[Cascade]

ArtK78 said:

I find the idea of upgrading a hand for no trump purposes due to extra controls to be backwards.

Controls are an indication that a hand should be played in a suit contract, not no trump. A hand with more high cards but less controls typically plays better in no trump than in a suit contract.

This is from another thread where I suggested a hand was only average based on controls for a 17 count.

I have done a simulation in which a balanced (4333, 4432, 5332) 17 count played in NTs opposite random hands constrained to have no 8-card or longer major fit and responder did not have a singleton or void. The double dummy tricks by the number of controls in the balanced 17 count are produced below:

Cont Tricks
3     7.895	
4     8.077	
5     8.29	
6     8.43	
7     8.523	
8     8.858

As you can see the average number of tricks increases with the number of controls in opener's hand.

[kenrexford]

I think it was Kilnger who came up with an adjustment up analysis that I like. Take total HCP (e.g., 17). Take the number of controls (e.g., 6). Multiple the control count by 3.3333 (e.g., 20). Subtract the HCP from the 3.33(control) count (20-17=3). If the number is 2-5, add a point. If 6-9, add 2. If 10+, add 3. If negative by the same degree, subtract.

Of course, this is just a rough guide and equires judgment and all. But, it illustrates the point well.

[jtfanclub]

What you did did nothing to agree or disagree with what he said.

Do a simulation in which a 17 count 5-3-3-2 played in NT opposite random hands constrained to have an 8 card spade fit and no 8 card heart fit.

Then do a simulation in which a 17 count 5-3-3-2 played in spades opposite random hands constrained to have an 8 card spade fit and no 8 card heart fit.

If the post you're quoting is correct, then you should do better in NT with fewer controls, and do better in the suit with more controls.

The idea that more controls produces fewer tricks in NT is a strawman. That's not what AK78 said, and not what anybody has said as far as I know. What has been said is that controls are worth more in a suit contract than in a NT contract, given the same numer of HCP.

[blackshoe]

In Romex (which doesn't have a natural 1NT opening), strong balanced hands have both a HCP range and a control requirement. The ladder is:

19-20 HCP, 6 controls
21-22 HCP, 7 controls
23-24 HCP, 8 controls
25-26 HCP, 9 controls
27-28 HCP, 10 controls
29-30 HCP, 11 controls.

The control requirement is considered more important than the HCP requirement. A hand might be adjusted upwards or downwards based on the number of controls, but not if it's more than 1 point "out". So a 23 point, 7 control hand would be treated as if it were 21-22 HCP, but a 24 point 7 control hand would still be treated as a 23-24 HCP hand. Similarly for excess controls. Rosenkranz says this is based on computer simulations and practical experience, but I don't remember the details.

In Romex, hands with 17-18 HCP are shown by opening 1 of a suit (usually a minor) and then jumping to 2NT. Hands with 12-16 HCP are shown by opening at the one level and rebidding 1NT. There are no control requirements for these hands.

[Cascade]

jtfanclub, on Mar 30 2008, 03:49 AM, said:

What you did did nothing to agree or disagree with what he said.

Do a simulation in which a 17 count 5-3-3-2 played in NT opposite random hands constrained to have an 8 card spade fit and no 8 card heart fit.

Then do a simulation in which a 17 count 5-3-3-2 played in spades opposite random hands constrained to have an 8 card spade fit and no 8 card heart fit.

If the post you're quoting is correct, then you should do better in NT with fewer controls, and do better in the suit with more controls.

The idea that more controls produces fewer tricks in NT is a strawman. That's not what AK78 said, and not what anybody has said as far as I know. What has been said is that controls are worth more in a suit contract than in a NT contract, given the same numer of HCP.

Read the first sentence of what he said:

ArtK78 said:

I find the idea of upgrading a hand for no trump purposes due to extra controls to be backwards.

What I have shown is that more controls for admittedly only one hcp value produces more tricks on average in no trumps other things being equal.

This suggests to me that "upgrading a hand for no trump purposes due to extra controls" is valid not "backwards".

Perhaps I shouldn't have quoted the next paragraph. I certainly tend to play a suit contract with a choice of contracts when I have lots of controls and tend to play no trumps with more queens and jacks. There are certainly exceptions e.g. when you have a running suit and side aces

One of my partner's puts this concept well. He says something like "bridge is a game about third round controls - if your third round controls are queens then you try to play no trumps if they are doubletons then you try to play suits".

[kenrexford]

Just as a side note -- does opening 1NT force you to play in a notrump contract? Or, does it show shape/strength and start the auction off?

[matmat]

pardon me, I am slow.

what exactly are we upgrading here? a 17 count with controls out of a 1n opener into a 1blah-response;2n rebid?

also, it would be interesting to see how these same hands do in their best major suit fit as a function of control count (8 card or longer, combined) and whether the expected number of tricks there rises quicker than in the NT contract.

[helene_t]

I am surprised by Wayne's results. I analyzed the GIB database using a logistic regression model
3NTmakes ~ A + K + Q + J + T
(the covariates referring to the number of aces, kings etc in the combined hands)
under constraints similar to those of Wayne, and I found that the optimal coefficients were almost identical to the MW scale (an ace being 4.2 rather than 4) and that switching to the optimal scale only improves your results by 0.03 IMPs per board.

Maybe this is not conflicting with Wayne's results. My intuition in that repect could be wrong (after all I didn't compute the average number of tricks so the results are not directly comparable) and there are some minor differences also (I did not require opener to have a 15-17 balanced, any deal without a major suit fit would do).

[skjaeran]

This is a complicated matter. But the general idea that controls are best for suit contracts and quacks are great for NT is not true. But you've got to see this in context wiht the whole deal.

If you've got a solid 6-card suit, it's better to have 3 aces in the side suits in stead of a combination of KQJ KQJ QJx. If opps lead your QJx suit, they have established their setting trick(s) and you're down as soon as you try establishing one of the other side suits.

Quacks are great as stoppers in NT, but normally suck when it comes to being a source of tricks (unless they're supported by ace and king of course). This is because they're so slowly established as winners - it will often be too late when opps already have established and cashed their suit.

In suit contracts you can more easily establish such suits, having the trump suit as a guard against opps cashing a long suit. Then again, having a source of tricks in a suit lacking both ace and king you might suffer early ruff(s).

I'd say controls are great in any contract.
Quacks are a more difficult matter, but in general I believe they pull a little more weight in suit contracts than at NT. But the margin seems to be small.

[ArtK78]

Just a comment about the scheme of hand evaluation advocated in the Romex system book, and not about the topic of controls versus quacks in notrump and suit play.

The purpose of bidding is to determine the trick taking potential of the bidding side's hands. For slam bidding purposes, especially with a balanced or semi-balanced hand facing another balanced or semi-balanced hand, the number of controls held by the partnership is of paramount importance. When it comes to the discussion of controls, most of the Romex system book is geared towards slam bidding. The discussion of game bidding involves modified loser count and cover cards, and controls don't usually even enter into the discussion.

So, it makes sense to propound control requirements for opening bids, especially opening bids representing strong hands, as slam bidding considerations usually outweigh game bidding considerations in auctions which start with strong opening bids.