Anti-robot systems
With EBU events being played online at present, many of them allow robots to participate. One notable property of the robots is that they don't pay attention to system summaries, or to explanations of bids, which means that it is possible to get a noticeable advantage over the robots by playing a system that they don't expect.
There are quite a few things that can be done at level 4 that would be absurd against humans, but would give a notable advantage when playing against robots. For example, swapping the meanings of 1!c and 1!d in an otherwise natural system is legal at level 4, and in robot tournaments, players have observed that opening their shorter minor often causes the robots to go haywire during the play (because they assume the actual layout can't possibly exist, and thus play for implausible possibilities).
As such, if an event is likely to be particularly robot-heavy, it may be that players decide to try out systems specifically designed to confuse the robots. I think it might be helpful to have guidance on whether or not this sort of thing is legal in EBU tournaments; in one sense, it's playing within the stated rules for legal systems, but in another sense, it's giving misinformation because the robots can't be successfully informed about such a system. (As a side note, I note that some bridge organisations ban agreements whose only purpose is to "destroy the opponents' methods", which appears to be mildly relevant here; however, the EBU appears not to have such a rule.)
Comments
I think we'll keep an eye on this and consider acting if it actually seems to be becoming a thing.
In a Precision system, opening 1D on a singleton is not uncommon.
They would also be surprised by a weak NT...?
My experience of playing a weak NT in Funbridge Robot games is that I seem to gain and lose in about equal measure from the robots expecting me to have more than I do.
I'm not sure the same principle applies in Funbridge games since you have to give it a system card and it normally responds to that as a human player would (subject to the normal weirdness you can sometimes get with any bridge-playing robots!). I can see how BBO robots might make weird plays since that same information is not available to them though admittedly they can also make weird plays some of the time, and the programming does account for "psychic" bids at least to some extent (i.e. if it expects you to have 4+ clubs and then show out on the 2nd round, it realises you've psyched (or in this case are playing a different system) and reacts accordingly, and equally will realise if a bidding sequence is inconsistent with their hand).
I suspect Gordon's comment sums it up. The trial of robot and non-robot games seems to have actually led to insufficient turnout in some of the robot games so there appears to be less demand for playing with a robot than there was, and in any case robots could easily be forbidden for higher level events.
It responds in the knowledge of the system you are playing but it defends on the assumption that you are playing the same system that it is.
Law 16A2 says ... Players may take account of the traits of their opponents.
So I don't see the problem (provided they explain what the bids mean).
I'm not playing an intentionally anti-robot system, but an example of this sort of thing came up recently, so I thought I'd post it here to get a better idea on what sorts of potentially problematic things can happen in practice. This happened in an EBU event where level 4 agreements are permitted.
{b... 1nx-2d x-3c-- x---}
Alerts and announcements:
1NT: 15-17
X of 1NT: artificial, showing 10+ HCP, 4+ spades, and a second 4+-card suit (not specified)
2D: a two-way bid; either natural with 3+ spades, or else "a hand that hates both clubs and spades" (typically both red suits)
X of 2D: takeout of diamonds
X of 3C: penalty (an ill-advised bid that was probably a mistake, but that was made at the table)
Unfortunately, for the opponents, North was a robot. Not being able to see any of the alerts or announcements, it concluded (based on the contents of dummy) that West must have a singleton King of Diamonds (the only way for West to be strong enough for the "penalty" double of a strong notrump was to hold the King and indeed all the other high honours, and for East to have enough diamonds to pull it was for the King to be singleton). So the robot declarer played the Ace of Diamonds at its first opportunity, losing a trick when the King turned out to be onside rather than singleton offside (dummy included !dAQ), and giving East/West a top rather than a bottom when 3!c went down exactly 1 as a consequence.
This particular artificial double is something I play in my partnership at matchpoints because it's very good at improving results by typically 10 or 20 points, and comes up very frequently (much more frequently than a penalty double does); it wasn't intended as a specific anti-robot measure, and I use it (and first used it) in robots-disallowed tournaments in addition to when playing against robots. It seems likely, however, that it will get a lot of random tops against robots that don't know what's going on. (One consequence I've noticed is that it may be distorting the NGS; the robots have a pretty good NGS score, and so something that confuses them tends to be a good way to shoot up the NGS rankings. This may help to explain why my NGS tends to go up in robots-allowed tournaments, and down in robots-disallowed tournaments.)