Winners of the Fall NABC Robot Individual Practice Tournament #2
Last weekend, November 16-17, ACBL BBOers were invited to play in the second NAB...
This month, we set out to evaluate the capabilities of the various bridge robots used or developed on BBO. To do this we expanded our usual robot simulations, where BBO's GIB robots compete against real players in daylong games, to include two highly competitive events: December's Stars and Platinum Robot Individual, and one of BBO's Zenith Daylongs.
In the first simulation, five different robots replayed the December Stars & Platinum Robot Individual to assess their performance against human players. The participating robots were:
The results were intriguing, demonstrating that the robots are formidable players, even against strong human competition. Click here for the full leaderboard.
| Rank | Score (%) | Robot Type |
| 60/618 | 56.30 | GIB Advanced |
| 93/618 | 55.23 | Argine |
| 123/618 | 54.29 | GIB Basic |
| 232/618 | 51.42 | Argine Basic |
| 486/618 | 39.47 | ACBL Ben |
Join the BBO Forums discussion about this experiment here.
In the subsequent Zenith Daylong Reward simulation, we introduced "Thinking Ben," another robot variation "trained" by Lorand Dali, to the robot lineup.
For the full leaderboard click here. Here's a summary:
| Rank | Score (%) | Robot Type |
| 522/1,672 | 53.91 | GIB Advanced |
| 646/1,672 | 52.58 | GIB Basic |
| 795/1,672 | 50.96 | Argine |
| 861/1,672 | 50.32 | Thinking Ben (ACBL) |
| 1,016/1,672 | 48.38 | Argine basic |
| 1,232/1,672 | 45.64 | Fast Ben |
To join the ongoing discussion and learn about the evolution of this experiment, click here.
Stay tuned for more information on robot development, Fast Ben, Thinking Ben, Lorand Dali and other exciting experiments.
In the world of bridge, if a partner is hiding information, that's not good. if the enemy is hiding information, TD can punish him. BBO, would you be so kind as to write down the name of the robot that is on the table. Also would you kind to give us a complete system and style gibs play by robots name. Thanks.
Who bots were playing against when replaying these two tourneys? Against other bots? If yes, the whole experiment shows very little