finalists, 6 with 1 loss, 8 out

 

R4: 1st final - 4 with 1 loss

 

R5 & 6: KO for the 4 with 1 loss

 

Add byes as needed.

[hrothgar]

Hi Jan

 

Back in April, Gerben Dirksen and Alex Ogen did some simulations comparing different types of tournament designs. The analysis was based on a technique known as Monte Carlo simulation. A group of teams with known strength was created using one of a variety of different distributions (normal, uniform, what have you). The teams were run through a tournament with format X. At the end of the tournament, the simulation compared the ordering of teams produced by the tournament (the sample statistic) with the known strength of the teams (the population statistic). The closer that the sample statistic mirrored the population statistic, the more accurate the tournament format. This specific set of simulations was created to study a problem that was originally posed by the Australian Bridge Federation. The same methodology could be (easily) modified for the USBF Team Trials.

 

With this said and done, I’m not sure whether it would be worth doing the necessary work. My impression of the USBF is that subjective considerations outweigh objective criteria.

 

Case in point: Let’s assume that the USBF needed to select a single team from a teams trial. Furthermore, let’s assume that I was able to conclusively prove that a single elimination tournament with format XYZ was much more accurate than a round robin followed by a (shorter) KO. My understanding is that the USBF isn’t interesting in using a single elimination format because teams don’t want to bother going to the time and expense of traveling to a tournament only to get eliminated in the first round. The objective goal “Select the best team” is ends up subordinate to the subjective goal “Make sure that teams have fun”. This same sort of issue crops up time and time again: You asked for advice regarding creating the Convention Regulations for the Team Trials. I suggested that you use the same set of regulations as the event that the team would be competing in. You responded that this wasn’t practical because you wanted to make sure that you selected the best team and couldn’t allow high variance methods.

 

In short, if you can provide a very specific description regarding the set of constraints, I think that folks can identify the most accurate tournament. Absent this type of description, the problem isn’t an attractive one to work on.

[Aberlour10]

Runner up & both semifinalists play a long "3-teams-match"

(with carry over for the runner up)

 

Robert

[TimG]

A double knockout? Since I am sure you want to finish with a knockout phase this seems like the normal idea.

If this is such that all losers are sent to a loser's bracket and the winner of that bracket is USA2, that seems right to me.

 

Filling out the losers' bracket with non-qualifiers from the RR phase can be done during the first round of the KOs to make the losers' bracket the right size in the event that there will be byes in the winners' side.

[cherdano]

A double knockout? Since I am sure you want to finish with a knockout phase this seems like the normal idea.

If this is such that all losers are sent to a loser's bracket and the winner of that bracket is USA2, that seems right to me.

 

Filling out the losers' bracket with non-qualifiers from the RR phase can be done during the first round of the KOs to make the losers' bracket the right size in the event that there will be byes in the winners' side.

Yes, it's the same thing that Gerben described, and that Adam called "double elimination" tournament.

 

I don't like uday's suggestion so much, I find it very odd when the question whether team C remains in competition depends on whether team A is beating team B.

[hrothgar]

Since only one trial, how can you do this? You do not know the winning team until the end?

 

This sounds like two trials :P

Mike has raised a very good point.

 

The fundamental constraint that you are operating under is time. The system that Uday suggests does not allow you to start the selection process for the second best team until the best team has been identified. I'd bet dollars to donuts that a double elimination format will make much more effective use of the available time.

 

Here's a snarky suggestion: Auction off the right to represent USA1 and USA2 to the sponsors willing to pay the most money. Let them hire whoever they want to round out the team. Feed the money into a marketing program, improved versions of "Learn to Play Bridge", what have you.

[TimG]

Here's a snarky suggestion: Auction off the right to represent USA1 and USA2 to the sponsors with the most money. Let them hire whoever they want to round out the team. Feed the money into a marketing program, improved versions of "Learn to Play Bridge", what have you.

Play for USA1; auction off USA2.

[pclayton]

Here's my feelings:

 

The qualifying should mirror the actual event, so I'm not a big fan of byes. Stamina should be a big factor in team selection, and if a team gets a free ride to the semi-finals, then stamina doesn't play nearly the factor that it should.

 

I'm also very skeptical about augmentation. If four (or even 5) warhorses grind it through and somehow gain one of the spots, I can see putting the two seats on the open market and letting someone buy their way onto our national team. Isn't that possible? Or does the USBF have a say-so in who is allowed to be added?

 

I agree with Justin that the two semi-finalists have the opportunity to play each other and then meet the loser to USA1 in the final for USA2.

[mike777]

Setting goals is important. As usual there will often be competing but conflicting decent goals.

 

As the vast majority, 99% or more of us who play and pay for the sport I never understood why so many countries try and select on the basis of choosing the "best team" or the "team most likely to win" as chosen by some selector or admin.

 

Why not seed as a reward for "regular season" performance, try and have some reasonable competitive format and let whoever wins, win! In other words you need to play and beat the other guys and gals to go to the Bermuda Bowl.

 

IF a buch of nobodies win...great...that should be a good thing for our sport and local country not a bad thing. :P

 

Of course this all means the main goal is some form of fair competition where the admin do not try and pick the best team.....

 

It seems for the past few years or decade, the USBF has done a good job in letting us nobodies have a shot, but rewarding the better teams for regular season performance.

[hrothgar]

I agree with Justin that the two semi-finalists have the opportunity to play each other and then meet the loser to USA1 in the final for USA2.

The problem with this format is that a single elimination KO provides no information regarding the relative strength of the different teams that were KOed by the eventual winner. For all we know, the team that lost to the winner in the very first match is actually better than three of the four semi-finalists.

 

Obviously, a good seeding system will decrease the chance that this happens. Then again, if you're that confident in your seeding system, there really isn't any reason to have a tournament. Just select your number 1 and number 2 seeds.

[Walddk]

In my view the closest you get to fairness is the format of The Schapiro Spring Foursomes in England. Conditions of Contest:

 

http://www.ebu.co.uk/competitions/may/spri...generalinfo.htm

 

with various appropriate links. "How it is scored" is one of those.

 

Perhaps you need to modify it a little in USA, not sure, but the general concept is excellent.

 

Roland

[jtfanclub]

Then perhaps the teams that lost to the winners in head to head play, and no one else, should compete amongst themselves for 2nd place.

Hmmm...suppose their were 5 rounds, of KO. The last round is double length.

 

After two rounds, for each team A still in the running, have the team who lost to team A in round 1 play the team who lost to team A in round 2.

 

After round 4, have the team who lost to team A in round 3 play the team who lost to A in round 4.

 

Then have the winner of the 1-2 losers play the winner of the 3-4 losers.

 

And finally, after the KO is over, have the finals loser play the winner of the 'defeated by A' guys.

 

I think that sounded more complicated than it is. Example, in which the higher letter always beats the lower letter, but the seeding is reversed, so that the two best teams meet in the first round. ' shows the number of losses, including that round.

 

Round 1:

A defeats B'

 

Round 2:

A defeats C'

 

Round 3:

A defeats D', and B' defeats C''

 

Round 4:

A defeats E'

 

Round 5, part 1

A plays F, D' defeats E''

 

Round 5, part 2

A defeats F, B' defeats D''

 

Consolation Round

B' defeats F''

 

A is USA1, B is USA2.

 

Of course, this means an awful lot of people who never played A will end up playing consolation rounds for nothing when the team that knocked them out gets knocked out themselves.

[jdonn]

I agree with Justin that the two semi-finalists have the opportunity to play each other and then meet the loser to USA1 in the final for USA2.

I just got the idea of sort of a comprimise suggestion between that and what Uday suggested, just throwing it out there. Go with your suggestion, but if the second place match is a rematch of one of the semifinal matches then include either a full or partial carryover based on the outcome of the semifinal. That way you can't beat a team by 100 the first time, lose by 1 the second time, and have them be chosen over you. It would have an added bonus (well I consider it a bonus but others may not) that the intergrity of the last portion of a semifinal blowout would be preserved, as a team losing by a ton wouldn't start either wildly swinging or simply trying less hard.

 

I admit it seems strange to make the match for US 2 potentially longer than the match for US 1, but that doesn't bother me as long as the match for US 1 is sufficiently long in the first place. Perhaps the answer is a 50% carryover and make the second match half as long as the match for US 1, which effectively makes the matches the same length.

 

In the case where the match for US 2 is not a rematch of the semifinals, then it is the same as Uday's suggestion anyway as it's the last two teams that got beat by the winner.

 

Any thoughts?

[mike777]

If you are going to do all of this, why not just have two trials with seeding and byes.

 

Sure there may be a bit more expense but so? We are here to play bridge and have fun yes, not pinch pennies?

 

If the number one goal is too save money start having tourneys on the web with proctors.

Reduce, Hotel, airfare, dining out expenses.

[barmar]

If you are going to do all of this, why not just have two trials with seeding and byes.

Isn't that what we currently have? BB is every other year, and we have a trial each year. The winner the first year is USA1, the winner the second year (which USA1 doesn't play in) is USA2.

 

Or am I thinking of something else?

[keylime]

I think having as few of byes possible is best for the entire event. One of the buggers about the USBF system are the 8 byes given. I strongly prefer 4 byes (or less - why not slog your way through it instead?)

 

I like the curling method of playoffs myself (called the Page method if memory serves) - during my time in Nanaimo the Vancouver Bridge League used it and I thought it was quite rational and balanced. It allows for the number 1 team to go to the finals yet does allow for a high seeded team to lose a match and still recover to win the entire event.