Are all hands played 16 times?

[andych]

I have a hand played (duplicate IMP) in BBO. When I left BBO, the hand has been played 15 times and not shown in the myhands page. But could be seen in movie in BBO. It is now a few hours later and the hand is still not seen in myhands page.

 

Would all hands be played 16 times and flushed to myhands page? If not when does a hand last?

 

HAPPY NEW YEAR :P :lol: :lol:

[uday]

16 times, then flushed when the last table that is using the hand has closed. Sometimes a hand gets 'stuck' for a couple of hours, not completely sure why.

[andych]

I played a 14 boards match with another pair 2 days ago. The hand mentioned which was played 15 times when I logout seems has died. It is still not flushed to myhands page.....

Other than the 16 times is there also a time limit of the life span of a hand?

 

;) :)

[andych]

The hands are finally played 16 times after a day or 2.

But it happens quite often the hands are stucked. I see it again just now so posting again.

Any chance to look at it?

 

 

<_< :angry: :) :o

[cherdano]

16 times, then flushed when the last table that is using the hand has closed. Sometimes a hand gets 'stuck' for a couple of hours, not completely sure why.

It's not infrequent that for quite a while, the hand gets stuck with only one table having played it. (Usually when I have bid a nice grand slam with partner, so that we never get the reward until we quit the table :))

Not a complaint, just a note that users would notice if this bug got fixed.

 

Arend

[hotShot]

16 times, then flushed when the last table that is using the hand has closed. Sometimes a hand gets 'stuck' for a couple of hours, not completely sure why.

If it really needs the table to close, maybe some tables live that long, changing players and hosts all the time.........

[andych]

Noticed that whenever the table is not playing at norm pace, either too slow or too quickly, the hand stucking problem seems more easily occured.

 

Really would like to see it fixed.

 

 

:P :)

[uday]

There are several legitimate reasons why a hand might not be played out and sent off the the database right away. Too boring to explain, tho I will if anyone cares. I'll sum up most of the reasons by saying they are a side effect of the way I separate human deals from robot deals. If I merged human/robot results, things would flow more smoothly.

 

One of the reasons might be a bug in the code. The next time a hand is stuck, give me enough information to identify the hands and i'll take a look. What i'd like is

 

Location (MBC? Tourney #? Team #/host )

table host username

Board number

contract achieved at your table

approximate date/time played (preferably in EST)

 

This is enough information for me to track down the deal, even retroactively, and find out why it is stuck, and where it is stuck (in BBO? In transit to myhands?)

 

Even completely stuck deals will eventually unstick, tho it might take a "maintenance" bounce to do this (we currently bounce once every few weeks or so).

[andych]

Board stucks again :( :(

Location Public Bridge Club (Rays bridge ladder) = MBC?

Table host andych

Board no. 13

contract 4♥ by E (making +3)

Approximate date time 5Feb2006/10:16AM EST

 

:( :P :( :D :( :(

[andych]

The hand got out finally after sucking for an hour.

Sometimes it doesnt take only an hour but a day!!

 

Hand played at my table 10:16AM EST.

Last played at 11:13AM EST.

 

Hand record here

http://online.bridgebase.com/myhands/hands...username=andych

 

Would be very grateful if it could improve. Also please let us know why it happens. Anyway to avoid?

 

 

:P :D :(

[jchiu]

Rather than doing further complaining that hands do not show up at lightning

speed on "myhands", I would like to ask whether the number of times a hand is

played before retiring could be increased.

 

Increasing this number improves accuracy. According to the central limit theorem,

the accuracy improves according to the square root of the number of independent

trials. My experience indicates that performance is much more closely reflected in

larger matchpoint tops. Likewise, an unusual result (you know, the guy who goes

for 1400 on a partscore deal) skews the IMP results on BBO by approximately 2-3

IMPs. I have seen many hands such that (1) achieving the normal result gives the

pair going plus a premium of about 2.5 IMPs and (2) had the exact hand been

played in a national IMP pairs game, the premium would be less than 0.25 IMPs

per table. A disproportionate number of these are due to the skew results.

 

There are several benefits to the accuracy of results, since it would help

- improving players who consider their main bridge club results seriously

- regular set games where players keep score for fun

- irregular set games where the players play for (say, $3) per crossimp

 

Jason

[cherdano]

Jason, I think the main problem is the quality of the field, not the sample size. If you bid a 33 hcp 6NT where the bidding doesn't start with 1NT, you will always earn 3-5 IMPs in the MBC. Same if you bid bid an obvious 23hcp major suit game.

 

In fact, MBC is a bit like rubber bridge. With good hands, you usually go plus.

 

Arend

[Sigi_BC84]

I am with Jason when he's saying that having the hands played more often will be beneficial (I'm also with Arend regarding getting a borderline hand will give you a good result if you bid it properly).

 

My suggestion: the possibilty of enabling "Butler scoring" for my own table, so that the most extreme results are disregarded for calculation of the base score.

 

I'd also like to have that option for tournaments.

 

Is there any chance we will see Butler scoring in the near future?

 

--Sigi

[Gerardo]

Being able to Cross-IMP, what would you want Butler for?

[Sigi_BC84]

Being able to Cross-IMP, what would you want Butler for?

For the reason I (and Jason) already mentioned: Using Butler, the extreme results are not going into the calculation of the IMP base, so if somebody puts a -7600 up or something like that, this does not skew the results of all the hands.

 

Or can one tweak the existing Cross-IMP setup accordingly (at least for private tables I haven't seen such an option)?

 

--Sigi

[hrothgar]

Jason, I think the main problem is the quality of the field, not the sample size. If you bid a 33 hcp 6NT where the bidding doesn't start with 1NT, you will always earn 3-5 IMPs in the MBC. Same if you bid bid an obvious 23hcp major suit game.

 

In fact, MBC is a bit like rubber bridge. With good hands, you usually go plus.

 

Arend

I think that Jason raises a valid point:

 

From a sampling perspective, 16 boards is a relatively "small" number. You're still able to get significant improvements in accuracy adding a small number of comparisons.

 

Increasing the number of comparisons from 16 to 32 would be a big improvement. Increasing the number of comparisons from 116 and to 132 would be almost meaningless.

[coyot]

I think it would be much easier to implement Butler than to have the hands played by 16 more pairs - looking from the GUI perspective. Those 16 results almost fit on the screen - and using 12 of them (after scratching two extremes on each end from the average) should be pretty OK.

 

Sure, playing each board 32 times would be possibly even better, but, then, with 32 tables, the likelihood of some moron claiming 7NTxx-13 is much bigger...

 

In my experience with online bridge, on almost every board you will see one or two overbidders, which means that with 32 tables, you'd probably see three or four of them. Butler eliminates this problem and I think that 12 results are a good basis for comparison.

 

(It will surely completely eliminate the occasional pairs who hone their virtual egos by combined-hands bidding and doubledummy playing using ICQ or Skype).