World Computer-Bridge Championship, XII

Las Vegas, Nevada, USA

July 20-25, 2008

COMMENTS

For the history and results of all the championships, including articles, reports and photos, go to www.computerbridge.com  or go to www.ny-bridge.com/allevy/computerbridge/wcbc2008.html

These championships are meant to publicize our game, promote the development of bridge software, highlight robotic play and add another dimension to bridge players experiences.  To that end, they are very successful.

The headlines... Wbridge5 has repeated as champion, defeated Jack in their final KO, 172-157.

Overview... The robotic play went along smoothly over the six days.  Play was generally good, the Individual was a lot of fun, and the championship was, as usual, very exciting.  Hits on our website (checking for results) were at 300 a day, lower than recent championships held outside North America.  The Daily Bridge Bulletin covered our activity, and a few kibitzers came to observe and comment.

Individual event...The first-time Individual was a lot of fun, with results bouncing all over the place.  When the robot partners aren't fully in tune with each other, odd things can happen.  Since the Convention Card was not totally complete, i.e., there were undiscussed sequences, even though the system played was relatively simple, there were misunderstandings.  If two strangers sat down and played a simple card without any discussion, they would draw on their experience when they weren't sure of their agreement.  For example, they would assume that fourth suit was forcing, at least for one round.  The robots don't have any experience to draw on, so if they forgot to mark fourth-suit forcing, they might, and did, pass.  Interestingly, the final results were extremely close, with only 5.5% difference between first and last place.  This suggests that the results were random, matching the randomness of the play.  If this experiment is to be repeated next year, more thought must go into a complete system and a complete convention card.

Estimate of relative performance... Wbridge5 and Jack have won the past eight championships.  Since 2001 Jack has won five times (2001-2004, 2006) and Wbridge5 three times (2005, 2007-2008), with second going twice to Jack, once to Wbridge5, three times to Bridge Baron and twice to Micro Bridge.

In this year's championship, Jack outscored Wbridge5 in their head-to-head matches (32 boards in the round robin and 64 boards in the final) 237-216.  Further, Jack outscored Wbridge5 in matches against the other robots by 131-110 VPs.  Jack performed better than Wbridge5, considering the robots aren't influenced by the importance of a board or the conditions under which a board or match is played.  Wbridge5 performed better when it counted most, which for robots is a random event.

Randomness clearly plays a role in the short-term outcome.  For a best estimate of relative strength, 1,000+ board matches would be informative.  The World Computer-Bridge Championships, similar to human championships, aren't meant to determine who is best, only to crown a champion after much play.  Nevertheless, recent performance is an indication of the robots' relative strength.

Based on recent world computer-bridge championship results, odds-makers would rank Wbridge5 and Jack tied for the top spot.  Bridge Baron and Micro Bridge have been consistently good performers and have challenged Wbridge5 and Jack for the top spot.  Q-Plus has also been a steady performer, but has not recently advanced to the finals.  Relative newcomer Shark Bridge has shown great promise, winning the Individual, finishing second in the round-robin and losing narrowly to Wbridge5 in the semifinals.  Based on this performance, Shark Bridge is close behind the top two performers and advancing.  Second-year entry RoboBridge had some good results and may be more competitive next year.  There are other possible competitors for next year's championship.

Summary of the 2008 Championship... Seven robots competed in this year's Championship.  The format was a 32-board round robin with the four top finishers advancing to the KO stage.  The semifinals and finals were 64-board matches with a carryover from the round robin.  For complete results go to www.computerbridge.com 

Some statistical snapshots...

The average swing-per-board (spb):

For the round robin: the average spb for all matches = 4.75
For the round robin: the average spb for matches between top four finishers (Jack, Shark Bridge, Wbridge5 and Micro Bridge) = 4.3
For the semifinal matches (Wbridge5 vs. Shark Bridge, Jack vs. Micro Bridge): the average spb  = 4.08
For the final match (Wbridge5 vs. Jack): the average spb = 4.86
For the top-four from the round robin and the semifinal and final matches: the average spb = 4.32, *Norm=4.25

* For the 2007 Bermuda Bowl/Venice Cup/Senior Bowl/Transnational Teams, the combined spb = 4.25

Imp swings:

Final match, Wbridge5 vs. Jack:

Imps Occurrence Percent  Norm*
0-2	46	  71.9	 50.0
3-5	  6	    9.4	 16.0
6-8	  5	    7.8	 12.0
9-12	  6	    9.4	 16.0
13+	  1	    1.5	   6.0

Semifinal matches, Jack vs. Micro Bridge and Wbridge5 vs. Shark Bridge:

Imps Occurrence Percent Norm*
0-2	61	 47.7	 50.0
3-5	27	 21.1	 16.0
6-8	18	 14.1	 12.0
9-12	18	 14.1	 16.0
13+	  4 	   3.1	   6.0

Combined Semifinal and Final matches:

Imps Occurrence  Percent  Norm*
0-2	 107	  55.7	  50.0
3-5	  33	  17.2	  16.0
6-8	  23	  12.0	  12.0
9-12	  24	  12.5	  16.0
13+	   5	    2.6	    6.0
*norm based on 2007 Bermuda Bowl/Venice Cup/Senior Bowl/Transnational teams combined

Round Robin 4a (boards 1-16) chosen at random

Combined Wbridge5 vs. Shark Bridge and Jack vs. Bridge Baron:

Imps Occurrence  Percent  Norm*
0-2	15	  47.9	  50.0
3-5	  3	    9.4	  16.0
6-8	  5	  15.6	  12.0
9-12	  5	  15.6	  16.0
13+	  4	  12.5	    6.0
Average swing-per-board = 5.3

RoboBridge vs. Q-Plus Bridge:

Imps Occurrence  Percent  Norm*
0-2	 2	  12.5	  50.0
3-5	 5	  31.3	  16.0
6-8	 2	  12.5	  12.0
9-12	 3	  18.8	  16.0
13+	 4	  25.0	    6.0
Average swing-per-board = 7.6

While this is very limited data, the results point to lower Imp swings per board and less high swing occurrences when 'more evenly matched-higher ranked' robots are playing, with the results approaching those for high level human play.