[Phil]

Youre on lead with:

63 T974 64 A9873

2N (20-21) - 3N - AP.

Its matchpoints

[Phil]

bump?

[bluecalm]

Heart, I am gonna run sim now to make sure
I am assuming 5-3-3-2 is in the range for 2NT ?

Lazy assymptions:
opener: 20-21, bal, any 5-3-3-2, no 5-4-2-2
resp: 5-10, no 5 major, no 4 major unless 4-3-3-3.
Results (the best lead is) after 1000 hands:

♠ - 681
♥ - 728
♦ -607
♣ - 627

From my experience tuning the assumptions to better represent resp hand or including 6-3-2-2's and 5-4-2-2's in opener hand won't change anything here.

[Phil]

bluecalm, on 2012-January-02, 09:53, said:

Heart, I am gonna run sim now to make sure
I am assuming 5-3-3-2 is in the range for 2NT ?

Lazy assymptions:
opener: 20-21, bal, any 5-3-3-2, no 5-4-2-2
resp: 5-10, no 5 major, no 4 major unless 4-3-3-3.
Results (the best lead is) after 1000 hands:

♠ - 681
♥ - 728
♦ -607
♣ - 627

From my experience tuning the assumptions to better represent resp hand or including 6-3-2-2's and 5-4-2-2's in opener hand won't change anything here.

Thanks Bluecalm.

Assumptions look reasonable to me. If anything the heart lead looks even better against quasi-balanced shapes like 6322.

I led a heart. Partner has ♣KQx. LHO (who has 39K MPs) sort of mumbled, "heart lead won't be the field's choice". At 21 tables, the heart was only chosen twice.

[bluecalm]

Quote

I led a heart. Partner has ♣KQx. LHO (who has 39K MPs) sort of mumbled, "heart lead won't be the field's choice". At 21 tables, the heart was only chosen twice.

I experienced the same sort of comments when playing with people who are better card players than me but don't get leads. Tough luck

[aguahombre]

Mama-Papa would do well.

[awm]

The club lead is probably better at IMPs.

A lot of players don't distinguish much between forms of scoring, and are taught to lead "fourth from longest and strongest." I don't think the field choice is that surprising here.

[bluecalm]

Quote

The club lead is probably better at IMPs.

Leads to defeat 3NT (1k hands):

♠ - 61
♥ - 51
♦ - 44
♣ - 88

[Free]

It's not the first time we've ran similar sims for 5-4 distributions. The result is always the same: to defeat the contract (= imp lead) you better lead your 5-card suit, but to maximize tricks (= MP lead) you better lead your 4-card suit. Doing the best (on average) thing doesn't mean you'll never lose ofcourse

[awm]

I suspect the MP lead has more to do with honor location.

If hearts were A987 and clubs T9762 then I suspect clubs may be the MP lead.

[bluecalm]

Quote

If hearts were A987 and clubs T9762 then I suspect clubs may be the MP lead.

♠ - 679
♥ - 695
♦ - 626
♣ - 687 (interestingly 7 or 6 is better than T or 9 by quite a few cases, 668 for T, after quick survey it's because of hands with stiff honor in partner's hand)

It's close so I run it again:

♠ - 663
♥ - 684
♦ - 623
♣ - 666 (and 658 for T/9)

It's close. Simulation has biases. I am not sure if we should believe that one or not.

[rogerclee]

I think the actual hand is close and interesting, I would lead a heart also. However, when commentating about double dummy lead simulations vs NT:

1) The results usually favor unusual leads. Making unexpected leads is hard on partnership defense, particularly if the lead is unreadable (here could be from HT9x) and/or you don't play some kind of smith. Your double dummy computer is always shifting back to clubs when it's right. In real life, partner often just plays you to have led from HT9x since he has nothing better to play for.

2) When underleading an honor blows a trick, it is likely to have the same effect double dummy as single dummy. The HT on this hand can be dangerous sometimes single dummy when it was never a problem double dummy. This can happen when partner has Qx sometimes, or just when declarer's hearts were Axx opp QJxx and the normal line would have been to play on another suit and pitch 2 hearts from hand and hope the heart hook is onside, instead of just solving the suit at trick 1 for him for 3 tricks.

[Cascade]

bluecalm, on 2012-January-02, 09:53, said:

Heart, I am gonna run sim now to make sure
I am assuming 5-3-3-2 is in the range for 2NT ?

Lazy assymptions:
opener: 20-21, bal, any 5-3-3-2, no 5-4-2-2
resp: 5-10, no 5 major, no 4 major unless 4-3-3-3.
Results (the best lead is) after 1000 hands:

♠ - 681
♥ - 728
♦ -607
♣ - 627

From my experience tuning the assumptions to better represent resp hand or including 6-3-2-2's and 5-4-2-2's in opener hand won't change anything here.

Is that using Jack?

[bluecalm]

Quote

1) The results usually favor unusual leads. Making unexpected leads is hard on partnership defense, particularly if the lead is unreadable (here could be from HT9x) and/or you don't play some kind of smith. Your double dummy computer is always shifting back to clubs when it's right. In real life, partner often just plays you to have led from HT9x since he has nothing better to play for.

2) When underleading an honor blows a trick, it is likely to have the same effect double dummy as single dummy. The HT on this hand can be dangerous sometimes single dummy when it was never a problem double dummy. This can happen when partner has Qx sometimes, or just when declarer's hearts were Axx opp QJxx and the normal line would have been to play on another suit and pitch 2 hearts from hand and hope the heart hook is onside, instead of just solving the suit at trick 1 for him for 3 tricks.

I agree with both points. Still I've found out going through a lot of vugraph hands that if I made a lead suggested by dd simul every hand I would be the best 1st leader in the world.

Quote

is that using Jack?

DM PRO. It would be cool if someone could run it in Jack (I don't have it).

[Cascade]

bluecalm, on 2012-January-02, 19:34, said:

I agree with both points. Still I've found out going through a lot of vugraph hands that if I made a lead suggested by dd simul every hand I would be the best 1st leader in the world.



DM PRO. It would be cool if someone could run it in Jack (I don't have it).

Does DM Pro do those numbers for you or are they manually calculated?

[bluecalm]

It does. It's double dummy though while Jack is more like real play. I don't go through 1k hand manually lol.

[Cascade]

bluecalm, on 2012-January-02, 22:31, said:

It does. It's double dummy though while Jack is more like real play. I don't go through 1k hand manually lol.

Thanks I will have to investigate. I haven't discovered that feature of DMPro.

By manually I really thought there could be a dump of data and you had to do some calcutions on it which might be reasonably easy with a spreadsheet or similar not that you were sitting there adding up 1K numbers by hand.

[bluecalm]

Quote

Thanks I will have to investigate. I haven't discovered that feature of DMPro.

This is major feature of that program and its main part. If you have it, by all means investigate it's great fun and very powerful.

[rhm]

bluecalm, on 2012-January-02, 09:53, said:

Heart, I am gonna run sim now to make sure
I am assuming 5-3-3-2 is in the range for 2NT ?

Lazy assymptions:
opener: 20-21, bal, any 5-3-3-2, no 5-4-2-2
resp: 5-10, no 5 major, no 4 major unless 4-3-3-3.
Results (the best lead is) after 1000 hands:

♠ - 681
♥ - 728
♦ -607
♣ - 627

From my experience tuning the assumptions to better represent resp hand or including 6-3-2-2's and 5-4-2-2's in opener hand won't change anything here.

I am not sure what these numbers really tell us. I repeated the simulation, and my numbers based on my assumptions about dummy and declarer were closer, but the ranking was like yours, except that a black suit lead was equally good.
The biggest issue I have is the following:

While it is obvious that over-tricks are more important at pairs than at IMPs claiming that avoiding over-tricks is as important as defeating a contract is simply an oversimplification, just as much as claiming that over-tricks do not matter at IMPs. (They tend to be worth between 10 and 20% of the IMPs avalaible in an undoubled contract, rendering many safety plays as useless)

In my experience the average cost of giving declarer a trick at pairs, he was not entitled to, is far less than letting a contract make, which could have been defeated.
While the cost can be the same my best guess is the cost is on average about half and that is true even if the final contract is going to be duplicated across the field.

There are reasons for that.
One is that the total number of tricks available to declarer tends to fluctuate, depending on errors and guesses on both sides. In that sense the opening lead is only the first of several guesses.

While you should not go out of your way trying to defeat an undoubled standard contract at pairs, there is still a premium trying to do so. Also leading your longest suit often simplifies the defense in notrump, while passive leads, often technically superior, tend to complicate it. (Will partner switch to clubs if he comes in, for example?)
Whether this makes a low club the best lead at pairs I do not know, but it is certainly closer than double dummy simulation suggests.

Rainer Herrmann

[Zelandakh]

rogerclee, on 2012-January-02, 18:02, said:

1) The results usually favor unusual leads. Making unexpected leads is hard on partnership defense, particularly if the lead is unreadable (here could be from HT9x) and/or you don't play some kind of smith. Your double dummy computer is always shifting back to clubs when it's right. In real life, partner often just plays you to have led from HT9x since he has nothing better to play for.

Just to point out that some players still use Journalist Leads against NT contracts where the 9 would show the 10 and no higher honour, thus removing the HT9x possibility. Of course it is easy to see how this could benefit declarer more than partner.

I have to admit that when I saw the hand a heart lead looked fairly obvious. But I probably make too many of this lead type sitting over a 2NT opener. Certainly partners sometimes do not understand why leading 4th highest from longest and strongest was not chosen when it does not work. And when it does work they usually do not notice.