[jcwla]

This suggestion applies to the games that include both all-human and human-robot partnerships, including a large number of virtual club games in which humans may choose (often for some additional fee) to partner with a robot.
It is based on the premise that these human players are entitled to a robot partner who, whatever its own limitations may be, at least understands the meanings of the opponents' bids to the same extent as do their human counterparts.

The problem it addresses comes on hands such as one this morning in which the robot's human LHO opens 2NT, explained in the chat box as "6-11 minors."
The robot does not understand this. It believes that its human LHO has shown a strong 2NT opening.
The robot's human partner has a creditable opening hand and overcalls 3H.
The robot's human RHO jumps to 5D.
Robot holds QJ73 / AQT84 / T / J75 but simply passes 5D, believing that LHO is sitting behind it with a moose.

It is entirely possible that the robot may have taken the same or even some worse action had it understood the meaning of 2NT, and in that case it could be chalked up to "you take the good with the bad when you play with the robots."

But I suggest that the software be modified (at least as to alerts to robots) such that the robot is alerted to the HCP range and the range of number of cards in each suit promised by the alertable call.

[smerriman]

The robots will assume all other players at the table are playing their system, and are incapable of understanding anything else.

(The 'alerts' themselves aren't even used by the robots for their own bids; they're more for humans and don't always reflect what the robot actually has).

[pilowsky]

The alerts made by robots are more for amusement value.


On the other hand, like the alerts made by humans IRL they have some relationship with reality but with a cloud of uncertainty that - like the universe - gets larger as the bidding continues.
Players making bids can be likened to players keeping time with each other and playing the same notes in an orchestra but without a conductor and with no music to follow.


When done well it sounds right; done badly it sounds like the Portsmouth Sinfonia - a cloud of sound surrounding a tune.

[Huibertus]

smerriman, on 2021-November-04, 17:02, said:

and are incapable of understanding anything else

Well, they are uncapable of understanding their own system too. It's a programmer making them aware, and that can be done for their own system AND for any other system.

THe alerting could be facilitated by a drop down box that list all options that were programmed.

[smerriman]

Huibertus, on 2021-November-06, 13:13, said:

Well, they are uncapable of understanding their own system too. It's a programmer making them aware, and that can be done for their own system AND for any other system.

I'll rephrase my response then to say that the programmers don't appear capable of doing it either Tiny bug fixes for the existing robot haven't been made for years. My point was that something like this isn't possible for the current robot due to the way it works, so one that understands human alerts would have to be an entirely new robot altogether - slightly harder than the already-too-hard tiny bug fixes.