Help
Playing on robot challenges or daylong tourneys, I experienced extremely biased distributins. 5-0, 4-0, 4-1 trump distributions are more common than normal, theoretically more probable 3-2 or 4-2 distributions. Why don't the BBO cdealer fulfill the statisticacs? Please, do not explain me the 'best hand' mechanist! That is not acceptable, that 5-0 and 4-1 distribution are 10 times frequent than 3.2.
Page 1 of 1
Biased distributions playing with robot
#2
- Group: Advanced Members
- Posts: 4,683
- Joined: 2014-March-15
- Gender:Male
Posted Yesterday, 13:56
You are wrong. This has been proven beyond doubt countless times by looking at hand records and actually counting them, vs human 'perception' which is highly flawed.
#3
- Group: Advanced Members
- Posts: 5,275
- Joined: 2008-September-10
- Gender:Male
Posted Yesterday, 17:10
Great point fimre!!! The reason you are getting bad breaks and finesses that don't work is that this is a way that BBO balances distributions between average members and the Super Gold Platinum BBO Executive Patron Club members. The Patron club members never get bad trump splits, and their finesses work about 98% of the time. In order to balance the overall percentages, BBO stacks the dealing programs so that the overall breaks and finesses work out according to what the statistics predict.
The only way to avoid the bad breaks and losing finesses is to join the invitation only Super Gold Platinum BBO Executive Patron Club. I've been trying to get into this club for years but haven't had any luck so far.
The only way to avoid the bad breaks and losing finesses is to join the invitation only Super Gold Platinum BBO Executive Patron Club. I've been trying to get into this club for years but haven't had any luck so far.
#4
- Group: Members
- Posts: 8
- Joined: 2009-December-13
Posted Today, 04:57
smerriman, on 2025-September-21, 13:56, said:
You are wrong. This has been proven beyond doubt countless times by looking at hand records and actually counting them, vs human 'perception' which is highly flawed.
I hardly think that I'm wrong. I made a refined counting during the play: counted ALL the distributions occured and counted when finesses succeeded or when not. The unsuccessful finesses were 3 times frequent than successful ones. 5-0 and 4-1 distribution were 10 times frequent than 3-2. The theoretical values are about 4%, 28% anmd 68%, respectively. Even the 2-0 occured more than 1-1. I'm a seasoned mathematican with deep knowledge of probability theory and statistics. The only reason for the many biased distributions is that the BBO dealer 'selects' the flat distributions for the paying members. The rest is for Free Daily Tornaments and other 'free' tourneys. So the overall statistical balance became right. These are behind the 'proof' you've mentioned.
Other argument for my statement is, that Free Weekly Instant Tournament has rarely bad breaks and NEVER when a slam is occures. That FWIT is a bait for the payed services of BBO. Nothing wrong with it but they should warn members, that free tourneys are more 'goulash-like' than ordinary bridge.
#5
- Group: Members
- Posts: 8
- Joined: 2009-December-13
Posted Today, 04:59
johnu, on 2025-September-21, 17:10, said:
Great point fimre!!! The reason you are getting bad breaks and finesses that don't work is that this is a way that BBO balances distributions between average members and the Super Gold Platinum BBO Executive Patron Club members. The Patron club members never get bad trump splits, and their finesses work about 98% of the time. In order to balance the overall percentages, BBO stacks the dealing programs so that the overall breaks and finesses work out according to what the statistics predict.
The only way to avoid the bad breaks and losing finesses is to join the invitation only Super Gold Platinum BBO Executive Patron Club. I've been trying to get into this club for years but haven't had any luck so far.
The only way to avoid the bad breaks and losing finesses is to join the invitation only Super Gold Platinum BBO Executive Patron Club. I've been trying to get into this club for years but haven't had any luck so far.
Unfortunately, I've nevr heard about that club. Where could I get informed about that?
#6
- Secretary Bird
-
- Group: Advanced Members
- Posts: 8,066
- Joined: 2003-July-12
- Gender:Male
- Location:Calgary, D18; Chapala, D16
Posted Today, 12:29
I'm sure you're correct. I'm also sure that you have the documented evidence of it you can show us which will clearly show a difference between the hands you're getting and all the studies that have been done and documented (hint: the BBO dealer can't generate 100% of hands due to size constraints. However, it has been shown that that doesn't produce a noticeable bias in honour location or break probability compared to 'all possible hands').
I am sure of this because the other 20 people (at least) that have claimed a bias on these forums (always on "finesses lose and suits break badly" for some reason. And they don't seem to be happy when it's to their advantage when they're defending, oddly enough) have been able to show that documentation.
Oh, I'm sorry, I meant *none of the other 20 people (at least)*. You'll be the first.
I am very much looking forward to the evidence.
I am sure of this because the other 20 people (at least) that have claimed a bias on these forums (always on "finesses lose and suits break badly" for some reason. And they don't seem to be happy when it's to their advantage when they're defending, oddly enough) have been able to show that documentation.
Oh, I'm sorry, I meant *none of the other 20 people (at least)*. You'll be the first.
I am very much looking forward to the evidence.
Long live the Republic-k. -- Major General J. Golding Frederick (tSCoSI)
#7
- Group: Advanced Members
- Posts: 4,683
- Joined: 2014-March-15
- Gender:Male
Posted Today, 15:26
fimre, on 2025-September-22, 04:57, said:
I hardly think that I'm wrong. I made a refined counting during the play: counted ALL the distributions occured and counted when finesses succeeded or when not. The unsuccessful finesses were 3 times frequent than successful ones. 5-0 and 4-1 distribution were 10 times frequent than 3-2. The theoretical values are about 4%, 28% anmd 68%, respectively. Even the 2-0 occured more than 1-1. I'm a seasoned mathematican with deep knowledge of probability theory and statistics.
Fantastic! Being a seasoned mathematician makes this far easier; most people just go on a gut feeling and when confronted with the actual numbers, ignore them and continue believing their false narrative.
Here are your last 12 daylong tournaments:
Free Super Sunday Daylong (Sep 21)
Free Weekend Survivor - Day 1 of 2
Free Just Declare Daylong (MP) - 2025-09-19
Free Daylong Tournament (IMP) - 2025-09-17
Free Super Sunday Daylong (Sep 14)
Free 2 Day Weekend Event (Day 2 of 2)
Free 2 Day Weekend Event (Day 1 of 2)
Free Daylong Tournament (MP) - 2025-09-11
Free Daylong Tournament (IMP) - 2025-09-10
Free Daylong Tournament (MP) - 2025-09-09
Free Super Sunday Daylong (Sep 7)
Free Daylong Tournament (IMP) - 2025-09-03
During these 112 deals, N/S received a total of 77 8 card fits. These broke as follows:
3-2: 51
4-1: 25
5-0: 1
Hm, that's odd - they seem to line up with what you would expect with a random distribution. But let's assume your hypothesis is correct, and that 5-0 and 4-1 occur ten times as often as 3-2.
This means for an individual 8 card fit, you should receive a 3-2 break 1/11 = 9.0909% of the time, and a 4-1 or 5-0 10/11 = 90.9090% of the time.
If this hypothesis is true, you can plug this into a binomial distribution to give the chance of receiving at least 51 successes from 77 samples where p = 1/11. A binomial distribution isn't 100% accurate, as the same pair of hands can have two 8 card fits, which aren't entirely independent - but the difference is completely negligible.
I may have a precision error, but I make this approximately 0.000000000000000000000000000000000007912408136921205.
I think that's *just* low enough to reject the null hypothesis.
Page 1 of 1