Help
Perhaps I'm in the wrong place for this, if so let me know.
When playing in robot tournaments, sometimes a single board is bid differently by the robot.
For example, in an Instant Touranment, on this sequence, 1H - 2D - P - 2NT - P
14 of the 15 robots bid 3D. The one board where the robot bid 3NT made a bunch of IMPS.
It would seem that when one plays robots, since the idea is that you are competing against other HUMANS, the robots wouldn't be that inconsistent.
For example, if it is a 40% - 60% I totally understand the robots bidding (and the same thing applies to playing) differently and there is a range of scores.
But when only one robot has a different bid, it tends to make for crazy results. Is that a "feature" to simulate the outlier bids you find in the real world? That seems contrary to the idea having a list at the end of how one stacked up against other humans.
What would be the negative to having similar bidding from robots unless a certain level of uneveness (say 25-30%) was reached?
Thanks for your patience,
Jay
Page 1 of 1
Robot Tournaments: One board bid differently than all the others
#2
- Group: Admin
- Posts: 21,584
- Joined: 2004-August-21
- Gender:Male
Posted 2015-June-21, 15:43
When you play an Instant Tournament, you're replaying a tournament that was originally played some time ago. Since then, the robots have been upgraded several times. All the robots at your table are running the current version of GIB, and they'll bid differently from the robots that played in the original tournament.
We're working on improving this situation, so we'll be able to use more recent tournaments that were played with the same version of GIB.
We're working on improving this situation, so we'll be able to use more recent tournaments that were played with the same version of GIB.
#3
- Group: Advanced Members
- Posts: 6,542
- Joined: 2010-February-01
- Gender:Male
- Location:Brooklyn, NY, USA
Posted 2015-June-21, 15:56
I actually kinda like the "earlier version" feature; it allows us to see the effects of modifications to GIB.
#4
- Group: Full Members
- Posts: 83
- Joined: 2010-October-29
Posted 2015-June-22, 01:44
"More recent tournaments that were played with the same version of GIB" would hopefully remove the random results - tops and bottoms and make these a truer reflection of skill. It cannot come soon enough for my liking! What sort of time frame are we looking at I wonder?
#5
- Group: Members
- Posts: 2
- Joined: 2013-August-12
Posted 2015-July-04, 13:30
Thank you for the reply. That makes sense. So if my board is bid differently than the other 15 boards, it is because GIB is bidding that situation differently than it had in the past. Having Instant Tournaments all on the same version of GIB would be useful IMO.
#6
- Group: Admin
- Posts: 21,584
- Joined: 2004-August-21
- Gender:Male
Posted 2015-July-04, 21:35
Nitschke, on 2015-July-04, 13:30, said:
Thank you for the reply. That makes sense. So if my board is bid differently than the other 15 boards, it is because GIB is bidding that situation differently than it had in the past. Having Instant Tournaments all on the same version of GIB would be useful IMO.
We're working on it.
#7
- Group: Admin
- Posts: 21,584
- Joined: 2004-August-21
- Gender:Male
Posted 2015-July-20, 09:31
Nitschke, on 2015-July-04, 13:30, said:
Thank you for the reply. That makes sense. So if my board is bid differently than the other 15 boards, it is because GIB is bidding that situation differently than it had in the past. Having Instant Tournaments all on the same version of GIB would be useful IMO.
We've implemented this change; since a few days ago, instant tournaments should be using only hands played with the current version of GIB (unless you somehow manage to play all of them).
Whenever we deploy a new version of GIB, it will take a couple of weeks for us to collect enough tourneys to start using them as the basis, so during that period it will continue to use the previous version. But it should never get older than that.
Page 1 of 1