[straube]

So we're playing 1M=5+ and a good 9 to bad 15 hcps (roughly rule of 19)

For 1S, looking at

2H-3 trump or 4 with a less offensive hand, between 8.5 and 7 losers
2S-9 or more losers
2N-4+ trump, ranging from 7 to 8 losers
3m-weak jump shifts
3H-4+ trump, 8.5 losers
3S-4+ trump, 9 or more losers

We're using the Modern Losing Trick count which is why we count a half loser...or a quarter loser, etc.

So it looks like 1S-2N, 3C could just about ask whether opener has 7, 7.5, or 8 losers

I'm thinking that if responder has something like 6 or 6.5 losers, he should just bid game. With a better hand, he might relay.

Criticism?

[ArtK78]

Never heard of "Modern Losing Trick Count." I use "Modified Losing Trick Count" as presented by George Rosenkranz in his Romex books (he uses the acronym MLTC, as you may also, for a different reason). Opener counts losers, responder counts cover cards, not losers.

So I really can't evaluate your response structure.

[straube]

ArtK78, on 2012-January-27, 09:47, said:

Never heard of "Modern Losing Trick Count." I use "Modified Losing Trick Count" as presented by George Rosenkranz in his Romex books (he uses the acronym MLTC, as you may also, for a different reason). Opener counts losers, responder counts cover cards, not losers.

So I really can't evaluate your response structure.

Ron Klinger wrote a book on Modern Losing Trick Count. He's basically looking at 9-cd fits or 8-cd fits with good trump. He subtracts half a loser for every ace in a long suit and adds half a point for every queen in a long suit.

So Axxx Kxxx Qx Qxx

would evaluate as 1.5 + 2 + 2 + 2.5 losers for 8 losers. Then he would modify this by subtracting half a loser for the Qx. So 7 3/4 losers.

To count tricks available, you subtract losers in both hands from 24.

How does Rosenkranz count losers and cover cards?

[ArtK78]

straube, on 2012-January-27, 09:56, said:

Ron Klinger wrote a book on Modern Losing Trick Count. He's basically looking at 9-cd fits or 8-cd fits with good trump. He subtracts half a loser for every ace in a long suit and adds half a point for every queen in a long suit.

So Axxx Kxxx Qx Qxx

would evaluate as 1.5 + 2 + 2 + 2.5 losers for 8 losers. Then he would modify this by subtracting half a loser for the Qx. So 7 3/4 losers.

To count tricks available, you subtract losers in both hands from 24.

How does Rosenkranz count losers and cover cards?

Losers are counted in the same way.

Responder counts cover cards. Potential cover cards are Aces, Kings and, to a lesser extent, Queens (often counted as a 1/2 cover card unless the auction showed that the Q was a full cover card). Shortness with adequate trump support can also count as cover cards. The number of potential cover cards morphs into actual cover cards as the auction progresses. Responder could usually determine the proper level of the contract by determining how many of his potential cover cards were working cover cards (they actually covered declarer's losers) and would act accordingly. In some situations (such as major suit raises) responder announced to opener how many cover cards responder had, and opener would determine the proper level of the contract.

In the version of Romex that I learned many years ago, the major suit raise structure included constructive single raises - 2 to 2 1/2 cover cards; limit raises - 3 cover cards; and forcing raises - 3 1/2+ cover cards.

Minimum opening bids were 7 losers (8 losers was a subminimum opening, and the original Romex system only permitted sub-minimum openings in limited situations). A six-loser hand was an above-average opening bid, and a 5 loser hand was a maximum one-bid (and could qualify for a Dynamic 1NT opening bid if other requirements were met). Unbalanced hands with 4-5 losers were opened 1NT (the aforementioned Dynamic 1NT). 3 loser hands were opened with 2♣.

If opener showed a 5 loser hand and responder had 4 potential cover cards, the partnership was in the slam range. Responder's goal would be to determine if all of his cover cards were working. Other situations were dealt with similarly.

By the way, loser count is not nearly as effective when a balanced hand faces a balanced hand. HCP valuation works much better in those situations. Rosenkranz had some facinating methods for slam bidding when a balanced or semi-balanced hand faced another balanced or semi-balanced hand. One of those was referred to as CONFI, which is an acronym standing for CONtrols and FIt. If the partnership was in the 31-32 HCP range, the goal was to determine if at least 10 controls were present (A=2 and K=1) and if there was an 8 card fit. Baron was used in CONFI auctions to find an 8-card fit if there were adequate controls for slam purposes.

But this is from the mid- to late-1970s. I haven't played Romex seriously since about 1980. I understand that Rosenkranz updated his methods several times since then, and I have a couple of friends who continue to play his methods (and won a National pairs game using his methods).

[awm]

There was some previous thread about bidding 1M-4M in a precision context.

A lot of people (including me) believe that you need to be careful with this. Basically, you don't want to blast 4M if you will have an excellent slam opposite an ordinary 5431 hand with the "right" singleton. For example:

♠Kxxx ♥xxx ♦AKxx ♣Kx is seven losers, modified down to 6.5 for an ace in a long suit. Opposite: ♠AQxxx ♥x ♦QJx ♣Axxx (which is a fairly ordinary 5431 pattern) you are virtually cold for slam.

The point is that 1M-4M hands that have good values normally won't have xxx or Axxx or the like in a side suit. This implies that bidding 1M-4M with all hands in some loser range is a bad idea; you need to pay at least some attention to the honor structure.

[straube]

awm, on 2012-January-27, 10:14, said:

There was some previous thread about bidding 1M-4M in a precision context.

A lot of people (including me) believe that you need to be careful with this. Basically, you don't want to blast 4M if you will have an excellent slam opposite an ordinary 5431 hand with the "right" singleton. For example:

♠Kxxx ♥xxx ♦AKxx ♣Kx is seven losers, modified down to 6.5 for an ace in a long suit. Opposite: ♠AQxxx ♥x ♦QJx ♣Axxx (which is a fairly ordinary 5431 pattern) you are virtually cold for slam.

The point is that 1M-4M hands that have good values normally won't have xxx or Axxx or the like in a side suit. This implies that bidding 1M-4M with all hands in some loser range is a bad idea; you need to pay at least some attention to the honor structure.

Good point. Can you give us a couple of example hands where you would blast to 4M?

Also, what do you think of the structure I proposed? I like the general idea of it but am wondering if we should tweak it somewhat. For instance, 3H might be better to promise 8 losers and 3S 8.5 or more losers. That sort of thing.

[akhare]

awm, on 2012-January-27, 10:14, said:

There was some previous thread about bidding 1M-4M in a precision context.

A lot of people (including me) believe that you need to be careful with this. Basically, you don't want to blast 4M if you will have an excellent slam opposite an ordinary 5431 hand with the "right" singleton. For example:

♠Kxxx ♥xxx ♦AKxx ♣Kx is seven losers, modified down to 6.5 for an ace in a long suit. Opposite: ♠AQxxx ♥x ♦QJx ♣Axxx (which is a fairly ordinary 5431 pattern) you are virtually cold for slam.

That's an interesting point. Perhaps, considering QPs in addition to loser count might provide the answer on whether it's advisable to blast to 4M?

For the uninitiated, QPs is A=3, K=2, Q=1...

[phil_20686]

LTC.

LOL.

[blackshoe]

IIRC, Confit was still in the Romex system described in Bridge: The Bidder's Game (1985), but had disappeared by 1990's Bid to Win, Play for Pleasure. Klinger mentions "cover cards" (and credits Rosenkranz for the concept) in The Modern Losing Trick Count. As for the word "modern" in that title, I suspect it was put in because until that book came out in 1986, no one had written on the subject since Maurice Harrison-Gray's articles in the '50s and '60s. The Modern Losing Trick Count has been reprinted several times, most recently, I think, in 2001.

In Bid to Win, Play for Pleasure and later books, Rosenkranz makes a point of specifying control requirements for balanced hands. A 19-20 balanced hand should have 6 controls, 21-22 seven, and so on. He talks about upgrading and downgrading based on the number of controls. I don't recall if those requirements were in Bridge: The Bidder's Game.

Romex still specifies raises in terms of cover cards. There's a scene in, I think, Godfrey's Bridge Challenge where Godfrey is reciting the trump length, loser and cover card requirements for a simple raise of 1M, and his wife says "you sound like a robot". These latest three books, with Godfrey teaching his wife and her regular partner Romex, are an attempt to ease the transition from 2/1, and to showcase the current (as of writing at least) "two card" system — Romex when vulnerable, Romex Forcing Club (a Precision variant) when not vulnerable (at match points — at IMPs RFC only at favorable). The idea is that if you already play 2/1, you can keep most of what you play now, just adding the cornerstone Dynamic NT and Mexican 2♦ systems and adopting Romex hand evaluation methods. From one of the books "We don't play Bergen Raises any more?" "Yes, you do, but George has his own flavor of these bids."

You can get all four of the latest Romex books — Bid to Win, Play for Pleasure, Godfrey's Bridge Challenge, Godfrey's Stairway to the Stars, and Godfrey's Angels, from Baron Barclay for ten bucks. The biggest problem with taking on the full two card system, aside from memory effort and finding a willing partner, is that three of the opening bids (2♥ when playing RFC, 2NT, and 4NT)* are Mid-Chart, and finding a game or tournament that allows the Mid-Chart, at least around here, ain't easy.

*2♥ is the Precision 2♦ opening, 2NT shows a "bad" three level preempt in either minor (so that responder to a 3m opening is better placed to bid 3NT), 4NT shows a "good" five level minor preempt, the idea being to facilitate slam bidding. BTW, the local expert who taught me the 4NT opening called it "the Reese 4NT" so I guess it goes back a ways.

[straube]

Still looking for feedback on our 2N, 3H, and 3S responses.

My basic question is that considering we're opening hands like AQxxx Qxx xx QJx (which is about 7 1/2 losers)and that we have 1S-2S as a bad raise (like xxx xx xxxx xxxx might qualify at favorable), then should we really be preempting 9-loser hands at the 3-level?

We could do...

2N-7 or 7 1/2 loser hand
3H-8-loser hand
3S-8 1/2 loser hand

and then slot all the 9+ loser hands into 2S.

Granted, it feels like with Kxxx x xxxx xxxx that one wants to jump to 3S, but most pairs don't have the option of a direct bad raise. Maybe that's better.

[awm]

Try Axxx Kxx QJxx Kx.

The best I can do with 5431 type shape is KQxxx AQx x Axxx and I still need some suit breaks (to ruff two clubs) despite the dead max. Basically I want a min GF with "wasted values" opposite any singleton.

[straube]

Good example. How do you feel about slotting all of the 9-loser hands into 2S?

[awm]

I do not think LTC is the right way to think about preemptive raises. LTC is a method for determining how many tricks you can make, while preempts are more about par. Some adjusted LoTT will serve you better; if you are going to play preemptive raises you definitely need to bid one on Kxxx x xxxx xxxx.

[dake50]

If you choose LTC as your tricks evaluator, do you just
accept C:Jxx w C:xxx(dummy) is a rare 3-losers that
doesn't combine well for you? Or do you have a "3+losers
here", do we combine for 3+losers? Do you find this 3+3 flaw?
Next, LTC for tricks, what finds trump tops, controls,
right K,Q when you are slammy? The follow-ups make this
playable or efficient. Publish them for critique.

[straube]

My impression is that different hand evaluation structures start to resemble each other as more adjustments are made. Also, it seems that the difference of one loser equates to slightly less than 3 hcps. So I'm suggesting Modern Losing Trick Count as a language for expressing the approximate trick-taking potential of a hand.

As far as whether hands mesh well...there's a lot of variation in that and it's hard to assess for getting to game. We have more room to assess mesh when looking for slam.

[awm]

straube, on 2012-January-28, 11:17, said:

My impression is that different hand evaluation structures start to resemble each other as more adjustments are made. Also, it seems that the difference of one loser equates to slightly less than 3 hcps. So I'm suggesting Modern Losing Trick Count as a language for expressing the approximate trick-taking potential of a hand.

As far as whether hands mesh well...there's a lot of variation in that and it's hard to assess for getting to game. We have more room to assess mesh when looking for slam.

There are basically three different styles of hand evaluation. These are: (1) Determining how many tricks we can take in a suit. (2) Determining how many tricks we can take in notrump. (3) Determining the total number of tricks available in total to the two sides in their best fits. It's quite possible for these three computations to have very different answers, and different evaluation techniques apply; the answers to these three measures will not resemble each other regardless of the adjustments you apply, because the "right answer" is different! As a very rough sketch, LTC is fairly good for the first, HCP is fairly good for the second, and LoTT is fairly good for the third. All three of these measurements are somewhat inaccurate, and can be improved in a variety of ways (by more accurate consideration of the value of certain cards, card combinations, and shape.. and by knowledge of how the two hands mesh).

The situations where you want to use these measurements are obviously different. They are:

(1) When we are playing in a suit and it is either an uncontested auction, or we have a substantial majority of the strength (such that a minus score is quite unlikely to be a good result).
(2) When we are playing in notrump.
(3) When we are in a contested auction, and it is at least conceivable that a minus score could be good for our side (relative to the field, or par, or whatever on this hand).

When you are making a preemptive raise, especially in a limited-opening system, you need to consider that you are in situation (3). Holding three points with a fit (i.e. Kxxx x xxxx xxxx) opposite partner's limited 1♠ opening, it is not unlikely that a score like -100 is actually quite good for our side, and our goal in bidding 3♠ is not necessarily to make the contract (although certainly that might happen) but rather to put pressure on the opponents and attempt to buy the contract (quite possibly for a small minus score) when they could in fact make something more substantial.

[straube]

I was only talking about evaluation methods that assess trick-taking potential in a suit. This thread is about major suit raises.

[awm]

straube, on 2012-January-28, 11:55, said:

I was only talking about evaluation methods that assess trick-taking potential in a suit. This thread is about major suit raises.

Consider two hands like this:

(1) ♠xxxx ♥x ♦xxxx ♣xxxx
(2) ♠AQxx ♥x ♦xxxx ♣xxxx

Obviously the second hand is much better than the first. LTC will count it as about two tricks better.

Yet it makes sense to bid 3♠ on both hands opposite a limited 1♠ opening. LTC is right about the strength; my expectation is that I will often be down two in 3♠ on hand 1, whereas I expect to make 3♠ fairly often on hand 2, and be down only one most of the remaining time. But when I bid a preemptive 3♠, I'm not necessarily bidding to make. Based on my shape and fit, I expect that there will be around 18 total tricks in the deck. On the first hand, this means the opponents can almost surely make a game in hearts (or maybe 3NT, or 5m) and bidding 3♠ right away makes things very difficult for them. The fact that I would also bid 3♠ on hand 2 makes it harder for them to double! On the second hand, it's fairly likely that we make 3♠, and if we're down they might make four of something. By bidding 3♠ right off I prevent them from finding their contract and I might score up +140.

The point is that if you restrict your 3♠ bids to include only hands with 8 losers (i.e. hands where you have a decent chance to make the contract) you are missing a lot of opportunities and arguably missing the entire point of a preemptive raise.

[straube]

Wait. I'm not arguing with LOTT. I asked your opinion as to whether we should shove the preemptive bids into 2S. I wasn't thinking that a preemptive 3S bid was not valuable.

I was thinking that...

1) a lot of folks don't even have a 3S preemptive bid. Some of them even pass with Kxxx x xxxx xxxx ! Just awful.
Others make a forcing NT and then try to play 2S. Shows opponents zero respect.

2) we have a direct raise of 2S which can be crap and is at least a little preemptive. Though raising to only 2S is not what I want to do with Kxxx x xxxx xxxx, I was considering the frequency of this sort of hand vs the frequency of a hand with four spades, high ODR, and about 8.5 LTC. If the latter were much higher frequency, then it might pay to yield 3S to that bid.

So I think you're right (about keeping 3S weak). Plus I looked at some hands and feel like I really would like a 3S weak raise and that the 8.5 LTC hand isn't frequent enough to assign 3S to that.

[straube]

btw, my earlier post today was in response to dake50 and not your prior post.