GreenMan

I think you're doing Bayes wrong. The statement "Black people are more likely to be criminals" is qualitatively different from "Criminals are more likely to be black people." The first, for reasons such as those you state, is true, but your horse-zebra argument is based on the second, which probably is not.

This may not address your point directly (I'm short of time so cutting some cognitive corners), but tons of studies have found that white and black people get treated differently in situations where people's judgment is involved -- getting car loans, house hunting with a real estate agent, etc. -- even if their clothing and supporting information such as finances are identical. Similar effects have been found in news media. So it appears you can control anything you like and a racial bias persists.

Undo requests don't bug me that much, but I mostly play with people I know and we don't mind reversing an occasional misclick. If I play with someone who abuses the privilege, or who denies them indiscriminately, then I just make a note not to play with them again. It's not that big a deal.

Most people play that sequence as 5-5. If it could be 5-5 or 4-5, opener with e.g. 3244 has a tough guess. Yeah there are plenty of other solutions if you like devoting larger chunks of memory to the problem, but most people considering Smolen aren't in that category.

Don't change the subject. If police get a report that some unknown person committed a crime they should stop and question everyone, white or black, but they don't. And if police or the TSA are just looking around for "suspicious behavior" they should look at everyone the same regardless of their color. But they don't. And when people object to that sort of profiling then they say that they're kept from doing their jobs by politically-correct people forcing them to include upstanding white citizens in the suspect pool.

It's not just that people who don't look white and middle class get targeted, it's that people who DO look white and middle class get passes. You may have heard of this experiment, where two bicycle thieves, one white and one black, tried to cut through a bike lock in a public park in broad daylight. The white guy was left alone, and one person even wished him luck. The black guy was hounded from the start, and almost everyone who saw him called the cops. That's the insidious kind of racism: Things that bring the heat onto a dark person get a pass if white people do them. If your enforcement strategy varies by race or ethnicity -- ever heard of "driving while black"? -- then yes, you're racially profiling.

Dark skin != suspicious behavior. :angry:

We showed up! That's the best you can say about my performance. :unsure: kristen33 and jillybean were their usual steady selves. My young partner did well. I was hoping he'd declare every contract we played, but it was not to be. (Check out board 8. I wouldn't have made that contract in a million years.) Throughout the match I noticed one important factor: The more active pair came out ahead on virtually every deal. We lost a lot of IMPs when jec/BridgeGoth wouldn't let us play in the contract that was reached in the other room. They just wear you down that way. We only had one double-digit swing, but all those 5s and 6s added up fast. (And every time I took a conservative view in the auction, we lost more IMPs.) Thanks again. We enjoyed it a lot. Onward! :)

Thanks, hrothgar. :) So if I'm understanding correctly, there may or may not be an equilibrium where a game-theoretical optimal strategy exists, and it may be possible in principle to figure out whether this strategy exists and, if so, what it is. This strategy could then be programmed into GIB. It would be a lot of work and no one is bothering with it right now. A human using it would probably want to be able to adjust it according to who the opponents are (and partner, if he or she is also an unknown quantity), and that gets even messier. So it's kind of an interesting question that is unlikely to be answered anytime soon. Bridge imitates life. :)

I think it's more accurately spelled "podnah". :) It's most often simply abbereviated "p" as in "wdp", but "p" by itself can be ambiguous, so when the mood strikes we add the other hard consonant in "pard" to clarify.

The OP asked for a game-theoretic solution that did not require knowledge of the opps' style. I suspect the reference to GT was out of place here. By my limited understanding of GT, any solution would depend on knowing the opps' strategy, e.g. what % of the time they bid strategically and falsecard. (e.g, restricted-choice calculations tend to assume random selection from equals, IIRC.) It also must account for the likelihood of opponents' errors -- the Principle of Restricted Talent, as Chthonic called it. Since these calculations change for each set of opponents, I suspect any such strategy would be effectively impossible to create. Anyone with a better understanding of game theory is invited to correct my mistakes. :) I find it a fascinating field that I haven't had much opportunity to explore.

The "problem" I was referring to is that it's a convention you have to devote memory space to that apparently makes no difference in your results. There's some theoretical advantage to it, but in practice I've never seen nor heard of a hand where Smolen had any effect on the outcome.

It could be made to apply only after trick 1, I suppose. That doesn't matter much to me, but I wouldn't mind an "auto-play to last trick" option.

Tell him Reverse Smolen isn't alertable.

:)

My way of thinking tends toward preferring advancer to bounce the pre-empt as high as possible, and this is facilitated by overcaller showing at least one actual suit. I can see the argument the other way. Thanks for the discussion. :)

:) I heard of someone actually asking how the auction went in a passout, and getting told, "Pass ... pass ... pass ... pass."

"You were in 1♣? How did the wait don't answer that." --actually said to me after score comparison

Are you sure the tournament wasn't identified as a Spanish-language event?

My wife signed me up for a bridge club. I jump off next Tuesday. --Rodney Dangerfield

I just now had a similar experience: showed Rank=1 for the last few boards, but I finished 3rd OA, 1st in B, 1st in C. The overall winner was 20 IMPs ahead of me, so I know it wasn't a last-board charge. It appears to me that that rank is where you stand in "your" stratum, not overall unless you're in stratum A. This tournament had 21 tables, and I was -2 IMPs after the first two boards but ranked 4th; my overall ranking would have been 12th or so.

I think Rodwell calls this the Sandpaper Coup.

It's been years since I played this system, but I don't recall any particular difficulties of that nature. Maybe that's just nostalgia. :)

Interesting ... when I played a strong-club system our explicit standard for 1♣ was 5 points better than a minimum opener with the same shape. So e.g. Kx AKxxx KQxxx x would qualify, but Kx AKxxx KQxx xx would not. I don't remember landing in an excessive number of hopeless 24-point games.

The ACBL's does: "vs Opening Preempts Double Is Takeout thru ___" But that's not near the space where negative doubles are indicated. They are, as you point out, different things.