[uday]

A user suggests that the hands recently ( since approximately our last power-down of the server itself) have been much wilder than usual.

I've heard no other complaints to this effect. Anyone else feel this ?

[inquiry]

A long time ago, I questioned the randomness of BBO... after a couple of 76, 66, and a 94 hand. But recently, the hands seem more or less normal to me. I guess it is the luck of the draw... thousands of people playing at one time, some group of us will get a wild hand....

Ben

[ack_hh]

What about something like "www.bridgebase.com/statistics" which will summarize the HCPs dealt to N/S/E/W, number of voids, stiffs, suit length, part scores, doubles, slams, grands, etc. etc. per week, month and year?

I guess the data is all there - it "only" needs to be collected.

That way BBO and its members can track the quality of the deals, and a bit of statistics can be fun too

Andreas

[Maaa]

There is currently an equal topic at the polish forums. Some players asking what kind of generator BBO uses... claiming that the hands are wild, that they see 6-6 or 7-4 distributions and a big number of sequences almost every day, while that does not happen such often at manually dealed cards.

Besides of that heard a few complains lately - "having cards" in the main bridge club belongs at what line you sit.

So YES - have heard of it.

I like the idea of collecting and posting statictics from time to time

[bearmum]

I played on another site for quite a while and DID keep SOME statistics ( as I was convinced that I was ALWAYS on the "wrong side" when it came to points - and GUESS what - over a six month period when I was playing an average of 300-400 hands a month the incidence of having an opening bid - AND partnership having at LEAST 21 points ( I KNOW -- a bit crude BUT I am not a computer whiz ) came out really CLOSE to 50% for BOTH --- a SLIGHT bias to me when playing N/S --- but probably explained from the small sample of hands --- so hopefully BBO using the same sort of hand generater programme that other site used. BTW our LOCAL club uses a hand generator programme - which OFTEN seems to have one 36 board game with LOTS of wild distributions --- and others with TAME ones

[Rado]

Hi all friends,
The main reason that people feel the computer dealt board are more "WILD" than hand ones is the 3-4 shuffles usually made at live play. Some mathematician sproved that for really "RANDOM" deal at least 7 shuffles must be done. That's why when playing home bridge we are used to more 4-3-3-3 hands tha to 7-6-0-0.
Regards
Rado

[bearmum]

Rado, on Jan 30 2004, 12:15 AM, said:

Hi all friends,
The main reason that people feel the computer dealt board are more "WILD" than hand ones is the 3-4 shuffles usually made at live play. Some mathematician sproved that for really "RANDOM" deal at least 7 shuffles must be done. That's why when playing home bridge we are used to more 4-3-3-3 hands tha to 7-6-0-0.
Regards
Rado

Interesting you should say that Rado - at our f2f club we sometimes have to actually deal the cards to play if too many boards in duplicate with "computer" dealt hands- AND have found that there are REALLY WILD distributions when only a couple of reshuffles done

[Trpltrbl]

I like wild hands but haven't seen any lately.

Mike

[TimC1982]

How does the random generator BBO uses work? If it uses information from the CPU clock cycle (of the host or one of the other players), there may be a side effect such as users with CPU's that have longer clock cycles seeing more random hands. I personally haven't experienced this, but I haven't been playing as often as I used to, and I am also on a very fast computer.

[Gerben47]

It's all relative. Last club night in a set of 26 hands I had TWO hands with 7-5 distribution. The first hand went 1♦ on my left I had 5♥+7♣, partner made 12 tricks in 5♥, unfortunately undoubled as some other tables did.

Then a few boards later I could not believe my eyes when I was dealt
♠- ♥Qxxxxxx ♦KJxxx ♣x

Another 7-5. All hand dealt! Anyway partner opened 1♠, I bid 1NT, and partner bid 2♦! Side question: How many diamonds do you bid?

Anyway, that proves these kinds of distribution happen in hand-shuffled tourneys as well. So there's nothing to worry about. And the only way to check is to keep statistics of many, many hands.

[helium]

Gerben47, on Feb 28 2004, 04:57 PM, said:

It's all relative. Last club night in a set of 26 hands I had TWO hands with 7-5 distribution. The first hand went 1♦ on my left I had 5♥+7♣, partner made 12 tricks in 5♥, unfortunately undoubled as some other tables did.

Then a few boards later I could not believe my eyes when I was dealt
♠- ♥Qxxxxxx ♦KJxxx ♣x

Another 7-5. All hand dealt! Anyway partner opened 1♠, I bid 1NT, and partner bid 2♦! Side question: How many diamonds do you bid?

Anyway, that proves these kinds of distribution happen in hand-shuffled tourneys as well. So there's nothing to worry about. And the only way to check is to keep statistics of many, many hands.

may i ask whay system youm play? dont understand the 1 nt bid

[the hog]

"may i ask whay system you play? dont understand the 1 nt bid"

♠- ♥Qxxxxxx ♦KJxxx ♣x
What else would you bid apart from 1N holding that hand? You can't bid 2H as that grossly overstates the high card content of the hand and forces the bidding almost to game on what might be a huge misfit.

[inquiry]

helium, on Feb 28 2004, 07:54 PM, said:

Gerben47, on Feb 28 2004, 04:57 PM, said:

♠- ♥Qxxxxxx ♦KJxxx ♣x
Anyway partner opened 1♠, I bid 1NT,

may i ask whay system youm play? dont understand the 1 nt bid

They are playing 2/1 Game force. Take a look on BBO library at the system notes for BBO Advanced (which is a 2/1 GF system) for more on forcing 1NT after 1 of major opening bid.

[JRG]

Using time, etc., when random dealing hands is normally only used to establish the "seed" for a pseudo-random number generator. If the generator is good and the shuffling alogrithm is also good, then it shouldn't matter.

However, I do know that some random number generators put restrictions on the "seed" if you want the generator to perform well. So, how it is seeded perhaps should be checked.

The other thing is there are statistical tests (I'm not knowledgeable about them, but I know they exist) that can be run to check the "randomness" of a random number generator.

I suspect Fred or Uday have run such tests (if not, I recommend they do so). They might also be wise to doublecheck the seeding. The server is a Unix variant and the kernel provides a seed for this purpose (saving information when it is shutdown so that the same sequence is not generated a second time).

I believe there is some kind of statistical testing that can be done directly on the generated hands (to check for "randomness"); however, I suspect it requires some programming. I would imagine several tens of thousands of hands would need to be generated to get a good measure (again, I'm not a statistician, but this has something to do with the "confidence" in the results). I may be greatly over-estimating, but I'm sure 20 or 100 hands are grossly inadequate for testing.

[luis]

Gerben47, on Feb 28 2004, 09:57 PM, said:

It's all relative. Last club night in a set of 26 hands I had TWO hands with 7-5 distribution. The first hand went 1♦ on my left I had 5♥+7♣, partner made 12 tricks in 5♥, unfortunately undoubled as some other tables did.

Then a few boards later I could not believe my eyes when I was dealt
♠- ♥Qxxxxxx ♦KJxxx ♣x

Another 7-5. All hand dealt! Anyway partner opened 1♠, I bid 1NT, and partner bid 2♦! Side question: How many diamonds do you bid?

Anyway, that proves these kinds of distribution happen in hand-shuffled tourneys as well. So there's nothing to worry about. And the only way to check is to keep statistics of many, many hands.

You've just violated one of my partneship golden rules:
"Never bid 1NT with a 7-5 hand"
And never means NEVER.

[Gerben47]

Same club yesterday. You get dealt
♠QJTxx
♥x
♦AKQxxxx
♣-

Sigh... That's three in my hand in 52 hand dealt boards. I'm going to suggest computer dealt boards, at least you get normal distributions with them! Yes, I checked that people were not dealing 5-4-4...

About not bidding 1NT with 7-5 distribution: what else would you suggest on this hand? 2♥ overstates your high card potential by a mile, even if it shows only 10 HCP (or a good 9).

[McBruce]

Isn't there a recorded instance from the early-mid 90s somewhere, where Fred discovered that the deals being used in a major Canadian championship were not random? Something about mirror images and the same suit combinations right down to the spot cards on deals in two consecutive sessions. I vaguely recall reading this in Canadian Master Point sometime around 1992-94. Maybe Fred has the article somewhere and can post it for us in the BBO library; it was quite entertaining.