[helene_t]

In the System Metrics thread, jdiana wrote:

jdiana, on 2025-February-28, 10:51, said:

...
As an example of something that I think would be really useful, Matthew Kidd has talked about adding a feature to BBO Helper as follows:

"A bigger idea is to provide a heads-up display of your opponent's style based on their playing history, classifying actions along both aggressiveness and wildness dimensions. So for takeout doubles, where a player averages on an (HCP + distribution points) histogram would measure their aggressive while their proclivity to make off-shape doubles would measure their wildness, e.g. a takeout double with 5-3 in the majors or a stiff in the clubs when 1♦ was opened, would drive up your wildness score, and a doubleton in an unbid major even more so. Similarly one can examine two-level overcalls, perhaps with a special statistic for how often they overcall on 5-3-3-2 shape—you see a lot of this even in open ACBL events; sometimes it can't be punished but sometimes it merely goes unpunished."

https://bridgewinner...2-2-se8mmlwks3/

Maybe it would be interesting to poll experts about what unanswered questions they think are most worth solving. It might also be useful to think about what BBO could do to help further this type of research. For example, if everyone was forced to click on a radio button to describe their basic system, would that make this kind of research easier? What else could BBO do to facilitate research?
...

I haven't personally mined the myhands data but I have done some mining of the vugraph archive. So far this has lead to an article in the IMP magazine about effectiveness of preempt styles, but there are lots of other things one can look into in the vugraph archive.

I know that other people are working on analysing real or simulated bridge data. But above all, as jdiana suggests, it would be good to hear from other BBO folks what kind of research questions they would find interesting.

[hrothgar]

helene_t, on 2025-March-02, 12:24, said:

In the System Metrics thread, jdiana wrote:


I haven't personally mined the myhands data but I have done some mining of the vugraph archive. So far this has lead to an article in the IMP magazine about effectiveness of preempt styles, but there are lots of other things one can look into in the vugraph archive.

I know that other people are working on analysing real or simulated bridge data. But above all, as jdiana suggests, it would be good to hear from other BBO folks what kind of research questions they would find interesting.

I'd like to see a test harness for evaluating the efficiency of various auction termination mechanisms

[helene_t]

Maybe my next project will be an AI tool that can predict Richard's posts

[jdiana]

This might be too simple of a question, but I would be interested to know whether, playing a typical 2/1 system, it really is better to open aggressively as opposed to having a sound opening hand. We could easily put some parameters on "aggressive" and "sound". The prevailing trend is clearly toward opening more aggressively but I'm not convinced that it's actually better, at least at my (average club player) level. (Both the level of the field and the card-playing skill of the opener's side probably are relevant variables.) When playing practice hands on Fun Bridge, it seems that opening light often leads to us going down in a game contract when we might have been allowed to play in a part score, or done better by defending. But that could be my selective memory.

[helene_t]

jdiana, on 2025-March-08, 08:58, said:

This might be too simple of a question, but I would be interested to know whether, playing a typical 2/1 system, it really is better to open aggressively as opposed to having a sound opening hand. We could easily put some parameters on "aggressive" and "sound". The prevailing trend is clearly toward opening more aggressively but I'm not convinced that it's actually better, at least at my (average club player) level. (Both the level of the field and the card-playing skill of the opener's side probably are relevant variables.) When playing practice hands on Fun Bridge, it seems that opening light often leads to us going down in a game contract when we might have been allowed to play in a part score, or done better by defending. But that could be my selective memory.

This is very similar to the topic of effectiveness of undisciplined preempts which I wrote an article about for the Dutch IMP magazine, so I have the code and the data for it already.

There 13536 deals in the Vugraph archive where at least one player opened 1H or 1S holding 8-10 HCPs, while at least one other player decided not do.
In that set, the light opening won on average 0.15 IMP, with a 95% confidence interval 0.07 to 0.24.

There are obviously a lot of limitations of this analysis. One thing that is very difficult to address is that even if the light openings work when they come up, they may backfire when you have a normal opening and partner doesn't dare to try for slam because you might have a light opening.

[jdiana]

helene_t, on 2025-March-08, 10:50, said:

This is very similar to the topic of effectiveness of undisciplined preempts which I wrote an article about for the Dutch IMP magazine, so I have the code and the data for it already.

There 13536 deals in the Vugraph archive where at least one player opened 1H or 1S holding 8-10 HCPs, while at least one other player decided not do.
In that set, the light opening won on average 0.15 IMP, with a 95% confidence interval 0.07 to 0.24.

There are obviously a lot of limitations of this analysis. One thing that is very difficult to address is that even if the light openings work when they come up, they may backfire when you have a normal opening and partner doesn't dare to try for slam because you might have a light opening.

Thanks! I wonder if form of scoring matters. I really only play matchpoint pairs.

[helene_t]

jdiana, on 2025-March-08, 11:03, said:

Thanks! I wonder if form of scoring matters. I really only play matchpoint pairs.

This is difficult to say from the VuGraph archive. Very few events are matchpoints and those that are often recorded habzardly with very little duplication.

[awm]

helene_t, on 2025-March-08, 10:50, said:

This is very similar to the topic of effectiveness of undisciplined preempts which I wrote an article about for the Dutch IMP magazine, so I have the code and the data for it already.

There 13536 deals in the Vugraph archive where at least one player opened 1H or 1S holding 8-10 HCPs, while at least one other player decided not do.
In that set, the light opening won on average 0.15 IMP, with a 95% confidence interval 0.07 to 0.24.

There are obviously a lot of limitations of this analysis. One thing that is very difficult to address is that even if the light openings work when they come up, they may backfire when you have a normal opening and partner doesn't dare to try for slam because you might have a light opening.

Can you separate 3rd/4th seat actions from 1st/2nd? It is much safer to open this way in 3rd/4th playing relatively standard methods.

I suppose it's also possible that some people are playing a system which better accommodates these openings (rather than, say, "standard 2/1" where you expect partner to game force with a misfitting 12-13). This would be hard to automatically figure out, but perhaps you could look for particular pairs who make these openings which much higher frequency and try to figure out what they play / if they have better results than people who do it only occasionally.

[awm]

I have a rather simple question about preempts. It used to be popular (in Europe) to play the combination of 2♦ multi (including major suit preempts with a six card suit) and 2M showing five in the major with 4+ in a minor. These days I see a lot more people shifting to preempts which simply show 5-6 cards in a major and unbalanced (whether 2♦ multi showing this, or 2M showing this). Here's what I'd like to know:

Suppose that partner opens 2M and this shows either six in the major or five in the major with 4+ in a side minor. How often does it make a difference to you which of the two hand types partner has (for example, you would pass a six-card preempt and run to a minor opposite the five-card preempt, or you would raise the six-card preempt and pass the five-card preempt, or you would bid 4M opposite the six-card preempt but 3NT opposite the five-card preempt)?

My estimate is that this actually isn't very often (which seems to be the assumption of other people playing this style also) but it would be good to get some actual data.