Wrong shape over preempt

[mtvesuvius]

I'm going to pass again :angry:

[skjaeran]

Passing twice.

[rogerclee]

I'd pass the second time, since I don't want to defend 3♦x.

[Codo]

I preefer a double pass in soccere to one in this hand, but I play 3 Diamonds undoubled.

[Free]

Anyone playing Raptor 2NT? :(

 

Pass twice

[Cascade]

GIB does some screwy things ...

 

however with these settings:

 

East has 6-9 HCP and six diamonds;

East does not have a four-card major; (not my style of pre-empting)

East might have up to five clubs;

 

NS vulnerable;

East dealer;

 

East is forced to open 2♦;

 

South has the exact hand in this thread's opening post - actually I gave it the worst spots possible ♠ AK32 ♥ A2 ♦ K2 ♣ 98432.

 

I played 100 hands three times.

 

The first time I forced south to overcall 2♠.

 

The second time I forced south to pass.

 

The third time I forced south to double.

 

GIB and its own simulations (judgement) took over from there.

 

Cross IMPed

 

2♠ +106 IMPs

Pass -95 IMPs

Dbl -11 IMPs

 

Head to Head

 

2♠ vs Pass +56 IMPs

2♠ vs Double + 50 IMPs

Double vs Pass +39 IMPs

 

The biggest losses:

 

1. GIB sold out to 3♦ after 2♦ 2♠ 3♦ with 1=3=3=6 11 count (Ace Ace-King). After pass and double both other tables played 6♣X for +1540.

 

-32 IMPs

 

2. 2♠ was hit for 1100 after pass and double EW played 4♠ on a 5=2 fit for -100 (EW). GIB with a seven count or so and seven hearts did not run but I suppose it did have doubleton small spade.

 

-30 IMPs

 

3. After 2♠ GIB raised to 4♠ with ♠ xxx and a 13 count which drifted 3-off. Amazingly after pass (and a delayed takeout double) and an immediate takeout double GIB played 4♥ +620 and +650 amazingly because the trump suit was A2 opposite Q973!!! Yes a 4-2 fit.

 

-28 IMPs

 

4. After 2♠ GIB raised to 3♠ competitively with ♠ QJT65 ♥ JT96 ♦ T86 ♣ K and made nine tricks. After pass and double GIB got to 4♠ Doubled and made ten tricks +790

 

-24 IMPs

 

5. After 2♠ GIB raised to 4♠ with ♠ Q875 ♥ KQJ653 ♦ 7 ♣ QT and drifted one-off when after pass and double GIB played the 6-2 heart fit making - so much for four-four fits ;-)

 

-24 IMPs

 

6. After 2♠ GIB raised to 4♠ with ♠ QJ4 ♥ KJ9763 ♦ 54 ♣ KQ which failed. 4♥ made after pass and double.

 

-24 IMPs

 

Two other times 2♠ cost more than 20 IMPs.

 

Thirteen times it cost between 10 and 19 IMPs.

 

Altogether it lost 496 IMPs on 43 hands.

 

Nine times there was a net 0 swing for 2♠ (seven of those times it was the 0 IMPs all around and twice what it gained against one action it lost against the other alternative.

 

Which left 48 hands on which it gained a total of 602 IMPs.

 

28 of those were 10 or more IMP gains.

 

12 were 20 or more IMP gains.

 

The biggest gains were:

 

5. 4♠ made opposite ♠ 876 ♥ QJ76 ♦ 985 ♣ AQJ. 2♦ was passed out for +50 and 4♥ was bid after a takeout double (I think GIB's auction was a bit silly) 3NT might have made but having already bid heart GIB ripped 3NT to rebid its 4-card heart suit.

 

+26 IMPs

 

4. GIB played 5♣ making after starting 2♠ but strangely 6♠ after passing (GIB made a worse canape overcall at the four-level on the second round after 4D came back - maybe its partner should expect a two-suiter.) and 4♥ after a double (2D.X.4D.4H.P.P.P) - both failed.

 

+27 IMPs

 

3. 2♠ Doubled made +670 while 3♦ drifted one-off after a pass and after double 3♥ was hit for 800.

 

+28 IMPs

 

1=. 3♦ made after a 2♠ overcall -110 after pass and double 3♥ Doubled was played -1400

 

+30 IMPs

 

1=. 4♠ Doubled made an overtrick +990. After pass 5♦ doubled made -550. After double 5♦ doubled failed one-trick (its after 1am here and I am not going through the tricks to see why although it looks like it should make the ♦ A was not with the weak two bidder so the king was finessable I am guessing for some reason GIB played for the drop.

 

+30 IMPs

 

Perhaps more of these big gains were from GIB idiosyncracies or maybe just idiocies. It didn't show in the extreme results here that I have looked at in more detail but when I was playing a few hands on BBO with dealer set to predeal this hand and a weak two in diamonds GIB did some other strange things that were bad for this 2♠ overcall.

 

The genesis of this idea of canape overcalls for me came around 20 years ago when a friend gave me a hand ♠ x ♥ A10xx ♦ Qx ♣ AKQxxx or similar (I looked it up a couple of days ago but my memory is hazy now and I can't be bothered searching for it again - I posted it on rec.games.bridge in 1990!!!). Internet traffic was much lower then I got about 6 replies over a period of about a month (or perhaps it just wasn't an interesting hand).

 

What I didn't write on rec.games.bridge is that I generated a large (considering I did most of the analysis by hand) number of hands and went through and looked at reasonable auctions after various starts and made ball park guesses at what would happen. I was very surprised when the big winner appeared to be overcalling 3♥. The big loser was overcalling 4♣ - basically even when it was right you gained nothing +130 just isn't worth much.

 

Since then a few times similar hands have come up and I have usually backed the simulations and overcalled the four-card suit with mixed results but I am pretty sure I am ahead.

 

Nevertheless I am a little surprised by the results I got here. Over a 3-level preempt to introduce your minor you need to go to the 4-level which is a big minus but that is not the case here. Although perhaps that is balanced by the fact that this minor is so horrid.

 

I might set GIB running for another 100 hands overnight and see how reproducible the results are for this hand.

 

As an aside the idea that 4minor is such a useless overcall was one of the most compelling reasons that I switched to playing leaping and non-leaping Michaels.

 

I particularly remember one hand that came up - well I remember the situation not the hand it was the first time at the table i had tried one of these canape overcalls - where the opponent opened 3♦ and I was 4=1=2=6 and overcalled 3♠. At the time based on the simulation it was a WTP? hence my "WTP?" in my first post in this thread. Then "LOL" since I was sure it was not going to be a mainstream opinion. I was not intending to mock anyone. :-)

[jdonn]

Thanks for the explanation, at least I understand now. :)

[awm]

[hv=d=e&v=n&n=stxxhxxxdaqxcaqjx&w=sqj8xhj98xxdjtckx&e=s9xhkqtd98xxxxctx&s=sakxxhaxdkxc98xxx]399|300|Scoring: IMP[/hv]

 

This was the full hand. Passing twice was not a success. 3♦ went down three tricks (it's down four on best defense). N/S are of course cold for 3NT, which was reached at the other table when east did not open.

 

Pretty much any non-pass call at either turn is likely to lead to success here; even if north decides to bid game in spades after a 2♠ overcall, the lie of the cards is friendly enough that this should make.

 

I suspect that having passed once, it is actually right to double at second turn. We have a lot of controls and four fairly likely tricks in hand, and the fact that LHO did not look for game suggests that partner is likely to make a contribution as well. But it may depend on your opponents.

[Cascade]

I might set GIB running for another 100 hands overnight and see how reproducible the results are for this hand.

Another 100 hands:

 

2♠ +387 IMPs

Pass -315 IMPs

Double -72 IMPs

 

Combined for 200 hands:

 

2♠ +493 IMPs

Pass -410 IMPs

Double -83 IMPs

 

Head to Head

 

2nd 100:

 

2♠ versus Pass +235 IMPs

2♠ versus Double +152 IMPs

Double versus Pass +69 IMPs

 

The 200 Hands Combined:

 

2♠ versus Pass +291 IMPs - nearly 1.5 IMPs per board

2♠ versus Double +202 IMPs - over 1 IMP per board

Double versus Pass +108 IMPs - over 0.5 IMPs per board

 

Maybe GIB has some inherent bias towards 4-card overcalls but maybe not.

 

Maybe its a WTP? These numbers are certainly starting to convince me and although I was quite flippant at the beginning I wasn't actually convinced 2♠ was right but did think it might be better than others would think. I thought it would be much closer than these numbers are suggesting.

 

But draw your own conclusions.

[awm]

Looking at the gains/losses here, it seems like GIB is doing a lot of crazy stuff.

 

You have one where GIB stopped off to double 4♠ (which made!) when 5♦ was making. Why do that? Sounds like terrible IMP strategy. And another hand where GIB doubled a cold 2♠ for no apparent reason.

 

There was another where partner has enough to bid 3♥ over our pass (and this goes for 1400 even though we have three quick tricks and never bid!) but cannot find a call after we overcall 2♠ and opponents raise to 3♦. How likely is that?

 

I think the GIB craziness is taking over here, and wouldn't read too much into the results.

[Cascade]

[hv=d=e&v=n&n=stxxhxxxdaqxcaqjx&w=sqj8xhj98xxdjtckx&e=s9xhkqtd98xxxxctx&s=sakxxhaxdkxc98xxx]399|300|Scoring: IMP[/hv]

 

This was the full hand. Passing twice was not a success. 3♦ went down three tricks (it's down four on best defense). N/S are of course cold for 3NT, which was reached at the other table when east did not open.

 

Pretty much any non-pass call at either turn is likely to lead to success here; even if north decides to bid game in spades after a 2♠ overcall, the lie of the cards is friendly enough that this should make.

 

I suspect that having passed once, it is actually right to double at second turn. We have a lot of controls and four fairly likely tricks in hand, and the fact that LHO did not look for game suggests that partner is likely to make a contribution as well. But it may depend on your opponents.

GIB which I assume is out of its book or database bids and so basing its actions on simulations seems to act after 2♦ Pass 3♦ Pass Pass ...

 

I haven't looked at all of the hands but what have looked at it seems to double although I have seen it bid spades - at least once spectacularly unsuccessfully over a 4♦ raise.

[hanp]

Very interesting simulation Wayne!

 

I don't have much experience with GIB, apart from playing on BBO. The big swings look strange, it would be interesting if you could go through some of the smaller swings and tell us what your impression is. (No need to post the hands, I would just like to know whether you think GIB's play and bidding is realistic or absurd or somewhere in between)

 

Given the huge difference between the first set of 100 hands and the second I think 200 is still a fairly low number to say anything with great confidence.

[Cascade]

Very interesting simulation Wayne!

 

I don't have much experience with GIB, apart from playing on BBO. The big swings look strange, it would be interesting if you could go through some of the smaller swings and tell us what your impression is. (No need to post the hands, I would just like to know whether you think GIB's play and bidding is realistic or absurd or somewhere in between)

 

Given the huge difference between the first set of 100 hands and the second I think 200 is still a fairly low number to say anything with great confidence.

Yes 200 is a very low number.

 

Although we are happy to decide major championships by many fewer boards and some conclude that the obviously the better team won on much less evidence.

 

Yes GIB does screwy stuff but it would be a mistake or at least a huge leap of faith to think that it only does this screwy stuff to benefit some whacky 2♠ canape overcall.

 

I'll take a look at a few hands around the +10 and -10 IMP marks - that is one game swing or two part-score swings. +/- 5 IMPs and we are looking at nothing boards +110 vs +50 (at both other tables) would give us +4 IMPs.

[Cascade]

1. 2♠ lost 10 when 5♦ was cold but on both contested auctions GIB got to 5♦ but on the uncontested auction GIB played a 4-3 heart fit which failed. The 2♠ overcall lost when partner with six spades dived unfavourable in 5♠ with six-card supprt. Double lost when 5♦ was hit and made - it wasn't completely unreasonable to hit 5♦ as it would have failed if the weak two bidder had the ♦ A.

 

This seems ten unfortunate IMPs out for 2♠. Well it could have passed 5♦ and saved a few of those ten.

 

2. 2♠ lost when 3♦ made. Partner did not compete with QJ86.74.Q93.JT65

 

But when we passed GIB responded 2♥ forcing which was raised to 3♥. We then doubled (reasonable) and partner bid 3♠ which made.

 

When we doubled GIB also sold out to 3♦ but this time beat the contract.

 

In diamonds there were three sure side suit losers and the trumps were Jx opposite A1087xx. After 2♠ GIB picked up Q9x onside where as after double GIB played ace and another.

 

I think this is a little unlucky for 2♠. Its partner might have competed to 3♠ although 3♠ was a lucky make - clubs were 2=2 missing ♣ AKQ - and GIB might have made 3♦ at the other table.

 

3. 2♠ won when partner raised to game which made.

 

After double we also got to a spade game.

 

After pass partner somewhat conservatively thought we didn't have enough for game with Q9876.KQ876.QT.6 after we made a second round double.

 

A little lucky for 2♠

 

4. 2♠ won when we got to game.

 

After pass we also got to game.

 

After double partner tried Lebensohl and then an invitational 3♥. We declined. Partner had 11 HCP with both majors some might have forced to game and forced partner to pick a major. But this is the sort of ambiguity that comes from an off-shape double. Curiously though if we had doubled with 2=4 in the majors we would have found our game.

 

5. After 2♠ we got to 3NT making.

 

After Pass we defended 2♦ down three. Partner had a balanced 13 HCP with ♦ QJ9x.

 

After double we also got to 3NT. When partner jumped in hearts (forcing I suppose in a Lebensohl context) and we corrected to 3NT which no double partner was very happy with.

 

I didn't do any particular selection on these hands except that I choose the five hands where 10 IMPs were won or lost out of the second simulation.

 

Overall while GIB does do some screwy things I don't feel or expect that it is particularly biased against one sort of action it seems that mostly its actions were ok on the hands that I looked at in more detail (that is more detail than +100 -50 +170).

[hanp]

Yes GIB does screwy stuff but it would be a mistake or at least a huge leap of faith to think that it only does this screwy stuff to benefit some whacky 2♠ canape overcall.

 

I am not sure what you mean by this comment. If GIB bids very poorly then it seems a leap of faith to draw conclusions from such a simulation. I wouldn't assume the results are false either, I would just ignore them. In that case a better way would be to go over a large number of hands manually and try to figure out how good players would bid after 2S, double and pass.

[hanp]

Although we are happy to decide major championships by many fewer boards and some conclude that the obviously the better team won on much less evidence.

 

In this simulation you randomly deal hands and then let four GIBs who do some screwy stuff bid and play the hands.

 

In a major championship you randomly deal hands but then you duplicate them and let two groups of world class bridge players play them.

 

I think that it is clear thatt in the latter case the standard deviation is much smaller.

 

So it seems natural to me that if your first simulation suggests that 2S wins 56 IMPs against pass, to keep in mind that the actual expected value may be considerably smaller or larger, while if one world class team beats another world class team by 56 IMPs over 100 hands then it is more reasonable to assume that the winning team played better.

 

There are also other practical issues, obviously it is easier to let a computer simulate 1000 more hands than it is to keep the world class players for several more weeks.

[hanp]

Overall while GIB does do some screwy things I don't feel or expect that it is particularly biased against one sort of action it seems that mostly its actions were ok on the hands that I looked at in more detail (that is more detail than +100 -50 +170). 

 

Your sentence isn't clear to me.

 

Do you mean to say that you don't have any reason to expect that GIB is biased in a particular way? Or do you mean to say that you expect that GIB is not biased?

 

I have some experience with bidding against robots on BBO. At the time I tried this (two months ago?) the robots were extremely reluctant to make a penalty double. It was incredible. We practiced bidding over 4H. Of course it was a big winner to bid 4S very aggressively, they would never double when you stepped out of line. Now I don't know if these robots ran on the same GIB engine, but if it did then I would keep in mind that the results of such a GIB simulation might have nothing to do with the real world.

 

If the GIB you are using is much more trustworthy then that would be good to know.

[whereagles]

what a revolting 2♦ opener!

 

but who are we argue with sucess?? LOL

[jdonn]

It's interesting to me that I think we should double 3♦ on the way back whether it's takeout or penalty (if it's takeout I would pull 3♥ to 3♠ since I think the double initially gives the wrong impression of our hand).

[Cascade]

Although we are happy to decide major championships by many fewer boards and some conclude that the obviously the better team won on much less evidence.

 

In this simulation you randomly deal hands and then let four GIBs who do some screwy stuff bid and play the hands.

 

In a major championship you randomly deal hands but then you duplicate them and let two groups of world class bridge players play them.

 

I think that it is clear thatt in the latter case the standard deviation is much smaller.

This maybe true but I am not so sure that it is obvious.

 

I am currently running a little experiment to check (well at least get a feel for this) by having GIB play the hands from a recent World Championship Final and comparing the variance (or standard deviation) of GIBs results with the variance of the results at the two tables in the World Championship Final.

 

I will let you know the results.

[jdonn]

I am currently running a little experiment to check (well at least get a feel for this) by having GIB play the hands from a recent World Championship Final and comparing the variance (or standard deviation) of GIBs results with the variance of the results at the two tables in the World Championship Final.

I think this kind of experiment is very interesting and a much better idea than trying to use GIB to test the correctness of bids early in a competitive auction. I am curious to see the results.

[hanp]

Although we are happy to decide major championships by many fewer boards and some conclude that the obviously the better team won on much less evidence.

 

In this simulation you randomly deal hands and then let four GIBs who do some screwy stuff bid and play the hands.

 

In a major championship you randomly deal hands but then you duplicate them and let two groups of world class bridge players play them.

 

I think that it is clear thatt in the latter case the standard deviation is much smaller.

This maybe true but I am not so sure that it is obvious.

 

I am currently running a little experiment to check (well at least get a feel for this) by having GIB play the hands from a recent World Championship Final and comparing the variance (or standard deviation) of GIBs results with the variance of the results at the two tables in the World Championship Final.

 

I will let you know the results.

Very interesting but of course it has little to do with my comment (which was poorly phrased I admit but a good reader only needs half a word as they say in the Netherlands).

 

Suppose you let a completely deterministic computer program play the same hand twice. The variance would be 0, no matter how well the computer program plays bridge.

 

In your simulation you forced the computer program to make 3 different bids with the same hand. This is entirely different.

 

To get a similar comparison with the world championship hands, force dealer at the first table to always pass, and force the dealer at the other table to always open 1X in their longest suit. See what variance you get then...

[hanp]

I am currently running a little experiment to check (well at least get a feel for this) by having GIB play the hands from a recent World Championship Final and comparing the variance (or standard deviation) of GIBs results with the variance of the results at the two tables in the World Championship Final.

I think this kind of experiment is very interesting and a much better idea than trying to use GIB to test the correctness of bids early in a competitive auction. I am curious to see the results.

OK I just wrote that it was interesting but now I am not sure why it is. This simulation tests whether GIB is consistent in its choices, not whether what GIB does is good bridge.

[hanp]

As for a constructive comment:

 

What would be more interesting (to me, as a bridge player) would be to compare GIB's results with the results obtained by the two tables and look at the variance of the differences.

[Echognome]

Just thinking out loud. Would it make sense to cross IMP GIB's results with, say, the results from the World Championship or something along those lines. I'm just thinking of the "Play with the Champions" type competitions.