Baffled. No sign of the Declare-only Group challenge. Did it expire before I played? Anyway, I just reissued it.
I win the regular group challenge with +15.25, a margin of 10. I lose to procrastinator by 19, beat NN by 48. All reissued.
Amazing. 8 board match against AA. All 8 boards tied. Exactly the same contract, exactly the same result. I didn’t drill down to see if the play was exactly the same, trick by trick. Reissued.
Interesting! Or not at all interesting, depending on how you look at it. One thing that makes that more likely is 1 was Passed Out, of the other 7, 6/7 we were on defense.
Looks like you didn’t finish the Declare-only. md won with a score of 25.67! 18 clear of 2nd place.
I would expect more card play variation on defense than when declaring. Overall more contract variation if we are declaring. Here we were in the auction (at least at may table) on 4 of the 6 where we ended up defending, and competitive auctions would often lead to contract variation;
Makes sense. My side was also competitive in 4/6 auctions.
New topic: I am a AI sympathizer. I tend to respect it a lot. This may be driven a bit by my chess knowledge, and how computers are clearly better than humans here, now. The ability to remember and calculate is so far superior to ours.
As has been said in another thread (on the AO I believe), the bots will never miss a count or a pip, which is huge. Compared to someone like me, who is good at counting and remembering for a human, they are so far superior. It makes their defense, their offense, their squeezes so surgical, I often feel them playing hands perfectly and as a result beating me by a trick. All that said, sometimes they make some absolutely baffling plays that frustrate me so much. Maybe I just notice it more, but it seems to come from my partner when we are in defense, especially against no-trump contracts.
For example, I lead a heart, that should clearly be my 4th highest from strongest suit. And they don’t lead one back despite multiple opportunities. And sometimes they do this and we set them because they have another angle - great. But other times we’ll go down like 12-1 when we could have set them if they just led back my opening suit once…
Yes, that is frustrating, and certainly happens. One reason is that the bots themselves often do not lead from their longest suit against NT and therefore (I suspect) do not think that we have usually led our longest suit.
AA wins the regular group challenge with the very low winning score of +3.25. procrastinator wins the declare only with +12.50. Neither was a margin of 10 (over second place), but procrastinator’s margin over 2nd place was larger than AA’s margin over fifth place. Quite surprising since I have hypothesized that declare-only contests should be more uniform than complete contests. That may well still be true on average, but not necessarily for all 16-board sets.
Both reissued.
[quote=“AbstractActuary, post:130, topic:417”]
For example, I lead a heart, that should clearly be my 4th highest from strongest suit. And they don’t lead one back despite multiple opportunities. And sometimes they do this and we set them because they have another angle - great. But other times we’ll go down like 12-1 when we could have set them if they just led back my opening suit once…
[/quote]
Yes, that is frustrating, and certainly happens. One reason is that the bots themselves often do not lead from their longest suit against NT and therefore (I suspect) do not think that we have usually led our longest suit.
The hand I actually had in mind I lead a Qh, where I had QJT9xx or something. And he won the second trick and didn’t lead it back.
Anyway, I need to probe this more. I suspect you are right, that bot is not considering my lead to be 4th from my longest and strongest suit. That’s the only explanation. Which begs the question. Is there a leak in the bot programming? Is there a leak in the default lead of 4th from longest and strongest? Why is there a disconnect here? Clearly in the most direct sense, it means I should play with the bots not assuming that they are leading from the suit they want returned, and not expecting them to return my suit. But (1) then what should I expect? Or how can I get them to lead a certain suit? How can I trick their algorithm? And (2) does this lead to the conclusion that I (/others) should change the way they play against NT contracts? Or that the bot should be reprogrammed?
From the system notes https://www.bridgebase.com/doc/gib_system_notes.php
“GIB usually leads passively against NT (read the book Winning Notrump Leads to understand why). Don’t assume it’s leading its longest suit. When you lead, it doesn’t assume you’re leading your best suit, which is why it doesn’t always return the suit like a human would.”
I don’t think much you can do to influence their defense. Note that attacking long suits vs NT is a better imp strategy and total point strategy than a matchpoint strategy (our challenges are always imps, which I prefer to matchpoints; total point strategy usually leads to the same decision as imps).
I haven’t read “Winning Notrump Leads”, or the related book “Winning Suit Contract Leads”. One of my regular partner’s did and thought it made sense. However I have read some reviews of its methodology, which is highly suspect. Both books look at a large number of randomly generated deals, consistent with the bidding, and then look at the “results” of various leads. That “result” is what happens, double dummy, after that lead. The biggest criticism of the mythology as that it concludes that leading an unsupported ace (in an unbid suit) against a suit contract is usually a good idea. Assuming that after seeing the dummy, you will always find the best continuation, does increase the value of holding the lead. In the real world, with the dual whammies of the chance that leading the ace is fatal to the defense, and the chance that you will not find the best continuation at trick 2 even if there is one, mean that most players are reluctant to lead unsupported aces.
Win by 15 vs NN, by 5 vs procrastinator. Both reissued
Our heads up challenge expired before you played it. I reissued another one.
I’m going to put this here.
I have a question about a bidding line and maybe more generally about this “game suit try” idea.
Bidding went
1H P 2H P
3D P 4H
3D description is “5+H, 3+D, Q+ in D, 16-17 points, forcing”.
I think I’ve also seen something similar described as “game suit try”. And if not, maybe someone can explain the nuance there. My main question is “why?”. Why is this better than a 3H bid, which presumably accomplishes the same thing of telling your partner you have enough points for game if they are at the high end of their range (“invitational”). Why does giving that little bit of barely valuable information in diamonds help your partner make this decision? I mean, I can see some small benefit to partner knowing this information, but from the will-be-closed hand, this seems like potentially valuable information to the defense. So why is 3D better than 3H here?
With the bots, perhaps it doesn’t matter. With a good partner, it is a significant help in hand evaluation. It says that honors in diamonds (especially relevant for the K and Q) would be very valuable, presumably in combination with honors you have in diamonds. Similarly Q and J in clubs or spades are quite likely worthless. So a hand with the spade or club A and the QJ of diamonds would be an absolutely definite accept (with a good human partner, even with only 7 HCP), but heart K, plus QJ of both clubs and spades and xxx of diamonds is a decline despite 9 HCP.
oops
Interesting. That is helpful. Thank you.
procrastinator first in both group challenges. In the regular, with +11.25, not a margin of 10. In the declare-only with +9.75, not a margin of even .25 (A tie for first, with NumbersNerd)
Both reissued.
ETA: + 2 vs AA, 5-3 with no swing > 2. Reissued
ETA: +5 vs procrastinator; -6 vs NN. Both reissued
Not only no swing >2. No difference of making a contract or not, or us declaring vs opponent. All a matter of overtricks one way or the other.