[mrdct]

PeterGill, on Oct 7 2009, 10:27 PM, said:

Also, the results of Australian teams selected by Pairs Trials are much worse than those selected by Teams Trials. Australia's dismal showing in Beijing last year
after Pairs Trials was not an isolated case. I could add a few hundred lines of actual Australian data for the last ten years here, but I will not. The data exists because we alternate Pairs Trials and Teams Trials from year to year.

Whilst this is a tread about USA team selection methods, Australia is an interesting test case because we dable in both pairs-based and teams-based selection methods.

I thought I'd have look at the data myself as it's all published on the internet and it turns out that over the ten year period that Peter refers to, the pairs-based teams have actually performed better than the teams-based teams in terms of average percentile placing in qualifying fields.

Since 1/1/2000 Australia has played in 6 Bermuda Bowls and 3 Olympiads/WMSGs. Of those 9 teams, 2 were selected from a pairs trial and 7 were selected from a teams trial.

The pairs-based teams contested the 2004 and 2008 Olympiads coming 7th/18 and 11th/18 respectively which on average is exactly mid-field (average placing of 50% of the field size).

The teams-based teams have had 3 quite credible performances making the KO stages in the 2000 Olympiad and the 2003 and 2007 Bowls but have had some poor performances also - finishing in the bottom third of the field in the 2000 (in lieu of 1999), 2001, 2005 and 2009 Bowls. On average the teams-based teams have come just slightly below mid-field with an average placing of 52% of the field-size.

Not much in it, but the statistics don't really support Peter's contention that "teams selected by Pairs Trials are much worse than those selected by Teams Trials".

Another quite telling observation that I've made in compiling this data is that in the aforementioned events Australia has fielded no less that 37 players on its open team whilst in the same events Italy has only fielded 11 players and at no time has ever made more than 2 changes to its line-up.

[gnasher]

mrdct, on Oct 22 2009, 02:35 AM, said:

The pairs-based teams contested the 2004 and 2008 Olympiads coming 7th/18 and 11th/18 respectively which on average is exactly mid-field (average placing of 50% of the field size).

The teams-based teams have had 3 quite credible performances making the KO stages in the 2000 Olympiad and the 2003 and 2007 Bowls but have had some poor performances also - finishing in the bottom third of the field in the 2000 (in lieu of 1999), 2001, 2005 and 2009 Bowls. On average the teams-based teams have come just slightly below mid-field with an average placing of 52% of the field-size.

It doesn't sound as though you have enough data to evaluate the different methods reliably. If this data means anything, however, it means that pairs-based trials lead to mediocre results, whereas teams-based trials give chances of significant success, but at a greater risk of failure. I know which I'd prefer, both as a player and as a member of a national bridge organisation

Take England's recent Olympiad performance as an example. I know that we lost in the final in 2008, and in the semi-final in 2000, and I know who played in each team. I have no idea where we finished in 2004 or who was in the team; all that I know is that we didn't reach the knockout stages.

[mich-b]

mrdct, on Oct 21 2009, 08:35 PM, said:

I thought I'd have look at the data myself as it's all published on the internet and it turns out that over the ten year period that Peter refers to, the pairs-based teams have actually performed better than the teams-based teams in terms of average percentile placing in qualifying fields. 

Since 1/1/2000 Australia has played in 6 Bermuda Bowls and 3 Olympiads/WMSGs. Of those 9 teams, 2 were selected from a pairs trial and 7 were selected from a teams trial.

The pairs-based teams contested the 2004 and 2008 Olympiads coming 7th/18 and 11th/18 respectively which on average is exactly mid-field (average placing of 50% of the field size).

The teams-based teams have had 3 quite credible performances making the KO stages in the 2000 Olympiad and the 2003 and 2007 Bowls but have had some poor performances also - finishing in the bottom third of the field in the 2000 (in lieu of 1999), 2001, 2005 and 2009 Bowls.  On average the teams-based teams have come just slightly below mid-field with an average placing of 52% of the field-size.

Not much in it, but the statistics don't really support Peter's contention that "teams selected by Pairs Trials are much worse than those selected by Teams Trials".

I think that it is very wrong to compare performance in Olympiad to performance in BB by terms of finishing place in the qualifying stage.
It is obvious that a qualifying group of 18 in an Olympiad (an open event!), would have quite a few weak teams (and only 1/4 of the strong teams), while the BB field has no weak teams and all the strong teams (including 2 from USA).

So certainly finishing 11/18 in the qualifying stage of the Olympiad is a poor result, equal to last in the BB. and 7/18 in the Olympiad group is equal to about 25th overall in the Olympiad , which also is far from being equal to average in BB.

So , it seems to me that the data as you present it, does very much support Peter Gill's claims.

[helene_t]

Obviously I know nothing from personal experience, but the Dutch model with a core group of 5-7 pairs from which teams are selected based on form curve and availability seems like a good model to me. It is more flexible than a team trial and avoids most of the social issues related to a pairs trial which might select pairs that don't really want to play together.

A few years ago there was a 4-handed team (Zia was one of the players I think) in the trials. The reason they didn't have a 3rd pair was that they had been playing with foreign pairs in other big US tournaments, so at the time of the trial it was difficult to find a good 3rd pair. So they hoped to qualify as a 4-handed team and then recruit a pair from one of the non-qualifiers for the BB. That seems awkward but probably doesn't happen too often.

[nige1]

mrdct, on Oct 21 2009, 08:35 PM, said:

Whilst this is a tread about USA team selection methods, Australia is an interesting test case because we dable in both pairs-based and teams-based selection methods. I thought I'd have look at the data myself as it's all published on the internet and it turns out that over the ten year period that Peter refers to, the pairs-based teams have actually performed better than the teams-based teams in terms of average percentile placing in qualifying fields.  Since 1/1/2000 Australia has played in 6 Bermuda Bowls and 3 Olympiads/WMSGs. Of those 9 teams, 2 were selected from a pairs trial and 7 were selected from a teams trial. The pairs-based teams contested the 2004 and 2008 Olympiads coming 7th/18 and 11th/18 respectively which on average is exactly mid-field (average placing of 50% of the field size). The teams-based teams have had 3 quite credible performances making the KO stages in the 2000 Olympiad and the 2003 and 2007 Bowls but have had some poor performances also - finishing in the bottom third of the field in the 2000 (in lieu of 1999), 2001, 2005 and 2009 Bowls.  On average the teams-based teams have come just slightly below mid-field with an average placing of 52% of the field-size. Not much in it, but the statistics don't really support Peter's contention that "teams selected by Pairs Trials are much worse than those selected by Teams Trials". Another quite telling observation that I've made in compiling this data is that in the aforementioned events Australia has fielded no less that 37 players on its open team whilst in the same events Italy has only fielded 11 players and at no time has ever made more than 2 changes to its line-up.

Thank you Mrdct. I don't think there's enough Australian trials data, especially pairs data, to arrive at a firm conclusion. I dare say we'll interpret them to suit our prejudices; but I agree with Mrdct that there appears to be no significant difference. If some pairs qualified only in teams trials or only in pairs trials, perhaps comparing their butlers would produce a more polarised result. Presumably, more data is available from other places and other years.

[cherdanno]

So aside from the fact that mrdct misleadingly compares Olympiad and BB performances, he also seems to disagree with PeterGill about whether the 2009 team was selected by a teams trial or pairs trial.

[JanM]

mrdct, on Oct 20 2009, 03:24 PM, said:

Perhaps a way to keep both camps happy would be to run a teams trial where the winner becomes USA1 and the runner-up earns a right to contest a play-off against a team determined from a pairs trial to become USA2.

Something very like this format was proposed for the 2009 US Trials, and was overwhelmingly rejected by the (very large and quite representative of the players in the Trials) International Team Trials Committee after a lot of discussion. At least some of the professional pairs said that they would not play in a pairs qualifying event under this sort of format. The ITTC chose instead to run what was essentially a double elimination KO (the teams that lost from the Round of 16 on played in the USA2 bracket). The winner of the USA2 bracket won the Bermuda Bowl, which probably will provide support for those of us who feel that a team trials is better than a pairs trials

[nige1]

Jlall, on Oct 20 2009, 05:32 PM, said:

I have a hard time believing there are people who don't play in the trials because it's a team trials rather than a pairs.

I disgree with JLAll. Suppose God would rank 48 players in order G1, G2, ... G48 (in terms of technical and partnership skills) and they play a tournament of 480 boards. In an individual event, you would expect a poor correlation between the actual results and God's order. In a pairs event where the better players tended to play with each other, you would expect a better correlation. In a teams event, where the better pairs tend to play in the better teams, you would expect an even better correlation. Arguably, this is almost self-selection. But unfortunate consequences of teams trials seem hard to avoid:

• Good teams (season-ticket holders) tend to stick together, year on year. With a practical monopoly on international experience, the top teams become better and better, as their members grow older and older.
  
• Good new pairs (seed-corn) will usually have to team up with inferior team-mates. With little prospect of selection, a good new pair may not deem it worthwhile to enter the trials; this is a pity because trials, themselves, are an excellent training ground. A pair may not even judge it to be worth the hassle to hone their skills to international standard. Hence teams-trials do not encourage new participants.

IMO, if you can ignore sponsorship, then pairs trials would produce better teams, in the long term, than teams trials. They would also make the game more interesting and exciting for new players. Sponsorship is here to stay however, so this discussion is academic, unless you adopt a suggestion like MrDict's

[awm]

There's certainly a point of view that says:

(1) The main reason to have a pairs trial is to get some good pairs involved who might otherwise not participate. Even though these pairs are in most cases not good enough to actually make the BB team, playing in the trials will be good experience for them, and if they do well it might help them get picked up by a good team in future years.

(2) The teams currently participating in the trials clearly find the format acceptable. Adding new pairs to the mix, or having a pairs trial of some form as suggested, would potentially reduce the chances of the currently participating teams winning the trials.

(3) The evidence that "everyone is happy with the current system" is based primarily on polls and votes of the current participants. It is unsurprising that the substantial majority likes things as they are, given the first two points.

[fred]

nige1, on Oct 22 2009, 04:13 PM, said:

Jlall, on Oct 20 2009, 05:32 PM, said:

I have a hard time believing there are people who don't play in the trials because it's a team trials rather than a pairs.

I disgree with JLAll. Suppose God would rank 48 players in order G1, G2, ... G48 (in terms of technical and partnership skills) and they play a tournament of 480 boards. In an individual event, you would expect a poor correlation between the actual results and God's order. In a pairs event where the better players tended to play with each other, you would expect a better correlation. In a teams event, where the better pairs tend to play in the better teams, you would expect an even better correlation. Arguably, this is almost self-selection. But unfortunate consequences of teams trials seem hard to avoid:

• Good teams (season-ticket holders) tend to stick together, year on year. With a practical monopoly on international experience, the top teams become better and better, as their members grow older and older.
  
  
• Good new pairs (seed-corn) will usually have to team up with inferior team-mates. With little prospect of selection, a good new pair may not deem it worthwhile to enter the trials; this is unfortunate because trials, themselves, are an excellent training ground. Thus a pair may not even judge it to be worth the hassle to hone their skills to international standard. Hence teams-trials do not encourage new participants.

IMO, if you can ignore sponsorship, then pairs trials would produce better teams, in the long term, than teams trials. They would also make the game more interesting and exciting for new players. Sponsorship is here to stay however, so this discussion is academic, unless you adopt a suggestion like MrDict's

I think you are wrong, Nigel, at least as far as the USA is concerned (where you can get plenty of "international experience" without playing in a single World Championship).

IMO "good new pairs", regardless of their bridge talent, who are not willing to pay their dues (by playing with inferior teammates during the time in which they improve to the point of being world class and to the point that they are recognized as such) probably do not have what it takes in terms of mental makeup to compete at the highest levels.

And, notwithstanding professional considerations, the best teams tend to stick together only as long as they continue to win.

Fred Gitelman
Bridge Base Inc.
www.bridgebase.com

[jjbrr]

Nige1, what do you mean by "good new pair"? If such a thing exists, they would never go unnoticed in America and thus would certainly have a team for the trials. If they were good enough to actually make a splash in the trials, there is no chance they would simply not play in the trials because of the format.

And what does the format have to do with "new participants"? There are lots and lots of other opportunities for a new player to establish his reputation. As so, why would the interests of a new participant be any concern whatsoever to anyone else?

[TimG]

JanM, on Oct 22 2009, 10:26 AM, said:

Something very like this format was proposed for the 2009 US Trials, and was overwhelmingly rejected by the (very large and quite representative of the players in the Trials) International Team Trials Committee after a lot of discussion.

It is unsurprising to me that a committee made up of team trials players favor a team trials format.

A decade or so ago, District 25 was considering early starting times (10 & 2 or similar) instead of late starting times (1 & 7). A survey was conducted at a regional where all events were run at 1 & 7 and the results showed that players preferred 1 & 7. No one should have been surprised.

This is not to say that the ITT Committee got it wrong. I just think that having the committee made up primarily of players who play in the event and may have a financial interest in seeing it run one way rather than another is a bad idea.

Tim

[mrdct]

cherdanno, on Oct 22 2009, 09:14 AM, said:

So aside from the fact that mrdct misleadingly compares Olympiad and BB performances, he also seems to disagree with PeterGill about whether the 2009 team was selected by a teams trial or pairs trial.

Addressing your first point, it is somewhat unavoidable to have to compare Olympiad and Bermuda Bowl results given that for at least the last 20 years Australia has used teams trials to select 100% of its Bermuda Bowl teams and has used pairs trials for most, but not all, Olympiad Teams. Pairs trials are also used to select our team for the PABF Championships (East Asia & South Pacific) in non-Bowl years, but I haven't included them in the data as I don't think the field is even remotely comparable (although Australia has failed to win that event for more than 30 years now so I guess it does provide some anecdotal evidence against pairs trials).

On your second point, Australia's 2009 Bermuda Bowl team of I Thomson, I Robinson, D Appleton, A De Livera, P Reynolds, R Brightling was defintely selected by way teams trial - results. With no disrespect to any of these players who are all fine bridge players and certainly much better than me, this team went on to finish 18th/22 in the Bermuda Bowl which I believe was Australia's worst performance in the history of that event. This team overcame the favoured team of M Green, P Gill, S Hans, T Nunn, A Peake, B Richman in a 64-board semi-final and then defeated the four-handed team of J Ebery, L Gold, T Antoff, A Simpson by 70 imps over 64-boards in the final.

Let me be quite clear - I am not an advocate of pairs trials, but I just don't see the evidence backing up Peter Gill statement, "teams selected by Pairs Trials are much worse than those selected by Teams Trials". What I do believe though is that for the size of its bridge population (32,707 which is greater than all but two European countries - France and Netherlands and I think is currently the 5th largest bridge population in the world) Australia has significantly and consistently under-performed at top-level bridge. So either we aren't very bright down here, we aren't developing our top players properly and/or we don't have our selection methods sorted out properly.

[mike777]

It seems in Aust. your selection methods are just fine. What is the problem? IN fact you mention Aust. has the fifth largest dues paying Bridge population. Aust. is doing something right! The number one priority is not to win but to let everyone compete in a somewhat level playing field. I have no issue with giving seeding points.

If that means Mr and Ms nonexpert make up the team to represent my country...great!

[awm]

Selecting a good team is part of the goal here. People like to see their country win, and there may be benefits to having world champions in your country too. I'd say the goals of selection should be:

(1) Select the best (or nearly-best) possible team.
(2) If there are many players of top quality (i.e. good enough to be on the team), make sure that all of them (or at least many of them) get the chance to represent the country at some point.
(3) Create at least the impression that the team is selected in a fair way.

The first goal should be pretty clear. There are several reasons for the second goal. One issue is that players get older, and if the same team has been sent for the last decade or two there may suddenly be a shortage of international experience if a player dies or retires from bridge. Another is that it allows for "mixing things up" if a particular team hasn't been successful recently. It also leads into the third goal -- people are more likely to support a national team in various ways if they feel that the selection process is fair. It also averts blame being placed on the selection method or committee if the team doesn't succeed.

The key to which is better (teams or pairs trial) depends on what we believe about the various things. For example, suppose there are three pairs that are much better than the others. Do we believe that these pairs will end up on the same team for the team trials (my impression is very likely they would not)? Do we believe that these three pairs would normally make up the best team (this is less clear, and probably depends on how well they get along)? And how realistic is this idea that there's a small number of pairs much better than the others anyway?

My impression is that in the USA, there are a fairly large number of pairs (more than six pairs anyway) which are of roughly equivalent level. Assuming that there's not substantial friction between teammates (some of these people are hard to get along with) pretty much any team composed of three of these good pairs will have a very good shot at winning a world championship. However, what happens in practice is that it's quite rare for a single team to have three of these pairs (because of sponsorship constraints). What frequently happens is that there are one or two "sponsor pairs" who are top quality (i.e. Nickell-Freeman). These top sponsors hire two other top pairs, and usually win the trials because they have three top pairs whereas other teams (with a weaker sponsor) have only two. In principle a team of three top non-sponsor pairs could form and contend for the event, but this only rarely happens (financial reasons). So the upshot is that the team trials often becomes "selection by the few wealthy sponsors who can actually play the cards well." This means the same teams often represent the US year after year (obviously they are good teams; certainly they are better than any other team with a sponsor that we could put together). But this has all the downsides mentioned by failing at goals 2 and 3 above. The problem is reduced a little by having two teams instead of one, but we do still see a lot of the same faces. Again, if these players are clearly head and shoulders above everyone else, it would be unavoidable to some degree... but it's not clear this is really what's going on.