[pilowsky]

There's an old Tom Lehrer song that starts out "Don't solicit for your sister that's not nice...". Which kind of reminds me not to double at IMP's but this hand was really really tempting.
I open a (for me very conservative) 1NT and the robot has the temerity to double. Well, what's a man to do: redouble. Of course, I mess it up and they get 1000 for -14.9. Except, I should make +2.
Here it is:
♠2 led

---

[Cyberyeti]

Come on pilowsky, playing a spade at trick 3 is deliberately kamikaze, it is absolutely no win and must be trolling. Just bang down ace and another diamond at that point and take your 7 tricks.

What were you thinking, the distribution is pretty much an open book.

[pilowsky]

I know right - it's pathetic - my only excuse is that I'd just woken up and I was in a bit of a hurry -- I'm sorry. I won't do it again.

[Vampyr]

I don’t know who told you not to double at IMPs. What you should avoid is low-level doubles intended to protect your partscore.

[pilowsky]

Vampyr, on 2020-September-15, 20:41, said:

I don't know who told you not to double at IMPs. What you should avoid is low-level doubles intended to protect your partscore.

It was Ron Karr from Palo Alto.
To be fair, I've never met Ron but I have met Warren and Pauline from the example in his pamphlet. Both charming people.
Of course, I also sometimes double at IMP's - in this example I even redoubled - I was going to give the robots a spanking. Typically, I copped a beating instead. I'll get my own back in a minute.

[smerriman]

pilowsky, on 2020-September-15, 20:57, said:

It was Ron Karr from Palo Alto.

He says nothing of the sort in the PDF; his advice aligns with Vampyr's above.

Of course, it's 100% irrelevant to this hand, since the double was artificial and not for penalty, and of course the passes were due to a bug. Specifically that GIB doesn't know what to do when you make insane bids as a human.

[pilowsky]

smerriman, on 2020-September-15, 21:53, said:

He says nothing of the sort in the PDF; his advice aligns with Vampyr's above.

Of course, it's 100% irrelevant to this hand, since the double was artificial and not for penalty, and of course the passes were due to a bug. Specifically that GIB doesn't know what to do when you make insane bids as a human.

Well, to be fair, he says something of the sort.
Here's what he says about IMP's:
Don't double unless you are very certain, particularly if doubling a partscore into game
Compared with his advice for Matchpoints where he says:
Double aggressively, particularly if they are vulnerable.

It sure sounds like different advice to me Stephen. sort of .

[Stephen Tu]

Of course strategy is different between doubling at IMPs vs doubling at matchpoints. No one disputes that.

But not doubling partscores into game without some margin of error (expect two trick set), is also a different strategy from "never double, ever".

[pilowsky]

Stephen Tu, on 2020-September-15, 23:17, said:

Of course strategy is different between doubling at IMPs vs doubling at matchpoints. No one disputes that.

But not doubling partscores into game without some margin of error (expect two trick set), is also a different strategy from "never double, ever".

Please tell me that you didn't really think that I was suggesting that people should NEVER double at IMPS EVER...

Everyone knows¹ (yes even me) that - but not the North robot sadly that there is a time and a place for doubling at IMPs - this board was not it. Thanks, Mr North robot. I manage to escape with a nice positive score overall - others were not so lucky.

This literary device is called "exaggeration for comic effect' it is quite widely used, along with irony, by non-software engineers, non-IT professionals and other people that don't play Bridge for a living in the English-speaking world outside America.
I know, it's a tough one to get your head around. A bit like a squeeze is for me. I'll keep trying though I'm sure we'll get there before I die of literary exhaustion.
Oy gevalt.

¹The bridge in the video is close to the Hunters Hill Bridge Club.

[msjennifer]

Sir,
Never double a contract unless you think they are stealing your hand.I had presented a hand in this column where an opponent doubled my 4♥ contract holding K1045 and expecting fully 2 trump tricks He did not get even a single one. and one overtrick was made by me.
One must not double a contract if one can not successfully double any other contract ,if they bid as an alternate one.Some days back an opponent holding J1098 of trumps doubled a contract of 7♠ and when opponents ran to 7NT could not defeat it.THANKS.

[DavidKok]

The semantic discussion on the title seems less than helpful. Personally at IMP scoring I only double partscores "if 2 down is in sight", which is conveniently open to interpretation based on how you view your own skill at defending. This is slightly more aggressive than Ron's advice, but then again most opponents are slightly more aggressive than they were in 2012 so it probably balances out.

I understand that East is a bot and the double was conventional, but if I was sitting West this board would be a beautiful end to the partnership. There is no hand not suitable for action on the first round that then wants to take out 1NT Vul against NV on the second round at IMPS, and even at matchpoints it is nothing short of a miracle. Double dummy EW can make 3♠ and South can make 3NT (and 3♦), but more often poor West gets to swallow -200 or -500 on an auction like this.

Playing real bridge the redouble is also a poor decision. South has described their hand, it is up to North to do something now. Perhaps this is different when playing against bots, but personally I wouldn't want to learn a method of play that only works on programmed tables.

[pilowsky]

DavidKok, on 2020-September-16, 03:05, said:

The semantic discussion on the title seems less than helpful. Personally at IMP scoring I only double partscores "if 2 down is in sight", which is conveniently open to interpretation based on how you view your own skill at defending. This is slightly more aggressive than Ron's advice, but then again most opponents are slightly more aggressive than they were in 2012 so it probably balances out.

I understand that East is a bot and the double was conventional, but if I was sitting West this board would be a beautiful end to the partnership. There is no hand not suitable for action on the first round that then wants to take out 1NT Vul against NV on the second round at IMPS, and even at matchpoints it is nothing short of a miracle. Double dummy EW can make 3♠ and South can make 3NT (and 3♦), but more often poor West gets to swallow -200 or -500 on an auction like this.

Playing real bridge the redouble is also a poor decision. South has described their hand, it is up to North to do something now. Perhaps this is different when playing against bots, but personally I wouldn't want to learn a method of play that only works on programmed tables.

I could not agree more about the semantic stuff, but I find that playing with Bots is giving me a much deeper understanding of how Bridge works mechanistically.
In Chapter 1 of his book on Precision in 1972, Reese discusses the difference between the two types of bidding systems.

The first, he calls approach-forcing - into this group come SAYC, Acol and 2/1. These systems are preemptive by nature, they announce to the table their holding with a thump. Nobody is in much doubt about what is in the opener or overcallers hand at the first bid. The problem is that most of the opening bids in these systems have an extremely wide range. Partner doesn't know where they stand.
Precision sought to remedy this by closing the gap on just about every bid except 1♣.

What I have learned playing with robots is that they really seem to like (or at least it seems that way to me) to know what HCP range I have on my opening bid.

If only my estimates of shape and my card play were up to snuff things would work out much better.

I must say, given that one is allowed to open '1♣ alerted as >16 HCP nothing to do with ♣' it really does seem weird to me that the some RA's get tetchy about opening 1NT without a specific handshape. I'm aware that this one has been argued about endlessly here and elsewhere, but there seems to be a clear inconsistency. To say nothing of the infamous forcing pass.

[DavidKok]

pilowsky, on 2020-September-16, 03:37, said:

I could not agree more about the semantic stuff, but I find that playing with Bots is giving me a much deeper understanding of how Bridge works mechanistically.
In Chapter 1 of his book on Precision in 1972, Reese discusses the difference between the two types of bidding systems.

The first, he calls approach-forcing - into this group come SAYC, Acol and 2/1. These systems are preemptive by nature, they announce to the table their holding with a thump. Nobody is in much doubt about what is in the opener or overcallers hand at the first bid. The problem is that most of the opening bids in these systems have an extremely wide range. Partner doesn't know where they stand.
Precision sought to remedy this by closing the gap on just about every bid except 1♣.

I have never played any strong club systems, although as far as I know they are/can be superior to natural systems. The core principle - assign most bidding space to the hand types that come up most often (around 10-15 points) and have some playing strength (a 5-card suit), and reserve the lowest possible bid for strong hands that need lots of room to explore what is going on, is extremely sound.

pilowsky, on 2020-September-16, 03:37, said:

What I have learned playing with robots is that they really seem to like (or at least it seems that way to me) to know what HCP range I have on my opening bid.

My partner taught (is teaching?) me a lot about the "shape first" principle, which states that shape of the hand is often more important for making decisions than strength (both high card and playing strength). Quite plausibly the bots like knowing your HCP range because that is easy to program, not because it is good bridge.

[pilowsky]

I hear that term "Good Bridge" a lot. To me only the score is important. What is good Bridge to you?

[Stephen Tu]

DavidKok, on 2020-September-16, 03:05, said:

There is no hand not suitable for action on the first round that then wants to take out 1NT Vul against NV on the second round at IMPS, and even at matchpoints it is nothing short of a miracle. Double dummy EW can make 3♠ and South can make 3NT (and 3♦), but more often poor West gets to swallow -200 or -500 on an auction like this.

I strongly disagree with this assertion. Maybe it holds if you had some weak opening available showing both majors, or super-light 1M opening bid system, but otherwise there are a ton of both major hands I'd want to come back in with as East, at IMPS unfavorable or really any conditions. Stronger than actual East hand, but not a *lot* stronger.

• West is marked with values, as North did not try for game
• With 5-5 in the majors, odds of finding 8+ fit are very high
• N/S are not going to be super eager to double your major partial into game at IMPs
• Shape is good for offense, not as great for defense as NS may be able to run a bunch of minor tricks.
• Even if you can beat NT in theory, partner will often lead wrong suit, declaring is often more accurate.
• At IMPS being -200 when the opps were going to be +120 otherwise is only -2 imps, not the disaster it is at MP

I think if you ran some sims of East's weak major 5-5s, < opening 1 bid, you'd find that it is much less of a miracle than you assert for bidding to work well, and in fact percentage action. Quite a few double partial swings available, and won't be punished very often.

[Cyberyeti]

Stephen Tu, on 2020-September-16, 05:54, said:

I strongly disagree with this assertion. Maybe it holds if you had some weak opening available showing both majors, or super-light 1M opening bid system, but otherwise there are a ton of both major hands I'd want to come back in with as East, at IMPS unfavorable or really any conditions. Stronger than actual East hand, but not a *lot* stronger.

• West is marked with values, as North did not try for game
• With 5-5 in the majors, odds of finding 8+ fit are very high
• N/S are not going to be super eager to double your major partial into game at IMPs
• Shape is good for offense, not as great for defense as NS may be able to run a bunch of minor tricks.
• Even if you can beat NT in theory, partner will often lead wrong suit, declaring is often more accurate.
• At IMPS being -200 when the opps were going to be +120 otherwise is only -2 imps, not the disaster it is at MP

I think if you ran some sims of East's weak major 5-5s, < opening 1 bid, you'd find that it is much less of a miracle than you assert for bidding to work well, and in fact percentage action. Quite a few double partial swings available, and won't be punished very often.

Yup, QJ109xx in both majors would comfortably fit the bill if you don't have an opener for that.

[DavidKok]

I honestly can't remember the last time I held a 6-6 in the majors, never mind QJ109xx-QJ109xx or better. The 'miracle' I claimed earlier was a mistake, sorry for that. Let's take it step by step and see if we don't happen to agree. I agree with most of the points Stephen mentioned, except:

• You list that N/S might be reluctant to double, but conversely if they have a hand to set you the double is close to guaranteed on this sequence. The normal argument for 'law protection' on the 2-level compares undoubled undertricks against defending, but I think that on this auction North in particular is in an excellent position to judge the hand. This increases the risk and also decreases the upsides, as N/S might run to 3m making. At any other vulnerability, or without an initial pass by East, I think the situation is not so clear and action is safer.
  
• I am a big fan of stretching to bid 5-5 in the majors (or longer), provided my points are primarily in my long suits. First seat V against NV I happily open starting at 9 HCP. If instead I have 6-4 or even 6-5 the requirements are even lower. This narrows the window for the sequence "P-(1NT)-P-(P); X" by eliminating precisely a set of hands on which the reopening double is relatively safe.

I agree with your other points. Your suggestion for a simulation sounds very interesting. My main point was that most hands that are worth acting on should have acted on the first round ("quick in, quick out"), and after eliminating those bidding is a lot more risky.

pilowsky, on 2020-September-16, 05:19, said:

I hear that term "Good Bridge" a lot. To me only the score is important. What is good Bridge to you?

The very same. To be more precise, during the play I want to take as many tricks as possible, and during the bidding I want to put my partner in the position to pick the final contract as quickly as possible (or, failing that, pick the right contract).

[bluenikki]

Stephen Tu, on 2020-September-16, 05:54, said:

West is marked with values, as North did not try for game

Yes West is marked with values. But enough to defeat a game if the opponents wake up?

YOU have no reason to think they can't make game, other than the fact that they didn't try to reach one.

Most obvious danger is that they have 10 cards in a minor and 24 hcp.

[Stephen Tu]

bluenikki, on 2020-September-16, 06:59, said:

Yes West is marked with values. But enough to defeat a game if the opponents wake up?

YOU have no reason to think they can't make game, other than the fact that they didn't try to reach one.

Most obvious danger is that they have 10 cards in a minor and 24 hcp.

This might be theoretically possible but in my experience it practically never happens. The opponents might get to their 3m/4m and find out they make 5 in practice, but they don't actually bid it. Certainly balancing them into making game is going to be insignificant compared to the frequency of partial swings in your favor, this wouldn't be a big enough risk for me to really worry about it at all.

[Cyberyeti]

DavidKok, on 2020-September-16, 06:52, said:

I honestly can't remember the last time I held a 6-6 in the majors, never mind QJ109xx-QJ109xx or better. The 'miracle' I claimed earlier was a mistake, sorry for that. Let's take it step by step and see if we don't happen to agree. I agree with most of the points Stephen mentioned, except:

I've held something like void, Qxxx, x, A10xxxxxx over a strong NT, but it was a 1N overcall rather than opener, not sure I'd have opened that given a chance.