Join Bridge Winners
All comments by Mike Dorsel
1 2
You are ignoring the author of this comment. Click to temporarily show the comment.
The analogy of a drug dealer giving a “free taste” comes to mind. The ACBL is a business, one the needs to attract/hook new members, and keep them as long as possible. Concentrating on the elite is not the way to make that happen. Frequent intermittent reinforcement (much like programmed slot machine payouts) is a superior strategy.

While the bigger games are getting watered-down, it would seem that there is an opportunity for a few side games with more flexible conditions of contest, ones that would draw tougher fields. The trade-off would likely be for these games to pay relatively little in terms of MPs, since the more popular games would not be subordinate, but competing against talent has its own rewards.
Feb. 28, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Thanks again Eric. You and Jeff obviously know each other very well. There must be times when you “feel” something about a hand, having nothing to do with agreements, especially when playing without screens.

How often are such “hunches” right, and how do you deal with these situations? Are you/Jeff consciously moderating your tempo and demeanor? How would you advise less experienced players to think about these things, given all the other stuff going on at the table?
Feb. 28, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Thanks Eric. Your response regarding shapely hands and LTC makes perfect sense.

How about for balanced hands? Does a more structured approach to valuation help when deciding whether to treat a hand as 11-13, 14-16, 17-19, etc? How about when deciding whether to treat a hand as a minimum response, an invite or a GF?
Feb. 28, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Thanks Eric, for participating in The Well. Love the book and am chewing through it slowly.

Given your development of a highly structured system … When it comes to hand valuation (pre-competition), do you take a more intuitive or quantitative approach? Is it more art or science? Do you and Jeff consciously value hands similarly or differently?

Feb. 27, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
It might seem strange, but it is possible to point out that the conditions-of-contest did not involve any ACBL-like restrictions without chiding the asker for a potentially limited perspective.

Feb. 25, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Most cards held by declarer in the suit led at a NT partial … ?

Declarer holds something like: Kx Ax KT987643 A … ?
Feb. 25, 2014
Mike Dorsel edited this comment Feb. 25, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
I have performed an additional regression to those referenced above. This yields an R^2 of 96.8% between (CCC-HCP) and (KnR-HCP), across all variations in AKQJT9 content in suit lengths from 1 to 5 cards.

The intercept was materially zero (<0.02), and the coefficient of the regression was 0.92 (CCC-HCP being less sensitive than KnR-HCP).

CCC for this regression included the SQA (for 3+ card suits).

KnR was derived by removing shortness value for singletons but otherwise using data obtained from the K&R Hand Evaluator (http://www.jeff-goldsmith.org/cgi-bin/knr.cgi).
Feb. 25, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Another key point (that I should have made earlier) is that after using this approach for a while, one's reflexes become better. Controls look better than quacks, pushers take on more meaning, and concentrations look better than scattered strength. After counting my HCP and looking at the overall texture, I like to guess at the CCC value; my guesses have gotten better over time.
Feb. 25, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Thanks Mike. You bring up some good points.

The suit quality adjustment (SQA) is clearly the most complex part of the calc. Luckily, when one has a 4432/4333/5332 hand, the SQAs are frequently immaterial to decision-making. These balanced hands are also the ones where you/partner are tending to limit your hand within a range, which is where refined valuation is most valuable.

For semi-balanced and unbalanced hands, one's first call is usually clearer, so one is more likely to have an extra round to perform a SQA, and by then the noise around the table is influencing hand value as well.

Your example hands are a couple of cards light, but you rightfully highlight that such hands have similar prospects (at least until the auction tells you otherwise). It should also be noted that modified/new LTC (a valuation approach that can be used once one has found a fit), treats these sample hands similarly.
Feb. 25, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
A 19.8 simplified KnR makes it a “bad” 20 in my book, so yes to the upgrade.
Feb. 23, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Thanks for asking Gary. I debated raising this issue in the original post.

I have not explicitly run simulations, but I have run regression analyses on the potential variations in high card content for 3-, 4- & 5-card suits. While these variations are not equally likely, this should still be largely indicative of any sim results, since both methods can be algebraically broken down into a combination of single suit cases without interaction effects. (I actually determined the Suit Quality Adjustment to nicely fill in the gap between the two approaches (i.e. the residuals from the three regressions).

I have also compared the calcs manually for a number of hands, and those results very much point me in the same direction with respect to table actions. That, of course, is anecdotal and subjective.

Bottom line: Not being a “sim guy” myself, I would very much welcome/appreciate it if anyone wanted to take such a simulation comparison on as a mini project. :-)

Feb. 23, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Both materially to blame. North tried twice to lose the board, and South put the final nail in the coffin.

Against North: Without discussion of 3D* rebid, 3S still better choice than 2H. 4S on third round creates wrong picture of hand and handcuffs auction.

Against South: 4NT RKCB over 4S is clear, even with some values that look wasted.
Feb. 20, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
The point of Responsive Doubles is generally to show equal length in the unbid opposite-rank suits, so the the takeout Xer can choose the best strain (since he/she may well have differing lengths in those suits).

In 1m-X-(2m or 3m) and like auctions, Responsive Xer is typically showing 4-4 majors and asking partner to choose.

In 1M-X-2M and like auctions, Responsive Xer is showing both minors and asking partner to choose.

Over 1M-X-2M, holding 4 other major and 5 clubs, advancer needs to choose. Otherwise, a fairly high percentage of the time, partner will bid diams when it is the wrong strain.

Using a X over 1M-X-2M to show 4 other major and 5 of a minor is probably workable (with enough additional discussion of overcaller's rebids), but that treatment/convention would not be a Responsive Double.
Feb. 18, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Interesting concept Phil. Making my head spin a bit …

Over 2NT-3; 3 you use 4M for slammish 4-4 hands. I suggest using 4, right-siding the major-suit contracts and leaving the other three 4-level bids for minor one-suiters.

I'd only want to use the non-fit transfer breaks (which others have mentioned as well) in GF auctions.
Feb. 15, 2014
Mike Dorsel edited this comment Feb. 15, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Thanks. I could have been clearer.
Feb. 14, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Thanks Benoit. Your change to the 2NT-3; 3 section and a 3 transfer f/b 3 could definitely add value.

You could switch the follow-ups to your 3 response rebid, but that would trade-off lead-direction availability for right-siding.

And I agree that the slam move over 3NT = 4-4 majors could be better.

It all comes down to how one balances the aforementioned criteria, including your rightfully added lead-direction ability.
Feb. 14, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Kai-Ching, I'm not seeing what in the 3 section is wrong-sided. (When I say 3=5 majors, I mean 3 spades & 5 hearts.)

In fact, majors are only wrong-sided in one case (2NT-3; 3-3 leading to a spade contract).

The one other “miss” is that 5=3 majors (5 spades & 3 hearts) is stuck starting with 2NT-3, so the 5-3 spade fit (with declarer having 5) could be missed. On a conditional probability basis, that seems like a small give-up relative to the other choices available …
Feb. 14, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Hi Kai-Ching. Loved your work on Denial Stayman and the associated GF-relay structure (which then pard & I called Lin Transfers) …

There are a number of “solutions” to finding all 8-card fits over a 2NT opening (or rebid after say 2-2). Each approach, consistent with the responses so far, either falls short of the goal, wrong-sides a number of contracts, has a non-intutive & memory-intensive feel, or some combination of these draw-backs.

The structure that comes closest to satisfying my implied criteria above is a modified version of something from Romex … over 2NT-3:

3 = <4 hearts, <5 spades
– 3 asks (3 shows 3 spades, 3NT shows 2, 4m & 4 show 4)
– 3 shows 3=5 majors (3NT shows 2 hearts, 4 & 4 show 3) ???
– 3NT = to play, <5 hearts, <4 spades
– 4 & 4 = single-suited minor (probably a 6331 type)

3 = 4 or 5 hearts, <4 spades
– 3 = asks (3NT shows 4, 4 & 4 show 5)
– 3NT = to play, <3 hearts, <5 spades
– 4 = single-suited clubs (probably a 6331 type) ???
– 4 = transfer to hearts, slammish
– 4 = to play, 4+ hearts
– 4 = single-suited diams (probably a 6331 type) ???

3 = 5 spades, <4 hearts
– 3NT = <3 spades
– 4 & 4 = single-suited minor (probably a 6331 type)
– 4 = transfer to spades, slammish
– 4 = to play, 3+ spades

3NT = specifically 4-4 majors
– 4 & 4 = Extended South African Texas transfers to hearts & spades ???
– 4 = single-suited clubs (probably a 6331 type) ???
– 4 = single-suited diams (probably a 6331 type) ???

All of this assumes that:
* 3 & 3 are Jacoby Transfers (3 followed by 3 showing 4=5 or 4=6 majors)
* 3 is minor-suit oriented
* 3NT is either natural or minor-suit oriented

There are a number of ambiguities above (at the very least the ??? items), which I hope to iron out via a forthcoming poll (or series of polls).
Feb. 14, 2014
Mike Dorsel edited this comment Feb. 14, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Wouldn't a Responsive X by East show shape much more in line with what West had?
Feb. 14, 2014
You are ignoring the author of this comment. Click to temporarily show the comment.
Either 1NT or X is reasonable on the first round; it's a guess. I would be interested to hear what folks who have lots of experience find works better with this hand type over time.

East has to get spades in at the 2 level while possible.

Playing BoP Dbls (as written up in Mel C's book), the second X is BoP, showing the extras and 3-2-y-z. Odds are, looking at the East hand, that West is 3-2-4-4 or maybe 3-2-5-3 (with concentrations of values in the blacks).

All of this makes it right for East to pass the second double at MPs.
Feb. 14, 2014
1 2
.

Bottom Home Top