[ovncylmz]

So nullve-zebutin 4 x 16.. Goodluck!
Btw, as soon as that match ends, I open up registration for the next tournament. The tournament style will include something where everybody plays at least 3 matches. I will post for the details.

[nullve]

nullve-zebutin: 70-96 (10-24, 24-16, 25-20, 11-36)

Set 1: http://webutil.bridg...aeb4-1473908067
Set 2: http://webutil.bridg...aeb4-1473944413
Set 3: http://webutil.bridg...aeb4-1473954691
Set 4: http://webutil.bridg...aeb4-1473963205

A well-deserved win for zebutin!

[ovncylmz]

Hello everybody, so we start the registration window for the new event. Please register through here or send me a message from BBO (make sure you get a confirmation if you send via BBO)! Here are the initial details:

I am thinking of some system with groups in which everybody plays at least 4 or 5 short matches (~12boards). It will be non-best hand and IMPs as Event #2. People already won at least a match in Event #2 will be seeded based on the number of wins - the rest and new entrants will be randomly placed in the groups.

The group stage will be ~48 to 60 boards and will be played in two weeks. Then top 8 players will go to KO stage for the championship. The players who cannot make it to KO will join a "consolidation swiss" (if they want to continue) where they start with some points - some percentage based on the number of matches in the group. At the end, we will just allocate some masterpoints.

Tomorrow frank0 will make a new post with the rules we thought of. We are open to ideas to improve this league, so we can discuss the system during the registration window (Today to Next Sunday, 9 days-window). We will make the draw and start the tournament. Please spread the word!


>> Btw, congrats Zebutin, also Nullve, Dadim and Diana_eva for making it to top four.

[hrothgar]

If there is anything that these events are showing, its the sheer amount of variance in the results of individual boards.

I think that moving to large numbers of short rounds is a big mistake

[ovncylmz]

hrothgar, on 2016-September-17, 03:22, said:

If there is anything that these events are showing, its the sheer amount of variance in the results of individual boards.

I think that moving to large numbers of short rounds is a big mistake

You will play against multiple opponents without being knocked out - 12/16 boards each x 4-5 matches. This is actually a good idea when there are no clear seedings (compared to KO when two top players may play very early and one loses immediately). Given you play around ~60 boards in the first stage (more than 32 of the current event) and multiple (2 or 3) players qualify from each group, actually we believe that it will be a lot "less random".

And of course, the KO stage (Top-8) will be starting from 64 boards with final being 96 boards.

The competition is cool, we also need to make this fun - so we thought it is a good idea to give people more interaction with others - groups then KO+consolidation. So when you lose in group stage, you don't wait 4 weeks for others to compete and finish - you will have a chance to play in consolidation if you want to play.

You will see the announcement today.

[nige1]

My tuppence worth:

The problem with a KO is that, even if luck is entirely eliminated, the second best competitor can be knocked out in the first round. If luck is a significant factor, then a KO is pretty crude.

The problem with a round-robin is that matches are too short, especially among the best players. A lot depends on hammering the worst competitors.

A round-robin followed by a KO mitigates some of these problems but equalisation of group-strengths is prone to error (witness the recent World Championships). this schema also Introduces the complication of tactical dumping.

Is it possible for everybody to play the same challenges? like an individual competition against robots? This would allow very long matches. And keep everyone playing until the end.

To further increase the skill component, we could use Match-points rather than imps scoring.

[ovncylmz]

nige1, on 2016-September-17, 09:13, said:

My tuppence worth:

The problem with a KO is that, even if luck is entirely eliminated, the second best competitor can be knocked out in the first round. If luck is a significant factor, then a KO is pretty crude.

The problem with a round-robin is that matches are too short, especially among the best players. A lot depends on hammering the worst competitors.

A round-robin followed by a KO mitigates some of these problems but equalisation of group-strengths is prone to error (witness the recent World Championships). this schema also Introduces the complication of tactical dumping.

Is it possible for everybody to play the same challenges? like an individual competition against robots? This would allow very long matches. And keep everyone playing until the end.

To further increase the skill component, we could use Match-points rather than imps scoring.

1) We are planning to seed people based on Event #2 results. New comers / early losers in the previous event will be randomly placed into groups after seeded players. After a few tournaments, seeding issue will be taken care of since the data gets bigger - and things will get better.

2) We had lots of logistics issues in the first tournament because of the "biggest difference" rule. We clearly don't want to deal with it again. Unless you have exactly 16 or 32 players, it is problematic to start a KO with ACBL style 3-way or 4-way matches:
--> 3-way is reasonable if it is played at the same time, which we cannot enforce for challenges. We may have issues people waiting for each other etc.
--> 4-way matches with 3-survivors are also tricky since we don't have a good seeding at this point. Usually top seeds play head-to-head vs. players who are at the bottom. In our case, the bottom does not say anything since it may be a player who is playing his first event. Therefore, it may unfair to assign head-to-head and four-ways to people.
Therefore, groups with RR and then KO looks the best.

3) You cannot make people play the same boards, that is not an option in BBO (and also again, people play different times and I can easily ask my friend to give the hands after he plays etc.)

4) Yes, the challenge events are random. I know most of the top-8 in last tournament. When I was talking to some of them, they all admitted that a lot of luck involved. So it is possibly better to keep it as a "ladder" and enjoy as much as possible while competing.

[frank0]

From now on please discuss event 3 here.

http://www.bridgebas...mat-discussion/

[smerriman]

nige1, on 2016-September-17, 09:13, said:

To further increase the skill component, we could use Match-points rather than imps scoring.

I think (as did several earlier in this thread) in this type of challenge event this actually increases the luck factor considerably, not the skill factor.

[cherdano]

smerriman, on 2016-September-17, 14:41, said:

I think (as did several earlier in this thread) in this type of challenge event this actually increases the luck factor considerably, not the skill factor.

Strongly disagree.

Still, thanks a lot to everyone running this - I'd like to sign up for #3.

[ovncylmz]

I wrote something first but then noticed that I really didn't understand what you guys are discussing.
@Smerriman, do you defend IMP?
@Cherdano, do you defend MP?

I mean, we really need to go IMPs - but we can make a MP tournament in the future.

[cherdano]

I think MPs makes it more bridge-skill-dependent, less luck-dependent, and also a little bit less specific-GIB-skills-dependent (such as avoiding crazy auctions, or provoking GIB into crazy auctions that you can double).
I'd also prefer MP for the event but that's not as strong an opinion.

[ovncylmz]

cherdano, on 2016-September-17, 17:26, said:

I think MPs makes it more bridge-skill-dependent, less luck-dependent, and also a little bit less specific-GIB-skills-dependent (such as avoiding crazy auctions, or provoking GIB into crazy auctions that you can double).
I'd also prefer MP for the event but that's not as strong an opinion.

Although I agree IMP may mean a lot of big plus and minuses while playing GIB, MP is also quite random with GIBs.

[cherdano]

ovncylmz, on 2016-September-17, 17:28, said:

Although I agree IMP may mean a lot of big plus and minuses while playing GIB, MP is also quite random with GIBs.

Of course, many boards are random. But when I get one more trick as declarer than the rest of the field, it will more often be because I did something good than when I am the only one to bid a making slam (that might only make because GIB makes a silly lead).

[smerriman]

cherdano, on 2016-September-17, 17:36, said:

Of course, many boards are random. But when I get one more trick as declarer than the rest of the field, it will more often be because I did something good than when I am the only one to bid a making slam (that might only make because GIB makes a silly lead).

That's where the problem is. With an actual field, that makes perfect sense and MP is far better. But here there is no "field" - you're competing one vs one. With GIB's bidding and play, even if you end up in the same contract, you're very likely to differ by a trick here and there regardless of what you did, making the results of each board quite random.

In MP, these tricks are highly important in every decision in every hand. In IMPs, they don't make as much of an impact.

(That's why I prefer MP for the daylong GIB tournaments, but IMP for challenges).

[barmar]

smerriman, on 2016-September-17, 18:57, said:

That's where the problem is. With an actual field, that makes perfect sense and MP is far better. But here there is no "field" - you're competing one vs one.

Right -- MP challenges are actually BAM, which is generally considered the toughest form of duplicate scoring. And GIB doesn't know that it's playing BAM, it's using normal MP strategy, which calculates expected field percentages.