[cherdanno]

inquiry, on Jul 27 2010, 06:00 PM, said:

This also is consistent with my mindset to award the highest scoring contract EW can play (or defend) at least a ten.

I think this is a mistake. It means that missing a 60% game becomes exactly as costly as missing an 90% game, which is very much unlike matchpoint bidding (or any kind of bridge bidding).
CTC have thought about this a lot and there scoring is based strictly on matchpoint expectancy.

[inquiry]

TylerE, on Jul 27 2010, 06:02 PM, said:

Just to make sure I understand...the 4♣ score would also apply for any lower ♣ partscore?

If not otherwise stated, the answer is yes.. 4C scores the same here as 1C, 2C, and 3C....

[TimG]

awm, on Jul 27 2010, 07:45 PM, said:

So we have 18 pairs in a ♣ partial, 8 in a ♣ game, and 1 in 3NT. Supposing that 3NT won't make (and is usually down two) we have:

If 5♣ makes, then 5♣ scores 22.5 and ♣ partial scores 9.5.

If 5♣ fails, then 5♣ scores 4.5 and ♣ partial scores 17.5.

Notice that if 5♣ made exactly half the time, the two scores would be equal in expectation. If anything, the effect of 3NT possibly failing by only one trick only hurts the 5♣ bidders (since ♣ partial always beats 3NT, but 5♣ will occasionally push with 3NT when both are one off).

Using Ben's numbers of 5♣ making 65% of the time the expected scores are:

For 5♣ 16.2 and for ♣ partial 12.3. If we normalize such that 5♣ is worth 10, we get a score of 7.6 for the ♣ partial. This is a lot higher than the score it received in the scoring.

I guess the point is, missing a 65% game at MPs is not a horrible result in expectation.

But, then everybody would be above average!

I see your point and understand the math, it's not really that the club partial deserves a better score, but that the club game doesn't really deserve a 10.

I agree with you: once you decide the top spot must get a 10, then you should adjust the other scores so that they are right on a relative basis.

[Siegmund]

Quote

I guess the point is, missing a 65% game at MPs is not a horrible
result in expectation.

This comment applies not just to this board, but to several others in the set. We were told that the scoring would reflect matchpoint expectations; but it seems, uniformly, to be a "find the best spot" contest instead.

There are quite a number of contracts in this set of 16 that appear to depend on finding a key card or something similar - but I don't remember ANY where the scores for a partscore and a game in the same suit were nearly tied.

Seconding the remark that, depending on the exact probabilities, 8 for 5C and 5 for 3C is more like what I expect (or 10-7 if we insist on normalizing them up, which we shouldnt).

[cherdanno]

Btw, I still don't understand why 5♣ is supposed to be a 60% game. When I tried simulations, I got 38%.

[inquiry]

cherdanno, on Jul 27 2010, 08:52 PM, said:

Btw, I still don't understand why 5♣ is supposed to be a 60% game. When I tried simulations, I got 38%.

If this is 38%, then the scores are all wrong and will need to be recalculated.

How did you come up with 38%?

Remember the following:
North has an opening hand
South has a responding hand
North does not have three or four card heart support (no mention of support dbl)
North has at least five diamonds
Sourth has, obviously, at least six hearts and I think no more than 7
North has at 11 hcp, but not all the missing hcp as south has a response

Did you apply anything like the above to the constraints? The more hcp you allow north to hold, the better the chances of making 5♣ (increases chances of spade KQ in north). There are also some squeeze chances (if spades are split).

When I apply those constaints, limiting north to 11-15 hcp, I find five clubs making 65% of the time.

[Mbodell]

inquiry, on Jul 27 2010, 08:11 PM, said:

North does not have three or four card heart support (no mention of support dbl)
Sourth has, obviously, at least six hearts and I think no more than 7

I question these assumptions. Is it really the case that North can not have 3 hearts? Not everyone plays support doubles! And if North has a good 6 card diamond suit and 3 hearts might well bid 2♦ anyways.

[awm]

Another thing that seems to come up is that often declarer needs to establish a second spade trick for a diamond discard. This often winds up with a series of spade plays where declarer leads the spade ten to an honor and the ace, then has to decide whether to lead up to the jack (playing opener for the king and queen) or try to ruff a spade (hoping RHO started with honor-third). Double dummy play will always get this right (for example always establishing the second spade trick when opener has four spades and responder has three, which is a fairly common layout) but at the table it is not so easy.

Of course, the double-dummy assumption can help the defense too, but the main issue there seems to be the opening lead, and I think a trump lead when you're defending a game-level contract bid on very thin values is a fairly common approach.

[655321]

Mbodell, on Jul 28 2010, 12:45 AM, said:

I question these assumptions.  Is it really the case that North can not have 3 hearts?  Not everyone plays support doubles!  And if North has a good 6 card diamond suit and 3 hearts might well bid 2♦ anyways.

The assumptions seem fair enough to me when the bidding has stayed low.

But if the auction was something like (1♦) X (1♥) 5♣ we would know nothing about North's shape.

[Hanoi5]

awm, on Jul 28 2010, 02:21 AM, said:

Another thing that seems to come up is that often declarer needs to establish a second spade trick for a diamond discard. This often winds up with a series of spade plays where declarer leads the spade ten to an honor and the ace, then has to decide whether to lead up to the jack (playing opener for the king and queen) or try to ruff a spade (hoping RHO started with honor-third). Double dummy play will always get this right (for example always establishing the second spade trick when opener has four spades and responder has three, which is a fairly common layout) but at the table it is not so easy.

Of course, the double-dummy assumption can help the defense too, but the main issue there seems to be the opening lead, and I think a trump lead when you're defending a game-level contract bid on very thin values is a fairly common approach.

Don't people avoid leading a trump with singleton?

[cherdanno]

Yes sorry I had a typo in my simulations, I now get similar numbers as Ben. They seem to be based mostly on the fact that declarer is picking up both KQx(x) and Hxxx with North, just as Adam wrote.

[TimG]

cherdanno, on Jul 28 2010, 06:24 AM, said:

Yes sorry I had a typo in my simulations, I now get similar numbers as Ben. They seem to be based mostly on the fact that declarer is picking up both KQx(x) and Hxxx with North, just as Adam wrote.

Can you see which is less frequent and subtract that likelihood from the making percentage?

[Fluffy]

when north fails to lead ♠K you have a clue, althou well you also don't know if they overruff your diamonds or not so probably evens out.

[cherdanno]

TimG, on Jul 28 2010, 06:44 AM, said:

cherdanno, on Jul 28 2010, 06:24 AM, said:

Yes sorry I had a typo in my simulations, I now get similar numbers as Ben. They seem to be based mostly on the fact that declarer is picking up both KQx(x) and Hxxx with North, just as Adam wrote.

Can you see which is less frequent and subtract that likelihood from the making percentage?

Constraints:
- North has at least 11 hcp, South at least 5.
- North has at least as many diamonds as clubs, and (no 5-card major) or (more diamonds than cards in each major)
- South has at least 4 hearts
- If South has 5 spades, he has more hearts than spades

West makes 5♣ 59.3% of the time (10000 runs)
North has ♠KQ: 30.1% of the time
North has ♠Hxxx: 27.8% of the time

[TimG]

cherdanno, on Jul 28 2010, 09:51 AM, said:

TimG, on Jul 28 2010, 06:44 AM, said:

cherdanno, on Jul 28 2010, 06:24 AM, said:

Yes sorry I had a typo in my simulations, I now get similar numbers as Ben. They seem to be based mostly on the fact that declarer is picking up both KQx(x) and Hxxx with North, just as Adam wrote.

Can you see which is less frequent and subtract that likelihood from the making percentage?

Constraints:
- North has at least 11 hcp, South at least 5.
- North has at least as many diamonds as clubs, and (no 5-card major) or (more diamonds than cards in each major)
- South has at least 4 hearts
- If South has 5 spades, he has more hearts than spades

West makes 5♣ 59.3% of the time (10000 runs)
North has ♠KQ: 30.1% of the time
North has ♠Hxxx: 27.8% of the time

I realize there are some other considerations -- 5C making isn't always due to the spade position -- but it does seem like the correct percentage is much closer to half the 60%.

[awm]

Hanoi5, on Jul 28 2010, 06:01 AM, said:

Don't people avoid leading a trump with singleton?

This doesn't actually matter much.

The thing is, LHO is known to have the long diamonds and RHO the short diamonds. Any time RHO has ♣Jx, you basically have to pull trump anyway because otherwise RHO will overruff one of the diamonds relatively early in the hand. Of course, LHO is also known to have the short hearts, and if my plan in the play is to cross-ruff (ruffing three diamonds in dummy before pulling a second round of trumps) then I need to ruff hearts in order to re-enter my hand. If LHO has ♣Jx, he is well-positioned to overruff one of my heart re-entries.

So the only position where I'm likely to make by ruffing three diamonds is where the club jack is singleton and the opponents don't find a trump lead. If LHO has ♣xx he may well lead a trump (it does seem like a normal lead when opponents bid to 5♣ on very few points) and leading from ♣J stiff is probably more common than leading from ♣Jx or ♣x.

Looking at Cherdano's results, it does seem like the chances of 5♣ making are very close to the chances of being able to pick up a second spade trick.

[TimG]

Perhaps a better way to judge how much the spade position plays in the success of the contract is to change the spade position so that two spade tricks are not possible and see how good 5C is then.

[TimG]

cherdanno, on Jul 28 2010, 09:51 AM, said:

Constraints:
- North has at least 11 hcp, South at least 5.
- North has at least as many diamonds as clubs, and (no 5-card major) or (more diamonds than cards in each major)
- South has at least 4 hearts
- If South has 5 spades, he has more hearts than spades

West makes 5♣ 59.3% of the time (10000 runs)
North has ♠KQ: 30.1% of the time
North has ♠Hxxx: 27.8% of the time

I've tried to duplicate these condition. My first run of 1000 deals produced a 5♣ making percentage of 56.6%, so I suspect I've got something slightly off, but pretty close.

I then replaced East's ♠J with the ♠6. Under this condition 5♣ made 9.3% of the time. I think it is safe to say that the double dummy results greatly overstate the chances of 5♣ making as a result of being successful against both ♠KQ and ♠Hxxx in north.

If you allow declarer to make 9.3% of the time plus 30.1% of the remaining 90.7% (cherdanno's percentage for KQ in north), you get 5♣ making 36.6%.

Under both conditions, 4♣ failed less than 0.5% of the time, so it doesn't seem to me that there should be any difference between the scores for 3♣ and 4♣.

[gnasher]

TimG, on Jul 31 2010, 06:41 PM, said:

If you allow declarer to make 9.3% of the time plus 30.1% of the remaining 90.7% (cherdanno's percentage for KQ in north), you get 5♣ making 36.6%.

Don't we get anything for a non-trump lead, or a bit of card-reading?

Also, Cherdano's conditions included "South has at least 4 hearts". but the opposition bidding implies that South has six hearts. That must significantly increase the chance of both spades being onside.

[TimG]

gnasher, on Jul 31 2010, 01:17 PM, said:

TimG, on Jul 31 2010, 06:41 PM, said:

If you allow declarer to make 9.3% of the time plus 30.1% of the remaining 90.7% (cherdanno's percentage for KQ in north), you get 5♣ making 36.6%.

Don't we get anything for a non-trump lead, or a bit of card-reading?

Perhaps you should, I was just reporting what I found through simulation. Did not mean to imply anything beyond that.