[jdonn]

whereagles, on May 6 2008, 03:36 AM, said:

If I allow for a 6 card heart suit, we get to a good game 32% of the time. If I demand a 7 card suit for the pree, game is good only 24% of the time. Those are bad odds, so I have to concede pass is statistically the better bid.

This is fine, but I'm saying game is worse since, lets say you get to a "good" game 28% of the time, but a good game will only make say, 75% of the time. So you are only bidding and making game like 21%, plus the few times it was a bad game but made.

[whereagles]

awm: good analysis. Thx for your time. Conclusion seems to be "it sucks not to have 3 spades"

jdonn: yeah, "good game" doesn't equal "making game".

I now think we probably must pass this one. Simulations show that lack of the stopper in this case is much more important than in general. Bid 3NT or dbl if you have to (or vs opps that open on raggy suits), but it's not the percentage play.

[awm]

Here's what I think is interesting. Suppose we just compare pass vs. bidding 4♣. We assume that partner will pass 4♣ holding two small hearts in a fairly flat hand and less than 10 hcp or so. In this case:

Pass wins boards: 1, 2, 5, 6, 12, 15, 17
4♣ overcall wins: 3, 7, 9, 13, 14, 19, 20
Tossup: 4, 8, 10, 11, 16, 18, 21

That looks pretty dead-even to me! Yet bidding 4♣ has received almost no discussion.

[han]

awm, on May 6 2008, 02:51 PM, said:

Here's what I think is interesting. Suppose we just compare pass vs. bidding 4♣. We assume that partner will pass 4♣ holding two small hearts in a fairly flat hand and less than 10 hcp or so. In this case:

Pass wins boards: 1, 2, 5, 6, 12, 15, 17
4♣ overcall wins: 3, 7, 9, 13, 14, 19, 20
Tossup: 4, 8, 10, 11, 16, 18, 21

That looks pretty dead-even to me! Yet bidding 4♣ has received almost no discussion.

I think 10 and 11 are raises so those are plusses for pass (on both 4C turns a plus into a minus).

You rank 13 as a win for 4C while 5C is not guaranteed and 3H is down several. As you point out yourself, pass is not necessarily the end on 14 yet you ranked it as such.

On the other hand, pass on 15 is not clearly a winner.

On 18 4C is indeed down 1 but partner is never passing 4C. Seems a winner for pass. On 21 again I'd expect partner to raise 4C so that would be another minus for 4C.

So I don't agree on exactly one third of the hands. Some of them I think you were quite wrong, on others it is much closer, and on the majority I completely agree with you. I guess it shows that the results of 4C on this hand depends very much on how partner will bid. I'd like my partner to expect a better suit and he shouldn't be afraid to raise on Kx if he has a decent hand or if he has a 7-count, 3-card support and heart shortness.

Also, the assumption that opener has a 7-card suit is a strong one and wouldn't apply in modern bridge white against red. If you lower it to 6 then partner's expected clubs go down a bit and hearts go up a bit.

Still very interesting to go over these hands, thanks for posting them.

[jdonn]

awm, on May 6 2008, 02:51 PM, said:

Here's what I think is interesting. Suppose we just compare pass vs. bidding 4♣. We assume that partner will pass 4♣ holding two small hearts in a fairly flat hand and less than 10 hcp or so. In this case:

Pass wins boards: 1, 2, 5, 6, 12, 15, 17
4♣ overcall wins: 3, 7, 9, 13, 14, 19, 20
Tossup: 4, 8, 10, 11, 16, 18, 21

That looks pretty dead-even to me! Yet bidding 4♣ has received almost no discussion.

Sorry 4♣ doesn't even merit discussion. It completely closes out the two biggest reasons for entering over 3♥ in general (3♥X and 3NT), and makes little progress toward the third (4♠). If clubs break badly you will get MASSACRED, not merely beaten. On some of the hands you show 4♣ winning or a tossup, I either disagree with your analysis of what will happen next, or note that LHO will raise anyway after which I'm certainly doubling for penalty when 4♥ comes back, for better or worse. I mean you really think partner is not raising if you bid 4♣ vulnerable on the four level on 10, or 11, or 18, or 21? I can't even take that thought seriously.

BTW giving RHO 7 hearts is not realistic. In modern bridge when they are w/r, it obviously depends on the opponents, but for some 6 is almost guaranteed, for others it's a tossup. If you increase partner's expected heart length passing does even better.

[awm]

On 10, I stated my assumption that partner "doesn't raise on balanced 8-counts." Your style may differ of course.

On 11, I do not raise partner to the five-level on Hx. I would bid 4♦ (forcing) here to cater to a possible 4-4 fit in spades or diamonds, expecting to reach 5♦. This is an okay game (but not a great one). Again, your style may differ. Obviously if you play that 4♣ in this auction "always shows six clubs and a really nice hand" you will tend to do worse by bidding it than if you fairly often bid 4♣ on good hands with only five cards in the suit (or minimum openings with six nice clubs).

On 13, 5♣ is actually extremely good. Say you get a heart lead, win the ace, pull trumps in two rounds. If trumps are 2-2 you are now ice cold. If LHO has 3 clubs, you continue with a heart to towards the jack. RHO wins and returns a spade (best) but you hop ace, cash two diamonds, ruff a diamond, and discard a spade on dummy's heart. A spade lead from LHO probably establishes a spade honor for a heart discard. So basically you just need clubs not 4-0 here.

Again on 18, I would have no problem passing 4♣. The heart king is waste paper and you're not ruffing anything. This hand is not much better than xxx xxx xxx Kxxx especially if partner's 4♣ bid is a six-bagger. If you raise to 5♣ on that then your 4♣ overcalls are much different than mine.

I can try some sims where opener has six hearts later today. I'd suggest that these hands probably should include six hearts to two top honors and not be 6322 type patterns.

[jdonn]

Think of it this way. Most experts avoid overcalling on the 2 level on suits like Axxxx because it's too dangerous (most would on a hand this good, but I just mean as a general statement.) And here you are overcalling on such a suit on the 4 level, two levels higher! At matchpoints where they will double you! Vul!! WHERE LHO HAS A PENALTY DOUBLE AVAILABLE which isn't even true on the 2 level!!!

You say you would bid 4♦ on 11? JTxx? Sorry I don't believe you. Are you going to tell me partner won't raise that with three card support?

One last point.

awm, on May 6 2008, 01:14 PM, said:

6.

♠8654
♥Q63
♦T876
♣T7

Any non-pass action here is probably awful.

By "probably awful" don't you mean "definitely completely awful"? This is not a 75% chance of turning -100 into -500, that would be "probably awful". This is a near 100% chance of turning +50 into -1400 or -1700. And you shrug it off as "probably awful"! Was dropping the atom bombs on Japan a 'potential inconvenience' for them too?

[han]

If your opponents never open 3H white against red on x QJ10xxx Kxxx xx or xx KQJ10xx xxx xx then you shouldn't include those hands.

If your partner bids J10xx of diamonds rather than raising with Kx then that certainly makes overcalling more attractive. I should warn my regular partners that I will make fun of them if they ever commit such a sin.

[awm]

A few points here:

1. It's matchpoints. All that matters is frequency, right? The fact that 4♣ occasionally goes for 1100 is just a "lose" at this form of scoring.

2. Even if it was IMPs, an awful lot of the hands where bidding wins are game swings (win 10!) While a few of the hands where bidding loses are huge disasters (lose 15 maybe?) a number of them are just turning a small plus into a small minus (lose 5 or so).

3. The argument that you want to be playing 3NT or defending 3♥X on this hand is really neither here nor there. If you pass 3♥ you are not getting to play 3NT or defend 3♥X either. The question is more: do we want to defend 3♥ undoubled or would we rather declare in one of 4♣/5♣/4♠/5♦?

4. This is both a very good hand (lots of controls too) and has fairly short hearts. I think it's pretty obvious to bid over a 1♥ or 2♥ opening regardless of the suit quality. The fact that RHO opened 3♥ doesn't make us much less likely to have game, especially at these colors where some people open 3♥ on weak two bids. It does slightly increase the chances of going for a number (because we are a level higher) but usually going for a number is an awful matchpoint result regardless of whether it's 800 at the three-level or 1100 at the four-level.

5. I think that raising 4♣ to 5♣ on hand 18 is a serious error. Your hand contributes next to nothing on offense; the ♥K is almost surely useless since partner would bid 3NT on almost all hands with heart A/Q that are good enough to bid 4♣. You basically need overcaller to have ten tricks in hand for a 5♣ call to be right; give partner ♠AK ♥xx ♦Ax ♣AJTxxxx, which I think is substantially better than a minimum 4♣ call, and you still can't make five clubs.

[Echognome]

I'll probably get flagged for piling on, but I think on hand 20 there's a good chance you'll be declaring 4♣ in the 5-1 fit.

[jdonn]

awm, on May 6 2008, 03:52 PM, said:

3. The argument that you want to be playing 3NT or defending 3♥X on this hand is really neither here nor there.

The point is the potential for gain is less, so there is less reason to do something risky. That point is both here and there. Neither passing nor 4♣ will achieve 3♥X or 3NT, but only one entails huge risk.

BTW going for monsterous numbers is both detrimental to morale, and worth 0 mps (whereas other disasters might give you a 10% board or something).

Quote

The fact that RHO opened 3♥ doesn't make us much less likely to have game, especially at these colors where some people open 3♥ on weak two bids. It does slightly increase the chances of going for a number

Slightly??? See there you go again Chalk this up to the same tendency by which you say "probably awful". 4♣ is both more likely than 3♣ to go down, by a fair margin, and more likely than 3♣ to be doubled, by a fair margin.

Quote

give partner ♠AK ♥xx ♦Ax ♣AJTxxxx, which I think is substantially better than a minimum 4♣ call, and you still can't make five clubs.

If partner has Jx like on the actual hand, the heart ace or queen (which they would surely lead) onside will let you make. Meanwhile if AQ is over the king they probably make 4♥ on this example!!!

[655321]

awm, on May 6 2008, 02:51 PM, said:

Here's what I think is interesting. Suppose we just compare pass vs. bidding 4♣. We assume that partner will pass 4♣ holding two small hearts in a fairly flat hand and less than 10 hcp or so. In this case:

Pass wins boards: 1, 2, 5, 6, 12, 15, 17
4♣ overcall wins: 3, 7, 9, 13, 14, 19, 20
Tossup: 4, 8, 10, 11, 16, 18, 21

That looks pretty dead-even to me! Yet bidding 4♣ has received almost no discussion.

I understand you have thrown out a number of hands (those where partner would balance). Yet, you are now using the remaining hands to justify a horrible 4♣ bid.

How does overcalling 4♣ work when partner would have balanced?

Also, agree with the posters who expect a 6 card preempt much (most?) of the time.

[han]

655321, on May 6 2008, 05:10 PM, said:

I understand you have thrown out a number of hands (those where partner would balance).  Yet, you are now using the remaining hands to justify a horrible 4♣ bid.

How does overcalling 4♣ work when partner would have balanced?

What??? I started looking at the hands without reading the starting assymptions, this is truly ridiculous.

You can't judge the merits of a 4C call if you throw out the hands where partner would balance with 3NT. Also, how often will partner move to bad slam expecting a better club suit?

All in all it seems a clear case of manipulating the data in order to draw the conclusion you want to see. First, you start with assumptions that are in favor of your chosen call (require that the opponents have 7 hearts which is highly unreasonable at these colors). Second, you throw out a number of hands that seem irrelevant but where pass will likely work better. Thirdly, you require partner to respond to your chosen call as if this is a standard 4C call, again making the outcomes look far better than they actually are (with the absurd 4D suggestion on the hand with J10xx Kx in the minors as ice in the pudding).

If this was done purposely then I think it is disgusting.

[Echognome]

han, on May 6 2008, 02:34 PM, said:

655321, on May 6 2008, 05:10 PM, said:

I understand you have thrown out a number of hands (those where partner would balance).  Yet, you are now using the remaining hands to justify a horrible 4♣ bid.

How does overcalling 4♣ work when partner would have balanced?

What??? I started looking at the hands without reading the starting assymptions, this is truly ridiculous.

You can't judge the merits of a 4C call if you throw out the hands where partner would balance with 3NT. Also, how often will partner move to bad slam expecting a better club suit?

All in all it seems a clear case of manipulating the data in order to draw the conclusion you want to see. First, you start with assumptions that are in favor of your chosen call (require that the opponents have 7 hearts which is highly unreasonable at these colors). Second, you throw out a number of hands that seem irrelevant but where pass will likely work better. Thirdly, you require partner to respond to your chosen call as if this is a standard 4C call, again making the outcomes look far better than they actually are (with the absurd 4D suggestion on the hand with J10xx Kx in the minors as ice in the pudding).

If this was done purposely then I think it is disgusting.

To be fair to Adam, the initial question was whether 3NT might work better than Pass. The whole 4♣ business came as an aside. One which many of us disagree with, but certainly wasn't the point of the initial analysis.

[awm]

There are a lot of questions that come up if you include hands where partner balances. Out of question, say you pass with this hand. What do you bid at next turn if:

(1) Partner balances with a double?
(2) Partner balances 3NT?
(3) Partner balances 3♠?
(4) LHO raises to 4♥ and this passes back to you?

[han]

OK I got angry so there is Matt telling me I should be nice to people and he is probably right.

Adam raised the point of the 4C call. When this was ignored because it was a terrible suggestion he posted a analysis showing that it was far better than what people thought. It really seems that he went out of his way to manipulate his data to make a case for 4C. If you think that this is all fine and dandy then you have different standards.

[awm]

I'll point out that:

(1) Yes, I disregarded hands where partner balances. The issue is that on these hands you will often have a complicated auction regardless of whether you start the bidding with pass or 4♣. For example, what do you bid if partner balances 3♠? Mightn't it depend on whether 4♣ now would be fit-showing, whether it would be forcing, how light partner tends to balance here, what a cuebid means, etc? If you bid 4♣ and partner has a really good hand, how much does he push for slam? Does he always raise you to slam on Hx with a good hand? Do you play transfer advances here? Would 4♦ over 4♣ be forcing? Is 4NT to play or keycard? On many of the hands where partner balances you are actually close to the slam range...

(2) Yes, I assumed seven hearts for the opening bid. If someone can give me a good guideline for what a six-card preempt looks like I will put it into my sim. The issue here is more that it seems like there is no guideline. I refuse to believe that most good players will open 3♥ on absolutely any hand with six hearts and 5-11 hcp at favorable colors. Of course there do exist 3♥ openings on six cards; my suspicion was that these would involve a good suit and some shape (and I'd be happy to do the sim) but others have shouted me down and insisted that KQJTxx and 6322 is a 3♥ bid. Certainly I know some people will open 3♥ with a five card suit. Anyway, if somebody gives me a reasonable standard for what these hands look like I'm happy to include them.

(3) I'm far from the only person who is purposely analyzing things in a skewed way. For example, many people like to point out that on hands where I already acknowledged that pass is the best bid, you could go for a big number by bidding 4♣. Okay, so what? It's still MP scoring. And a lot of people have suggested that it's right to raise 4♣ to 5♣ on various balanced 8-counts with 3-4♣, even if including an awful heart holding like Kx. Well, if your agreement is that a 4♣ overcall shows "ten tricks in hand" then very well, but is this agreement standard? Is it even reasonable? I stated my assumption up front that "8 points balanced with xx in hearts is not a raise" -- no one else has made any statement about what a 4♣ bid "should look like" or what qualifies as a raise, except to say that they would raise on a wide range of hands (even balanced 8-counts or balanced hands with two clubs) where I think raising is awful. One person (Han) is questioning my judgement for saying 4♣ is better when it gets you to a game that's roughly 87% to make (clubs not 4-0). Another (Matt) is suggesting that passing 4♣ with a 5161 six-count is reasonable, at the same time that another (Han) is suggesting that we should raise 4♣ to 5♣ on what's effectively a balanced five-count. And then we have another (Josh) stating that reaching a no-play 5♣ opposite a 4♣ overcall of AK xx Ax AJTxxxx -- a super-max 4♣ call with a seven card suit and prime cards -- is a good idea because, well, overcaller could have the ♥J too and one of the two heart honors could be onside with the other being lead. But couldn't overcaller also have, say, the ♦K instead of ace, or the ♠Q instead of king, or one fewer club and one more card in any other suit?

No one is really being objective here. I include myself in that. But I am not trying to argue that 4♣ is better than passing, only that it is close enough to be worth some consideration. The possibility of a 3NT call, which seems honestly quite awful, received some support and discussion. I don't understand why a 4♣ call is so ludicrous as to not be worth at least the same consideration. You have a making game on these hands seemingly about 1/3 of the time that partner can't balance all of which games you miss by passing 3♥.

[jdonn]

awm, on May 6 2008, 06:15 PM, said:

And then we have another (Josh) stating that reaching a no-play 5♣ opposite a 4♣ overcall of AK xx Ax AJTxxxx -- a super-max 4♣ call with a seven card suit and prime cards -- is a good idea because, well, overcaller could have the ♥J too and one of the two heart honors could be onside with the other being lead.

I didn't reaching a no play game is a good idea, I said raising is a good idea. You gave a carefully constructed hand that is a very good hand on which game has no play if the heart is off, as it rates to be. I pointed out that a very minor and irrelevant-looking change gives you a chance to make, and that even on the hand you provided it was likely a good save. What is so objectionable about that? Fine no changes just your actual example. If clubs are 2-0 then either 5♣ or 4♥ makes so raising is very right. If clubs are 1-1 I traded +50 for -200 or +100 for -100, if the heart is off, but sometimes the heart ace will be on and game will make anyway. So it's a good save about half the time and makes some of the rest of the time opposite your own example. That is the raise you dub a serious error. Of course when you go for 1700 it's just "probably awful" and "only one board" (that second one is not a direct quote of your words, but certainly of your belief).

You aren't even being objective when trying to defend yourself against accusations of not being objective by pointing out that other people are also not objective. I mean you admit some open 3♥ on a five card suit but force a seven card suit on opener. You say why am I being treated so harshly when 3NT was discussed. It had one person backing it, who probably knew no one would agree with him, and he eventually concluded based on his own simulation it was wrong. Sorry for everyone being so kind to him! Don't complain that your 4♣ bid isn't getting "the same consideration" as the 3NT suggestion. It is, which consists of everyone else thinking it's a bad bid. Well ok at first everyone ignored it, just as they ignored a suggestion near the top of the second page to double, because neither call merits a response. Then you came back again essentially asking everyone to please consider 4♣. So we did.

Sorry I don't know what else to say. It's so evident that overcalling on the 4 level in a minor on Axxxx is ridiculous that I'm not sure how else to respond any more. When you replied partner will bid 4♦ on JTxx with Kx of clubs, then you brought a lot of the harsh treatment on yourself.

awm, on May 6 2008, 05:51 PM, said:

There are a lot of questions that come up if you include hands where partner balances. Out of question, say you pass with this hand. What do you bid at next turn if:

(1) Partner balances with a double?
(2) Partner balances 3NT?
(3) Partner balances 3♠?
(4) LHO raises to 4♥ and this passes back to you?

(1) 4NT
(2) 4NT
(3) 4♠
(4) DBL

I don't claim these are sure to work brilliantly, but I also don't consider any of them brutally stressful bidding problems. Maybe I should cuebid on 3, it shows support but A AK A might compensate.

[Echognome]

awm, on May 6 2008, 03:15 PM, said:

Another (Matt) is suggesting that passing 4♣ with a 5161 six-count is reasonable, at the same time that another (Han) is suggesting that we should raise 4♣ to 5♣ on what's effectively a balanced five-count.

Yep and I don't think they are inconsistent. I'm not saying you WILL play 4♣, I just think it's entirely possible. If I have a six count and what looks like a big misfit, why am I bidding again? Obviously in hopes that we have a fit. Partner's expected shape is something like 2=2=3=6 or 3=2=2=6, but obviously it could be better or worse. I certainly wouldn't expect 2=2=4=5! So what if I bid 3♦ or 3♠ even (the choice made when I gave this hand to a friend) and now I hear 5♣ or even worse, double by the opponents. It's all a bit hairy.

Now on the other hand, when I have a known fit, things are suddenly not so scary.

[han]

awm, on May 6 2008, 06:15 PM, said:

(3) I'm far from the only person who is purposely analyzing things in a skewed way.

I agree, in fact, we have seen posts that were far worse in this respect. Often these are from posters that come, get laughed at and leave. You are a frequent poster with a very good understanding of statistics, should we not hold you to higher standards?

I feel very strongly that those who have some understanding of statistics should treat simulations with more care.