[fred]

A couple of new ideas:

1) Run MBTs where the human (South) is always dealt the "best hand" at the table.

2) Run MBTs where the South hand on each board is the same at every table, but the other 3 hands are randomized.

What do you think?

Fred Gitelman
Bridge Base Inc.
www.bridgebase.com

[MickyB]

I don't like the first idea - it will encourage quick play to an even greater extent. I really like the second idea though.

[hotShot]

The GIB'S play at a good speed, and the setup is fun.
Just a few remarks what can be better.

Let us assume that 10 boards are played in 25 minutes, then there would be an average of 4 partscores, 5 makeable games and a makeable slam. (Note that makeable does not mean it can be bid!)
In half of the cases the slam is on GIB's side, those players have no chance to win. So luck has a big role in this kind of tourney.
This can be reduced a lot, if the worst result of a player would be canceled or if only the 10 best hands would be used.

If i get a good score e.g.
- slam/grand in the first board or better at the first board vul.
- make 2 games in a row
my best strategie is to stop playing. This way i will be better as those where the GIB's have a slam or a full game more than the human (about 6-7 of 10).
There should be regulation against stopping to play.

[Poky]

fred, on May 5 2006, 12:19 AM, said:

A couple of new ideas:

1) Run MBTs where the human (South) is always dealt the "best hand" at the table.

2) Run MBTs where the South hand on each board is the same at every table, but the other 3 hands are randomized.

Although I think both ideas are worth to implement, the 2nd idea (same hand for all Souths) sounds much better to me.

The only better idea I can see is to implement "tournaments" (BFM area) at two tables with same hands where one player can compete against another. Same hands, same vulnerability, same position, same opponents, only skill makes the difference...

[uday]

Quote

two tables with same hands

People will cheat

[barmar]

hotShot, on May 5 2006, 02:59 AM, said:

If i get a good score e.g.
- slam/grand in the first board or better at the first board vul.
- make 2 games in a row
my best strategie is to stop playing. This way i will be better as those where the GIB's have a slam or a full game more than the human (about 6-7 of 10).
There should be regulation against stopping to play.

Maybe have a minimum required number of boards played in order to scratch.

I also think the idea of using the best N boards should be given some thought.

[Poky]

uday, on May 5 2006, 03:52 PM, said:

Quote

two tables with same hands

People will cheat

People could cheat if 3+ tables play the same board, but not if only two tables are involved.

Instead of playing with GIB as partner against two GIBs, I find much more attractive (and fair) to play with GIB against 2 GIBs, comparing my own score against another player in same position with same cards.

Why is this so difficult to implement?

[Echognome]

I also enjoyed the tournaments (I liked the speedball aspect of it). However, I also felt the luck factor was too high. In the first tournament I played, it went GIB bidding game game slam game slam. Four of these made and one of the slams on a finesse. There was no chance to recover after that.

So getting rid of the luck factor would help. I am just throwing around some ideas here, and I do like that cheating is not possible. However, what if you just made cheating really hard? Here's what I had in mind:

Suppose that you made the mbts as fixed board rounds with time pressure? So everyone played the same hands, but you were given say 3 mins a board, one board rounds, say 10 boards. So yeah, two people on MSN might be able to say "the finesse is off" or blah blah, but it would be difficult on most hands for them to collaborate. And since no one can start the next board until everyone is finished, they cannot race through to find out the hands. Now I'm not sure whether 3 mins/board or 2 mins/board or whatever is best for this. Of course at 2 mins/board you will have tremendous time pressure already. So this method will not eliminate cheating alltogether, but it will certainly make it very difficult to cheat and, at the same time, eliminate the luck factor.

Just a thought.

[inquiry]

Poky, on May 5 2006, 11:07 AM, said:

uday, on May 5 2006, 03:52 PM, said:

Quote

two tables with same hands

People will cheat

People could cheat if 3+ tables play the same board, but not if only two tables are involved.

Instead of playing with GIB as partner against two GIBs, I find much more attractive (and fair) to play with GIB against 2 GIBs, comparing my own score against another player in same position with same cards.

Why is this so difficult to implement?

Team Game Money Bridge... No human at your table.... but other human holds exactly same cards as you. The problem would be speed of play. What if you are two tables ahead and other fellow gets disconnected.

However, in theory, I like this concept. No unlucky 7NT bid against you vul and you get zapped for a ton of money.... you have a chance to bid the same 7NT...

:-)

[uday]

Quote

So everyone played the same hands

Even an occasional edge ( easy to achieve here) will give the perp-pair a huge advantage over the civilians.


Things I'm sort of thinking about are: longer tourneys (flattens out HCP better), imping results against old hands from the MBC, using russian scoring.

We'll see. We'll start as is, try longer Ts next (they are trivial to deploy, unlike the other two options), maybe try russian scoring next.

[barmar]

How are people playing so quickly? The thing that I find that slows it down is the inability to claim. I spend a good 30 seconds playing out hands that should be claimed. With hands that can be claimed early, GIB can spend a while thinking about its discards that I know don't matter.

When playing against GIB in the MBC I thought there was a claim button, why isn't it available in MBT?

[asdfg2k]

I think I'm running an average of 14 or 15 hands in the 25 minute period, unless I have connection problems. I think I could move that to 16 or 17 if GIB didn't hesitate when deciding what to play from 432 of a suit sometimes.

I don't stress over setting an extra trick or giving an overtrick. Decide in advance what you are going to do and then do it. Don't even think about the hand you just finished playing until the 25 minutes are up - what is done is done.

I have won 2 out of the 6 or 7 (8?) that I have entered, in case that matters.

[fred]

barmar, on May 5 2006, 07:59 PM, said:

How are people playing so quickly? The thing that I find that slows it down is the inability to claim. I spend a good 30 seconds playing out hands that should be claimed. With hands that can be claimed early, GIB can spend a while thinking about its discards that I know don't matter.

When playing against GIB in the MBC I thought there was a claim button, why isn't it available in MBT?

When you play against GIB in the Main Bridge Club, there is a copy of GIB running on your hard disk that does that thinking. It is relatively simple for the BBO client to interact with the local copy of GIB in terms of claims.

It is more complicated when the GIB is running on the server (as it does when you play in MBTs). I suspect we will do the necessary work before too long, but most likely it won't happen for the first release of the MBT-enabled client.

Fred Gitelman
Bridge Base Inc.
www.bridgebase.com

[dogsbreath]

hi
I've tried several of the introductory tourneys and the sofware seems to work fine.
However , i just dont see the point .. total points with everyone playing different hands seems a complete lottery and will not induce me to pay to play .. sry i may be missing something but it seems completely pointless to me.

Rgds Dog

[fred]

I think you are wrong about it being a complete lottery.

Luck is important for sure, but it tends to even itself out in the long run.

Skill will always prevail and there is a lot of skill to this game.

Another way to look at this is that a strong player's expected return on his $1 (or whatever) entry fee is much higher than that of a weak player.

MBTs may seem pointless to you, but I think many people will find these tournaments to be fun and exciting. The tournaments will not seem pointless to them regardless of what they think of the relative importance of luck and skill.

Fred Gitelman
Bridge Base Inc.
www.bridgebase.com

[jchiu]

barmar, on May 5 2006, 07:59 PM, said:

How are people playing so quickly? The thing that I find that slows it down is the inability to claim. I spend a good 30 seconds playing out hands that should be claimed. With hands that can be claimed early, GIB can spend a while thinking about its discards that I know don't matter.

When playing against GIB in the MBC I thought there was a claim button, why isn't it available in MBT?

How To Play (Insane) Speedbridge:

1. Given a new hand, just approximate its HCP value and bid. Missing the target by one or two HCP is not a big deal usually, but be careful to not miss by more. Then, when the auction goes around the table, actually count the HCP.

2. After your second turn, plan out the rest of the auction if necessary, while waiting for the bid to come around a third time. For most experts and most hands, the first two actions are sufficiently automatic that they should be instantaneous. Use the in-between time to decide your next move (like in chess), so that all bids are limited only by your reflex time.

3. When GiB-GiB (E/W) are in an uncontested auction, and you do not plan to enter or make a lead-directing double, keep the mouse over "pass". Keep passing like a good puppet, then review the auction at the end. Half of the time, you will need to look at it before your lead. The other half, use GiB's lead tanking time at the lead to actually look.

4. Roughly plan your play at trick one. Usually this takes the longest (1.5 to 1.75 seconds) for me, but saves time later. Continue planning the play when GiB is thinking and you are not on lead.

5. Play suits in "beginner order". Make sure you leave them free enough that you can play one card every 0.25 to 0.40 seconds later after you would claim otherwise. Ruff winners if necessary to make sure all tricks in one hand are played and then the other. This way, you can randomly click cards from the other hand without blowing tricks.

6. Declare, declare, declare. Besides the obvious importance of high cards, this is the main reason you want big hands. GiB is a decent and slow declarer, but can be a sloppy and fast defender often.

7. Never post-mortem. It just frustrates you and slows you down.


But I digress into real bridge advice .... Using these techniques, I am averaging about 88 seconds per hand, with some records of about 81 seconds per hand (computed over 25-minute sessions). It takes practice to play well at this speed, but the law of averages definitely favors good players who are faster.

[uday]

We've moved the schedule around and added "pay" mbts

We will be running free mbts every hour, 15 minutes past the hour
We will be running pay mbts twice an hour : at :00 and :30

Initially, the pay mbts will cost $1 and return 80%. The logic for prize division is approximately this:

Prize pool = 80% of all card fees.

Winner gets 50% of prize pool.

Next place gets half that ( but minimum must be the card fee -- $1)
Next place gets half that ( but minimum must be the card fee -- $1)
etc

if "half that" is not enough to make up a card fee, it is added to the winner's prize.



I welcome suggestions for division of prizes ( we can also let winner take 100% of pool, with no one else getting anything ). We are also interested in opinions about the 80% prize pool. Is this too high? Too low ? Just right ? Finally, is there any interest in pricier tourneys (say, a $5 card fee ) ? Should the prize pool percentage go up/down/be-unchanged if the fee goes up ?

[barmar]

More expensive pay tourneys should last longer -- you're risking more money so you'd like the luck factor to be less significant.

Which reminds me. Does it make sense that among regular tourneys I often see both 6-board and 12-board tourneys charging the same $1 entry?

[Deanrover]

You should definitely offer higher buy ins. If you enter a $1 then you are looking at winning $3 net at the most....not that thrilling for most of us. I'm sure you will eventually get business for $100 tournaments (and likely higher too).

I think the % cut you take should go down as the buy in goes up.

FWIW the schedule of most poker rooms is something like

1 + .2
5 + 1 or .5
10 + 1
20 + 2
30 + 3
50 + 5
100 + 9
200 + 15
500 + 30

where the first figure represents the amount that goes to the prize pool, and the latter the amount that goes to the house

[jchiu]

uday, on May 6 2006, 04:54 PM, said:

Prize pool = 80% of all card fees.

Winner gets 50% of prize pool.

Next place gets half that ( but minimum must be the card fee -- $1)
Next place gets half that ( but minimum must be the card fee -- $1)
etc

if "half that" is not enough to make up a card fee, it is added to the winner's prize.



I welcome suggestions for division of prizes ( we can also let winner take 100% of pool, with no one else getting anything ). We are also interested in opinions about the 80% prize pool. Is this too high? Too low ? Just right ? Finally, is there any interest in pricier tourneys (say, a $5 card fee ) ? Should the prize pool percentage go up/down/be-unchanged if the fee goes up ?

Raking 20% is very high compared to a minilimit poker table which rakes about 6-10% (I think, the more serious poker players could check this). However, the cheap entry definitely justifies making sure BB keeps a few cents. Right now, the tournaments are two to four tables, and it is fine to rake 20%. However, with larger events, I would suggest raking 20% of the first (five, ten) tables, and less for every table thereafter.

Pricier tourneys would make it more worthwhile to play. I would think many more players would want to put down $20 into a two-hour event (say four sessions of 25 minutes and 5 minutes). The rake here can be considerably less, say 5%, and BB would end up making more per tournament.