[pescetom]

This board had me frustrated about the robot's combination of 5cM 1NT and 4cM Stayman, but also about play of diamonds at trick 4?

[smerriman]

Bidding completely normal.

East's lead of the T is either precisely T3 doubleton, or any other combination containing the 9. Everything else being equal, there are considerably more of the latter than the former, so it sees it as completely clearcut at MPs.

I guess the human logic is that West would have ducked with other combinations, but GIB doesn't consider what ifs like that.

Of course, humans would have led a low club, making this moot at other tables, but GIB also assumes other tables are GIB.

[johnu]

What's the format of the tournament? One human N/S and one human E/W? Are there robots sitting North at the other tables, in which case the bidding should be replicated across those tables.

Why is the tournament set up to let robots declare the hand when there is a human available to play that hand. IMO, the purpose of a tournament is to test the bidding and play of the humans, not the robots whose play should be mostly identical.

[pescetom]

johnu, on 2024-August-30, 18:07, said:

What's the format of the tournament? One human N/S and one human E/W? Are there robots sitting North at the other tables, in which case the bidding should be replicated across those tables.

Why is the tournament set up to let robots declare the hand when there is a human available to play that hand. IMO, the purpose of a tournament is to test the bidding and play of the humans, not the robots whose play should be mostly identical.

The format of the tournament is MP pairs, with human players free to partner another human of their choice or a robot.

The tournament is set up to allow the declarer to play, whether human or robot. Which follows the Laws of bridge and is considered quite normal by the players (no robot has complained either).

[pescetom]

smerriman, on 2024-August-30, 16:06, said:

East's lead of the T is either precisely T3 doubleton, or any other combination containing the 9. Everything else being equal, there are considerably more of the latter than the former, so it sees it as completely clearcut at MPs.

I guess the human logic is that West would have ducked with other combinations, but GIB doesn't consider what ifs like that.

Not just the fact that West did not duck, but surely everything else is NOT equal? If the lead was from T3 then playing the 7 will be a MP disaster, as indeed it was. You have to consider the gravity of consequences, not just the frequency.

[jillybean]

pescetom, on 2024-August-31, 06:26, said:

The format of the tournament is MP pairs, with human players free to partner another human of their choice or a robot.

The tournament is set up to allow the declarer to play, whether human or robot. Which follows the Laws of bridge and is considered quite normal by the players (no robot has complained either).

Which law covers robots?

[pescetom]

jillybean, on 2024-August-31, 09:28, said:

Which law covers robots?

No law currently. What the law does cover is which player plays the contract, namely the player who for that side first bid the denomination of the final contract (see Definitions and 41a). No obvious reason why that should change once the laws cover robots specifically, either.

[smerriman]

pescetom, on 2024-August-31, 06:43, said:

Not just the fact that West did not duck, but surely everything else is NOT equal? If the lead was from T3 then playing the 7 will be a MP disaster, as indeed it was. You have to consider the gravity of consequences, not just the frequency.

Not following - if the lead was from T3 you score 0%, but if it was T9x3, you score 100% don't you? *If* the latter were a slight favourite, which is our conditional, isn't that a better MP play than a guaranteed 50%?

I guess it depends on your position in the tournament to date.

[pescetom]

smerriman, on 2024-August-31, 13:24, said:

Not following - if the lead was from T3 you score 0%, but if it was T9x3, you score 100% don't you? *If* the latter were a slight favourite, which is our conditional, isn't that a better MP play than a guaranteed 50%?

I guess it depends on your position in the tournament to date.

Partly the latter (if you are hoping to win, and I was close), but also if it was T9x2 I do not expect anywhere near 100%... this is not the US and half the field will be in 2♠, even if not there is leeway for level of declarer play in the remaining tricks and robot is not the best (although above average) in this field.

[smerriman]

Ah yep - so that was my original point of GIB assuming GIB at other tables. The fact it doesn't know when its "ahead" makes it a poor MP player, though I wouldn't have the foggiest idea how to change that.

[johnu]

pescetom, on 2024-August-31, 06:26, said:

The tournament is set up to allow the declarer to play, whether human or robot. Which follows the Laws of bridge and is considered quite normal by the players (no robot has complained either).

IIRC, when BBO first started robot tournaments, the robots always declared the hands. Later, human declare tournaments were offered as an option, which became a lot more popular, even though the robots probably played better than a lot of the human players. That's probably because humans have selective memories and quickly forget when the robot declarer played the hands without incident, but the hands where the robot completely misplayed in a manner that no human would really left a mark.

As for the Laws of Bridge, they don't contemplate having robots being part of a table. I stand by my belief that tournaments are designed to test who the best human players are, not whether a robot is a better player than a human.

[pescetom]

And I stand by my belief that pairs tournaments are designed to identify the best pair, not the best player or the best human player.
As does the tournament organizer.

[pilowsky]

 jillybean, on 2024-August-31, 09:28, said:

Which law covers robots?

Apparently there are 3.

[johnu]

pescetom, on 2024-September-01, 06:28, said:

And I stand by my belief that pairs tournaments are designed to identify the best pair, not the best player or the best human player.
As does the tournament organizer.

Robots aren't awarded "masterpoints", or whatever prizes are awarded.

[jillybean]

pilowsky, on 2024-September-02, 01:30, said:

Apparently there are 3.

?

[johnu]

jillybean, on 2024-September-03, 18:22, said:

?

Asimov's "Three Laws of Robotics"