BBO Discussion Forums: One sided points distribution - BBO Discussion Forums

Jump to content

  • 2 Pages +
  • 1
  • 2
  • You cannot start a new topic
  • You cannot reply to this topic

One sided points distribution One side gets the best hands

#21 User is offline   smerriman 

  • PipPipPipPipPipPipPip
  • Group: Advanced Members
  • Posts: 3,726
  • Joined: 2014-March-15
  • Gender:Male

Posted 2023-January-10, 18:48

Humans are well known to be extremely poor judges of whether something looks 'random' or not. There are some classic studies where people are asked to make up a "random" sequence of coin tosses - all of the results turned out extremely biased due to the humans almost always 'balancing out' heads and tails over short sequences, whereas in reality, random sequences are much more imbalanced, with long streaks occurring regularly.

Of course, who declares in bridge is not equivalent to flipping a coin. Some bidders are overly aggressive, and some overly passive.

According to MyHands, you played 14 hands yesterday, (not 18), and North did declare on 12 of them. But quite a few of these were a result of poor bidding. For example, on one hand, when you had bid to game in 4, North made an insane sacrifice in 4 (undoubled) with a 5062 10 count and his partner silent. Once North got to be declarer because South opened 1 with a balanced hand in their NT range - normal bidding would have left South declarer.

On another, you let North declare in 2 when your partnership had 26 high card points between you! (And there was another occasion where he was also left in 2 with your side having the bulk of the cards).

You can't do anything about an opponent overbidding to become declarer, but underbidding like that will definitely bias your results significantly towards defending much more than declaring, so it's something worth working on.

Yesterday, you didn't have E/W declare 21 times in a row; you're probably thinking about the session several days before that, where E/W declared 17 times in a row, when you were sitting N/S. Yes, you had a bad run of cards that session (though again there were definitely occasions of underbidding when you should have been declaring) - but nothing out of the ordinary at all.

Actually yesterday, North declared 7 times, South twice, East 6 times, and West 4. You'll always be able to find patterns if you only pick and choose the ones that don't look random to you.

You may think that the fact there is an occasional forum post about bias implies maybe there is something going on. But consider that there are 12000 players online right at this second.. one of those players right now is going through such a bad streak of cards that will only happen 0.008% of the time! If they're underbidding on top of that, they'll definitely think BBO is against them.
2

#22 User is offline   wuudturner 

  • PipPipPip
  • Group: Full Members
  • Posts: 80
  • Joined: 2011-November-07

Posted 2023-January-14, 15:01

View Posttadyan, on 2023-January-06, 18:27, said:

I absolutely agree with the point raised (one sided distribution). I play 25 casual boards twice a week with 3 friends. I would say in 90% of the games one side is favored. It is very annoying for the side which doesn't get points. I think the algorithm should distribute cards more evenly.


NO. You don't understand what random means. Almost ALWAYS, one side will be favored by some amount. Sometimes it is us, sometimes them. Sometimes more than others. The odds of an exactly equal spread of points WOULD be surprisingly low. And the code to do that would be nasty to write. For example, board 1, E-W get lucky, and pick up 35 points between them. Now, in order to even things out, the code needs to give N-S 35 points on the very next hand. And the code needs to remember who got how many points on all previous hands. And what if someone sits into a game midway through? The code needs to now figure out how to equitably distribute points to everyone so they are all perfectly happy they got an equitable distribution of points?

Or, suppose NS gets 35 points on board 1. Is it ok if EW now average 21 points over the next 15 boards?

If the algorithm did try to distribute cards perfectly evenly, then that WOULD be non-random. And it would be way harder to do well, whereas random sets of cards are trivially easy to generate. And BBO would have no reason to favor some people over others. OK, I take that back. I do think there is a trap in the code, that looks something like

if user = tadyan
generate hard hands.
watch'em sweat, while the devs laugh and laugh and laugh.
end

Not true of course.

Does it seem like the odds are always against you? YEP. You remember the bad days. They stick out. Do you complain about the good days? Of course not. You won! You are happy. Sadly, this is an argument nobody will ever win, because those who think they see a pattern against them will always remember those bad days more than the good days.
0

  • 2 Pages +
  • 1
  • 2
  • You cannot start a new topic
  • You cannot reply to this topic

1 User(s) are reading this topic
0 members, 1 guests, 0 anonymous users