Join Bridge Winners
All comments by Doug Bennion
You are ignoring the author of this comment. Click to temporarily show the comment.
Here are large-sample comparisons human vs double dummy results http://www.rpbridge.net/9x29.htm
Dec. 30, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
I play a Stayman scramble with hands this shape and weaker, and if opener responds 2, will pull up in a Moysian
(opener will not be 22(45)). The Moysian holds its own vs 1NT, and the found 8+ fits are huge winners.
Dec. 30, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
My hand generator deals QJxx Jx KQx AQxx about once every 7 million deals ('x' could be ‘T’).

Making the arithmetic easy, one could play 7000 deals a year, so one could see a close-matching pair of hands every 1000 years, and 2017 was your lucky year.

Are there 70000 different visitors to BW? 10 of them might see a close-matching pair of hands in a year, and you're the first to write yours up.

I think what I just wrote applies only to consecutive deals — your memory sucks and you can recall only the previous hand. If your memory is awesome and you're comparing your current hand with thousands of previous hands, the math changes a lot. If you can recall 7000 hands, you will see a match every 1000 days, or say 3 years.

I'm quite sure someone will tell us if I'm completely out to lunch with this line, but at least I gave it a try lol.
Dec. 27, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
I should perhaps have reiterated that north has the specified 5-3-3-2 shape in all those simulations.
Dec. 15, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
If south can be some 64xx, then south will have 0-1 spades about 4% of the time, and 2 spades about 13%.
Dec. 15, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
I simmed what spades the south hand might hold given a few crude constraints.

I don't know what East's 1 means, but I assumed it showed 3+ clubs, diamonds were not longer than clubs, and majors were 0-4.

South has anything in any suit, short of a 6-carder.

West has 5+ clubs, and 0-4 majors.

Given those parameters, I was surprised to find that south will have 0-1 spades just 2% of the deals, and a doubleton just 12%. North seems to be fully justified to try 4.
Dec. 15, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Overcall 1NT with hand #1. When RHO opens 1 and you hold a weak notrump, it's more your hand than theirs, and you sit over RHO, so get in there. Quite often 1NT will be the optimum contract by either side, so get there before they do. Play systems on, including some kind of scrambly Stayman, and you'll need some kind of run-out routine as well. Makes it difficult for opps to negotiate their 4-4 major fits. You'll find those 5-3 major fits when partner has the 5.At least at matchpoints, this is a great tactic.
Dec. 14, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Heh sorry I ironically got the position wrong. The writeup is incorrect, no wonder you were confused. The position is opener's rebid:

X = clubs
1NT = natural
2 = next suit = diamond rebid (and might be very strong)
2 = ‘good’ heart raise, better than
2 = heart raise

You do what you want with the two raises. In support-X style, you might make 2 = ‘good’ = the 4-suit raise. Or if you don't play support doubles, then 2 might be extras and when responder is weak you can get out in 2 when everybody else got too high with an invitational jump to 3.
Dec. 13, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
I'm the originator and we still play them. The transfer is always to the next suit up, so 2 would be whatever 2 would mean, here a cuebid of opener’s call.

They arise most often in Snapdragon and Support X positions. The most frequent gain is the additional raise type. The prime drawback is occasionally not recognizing the positions dammit.
Dec. 13, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
When using a DD engine, you can reduce statistical fluctuations by (1) tightening constraints, or (2) by running jillions of deals. Since it looked to me the two hands were pretty similar, I limited responder to 5-7 HCP just to save some time because I wanted a decent sample, 2K hands. Prefer another range? OK I just gave responder 10-12 HCP and nothing changes, both hands make 4 37% of the time. The two hands are freakily similar.

A look at the Kaplan-Rubens evaluator (admittedly just a general hand evaluation, zero context), also says the hands are equivalent, and actually gives a small edge to the hand you think is so weak. That evaluator makes a downward adjustment for 4333 hands, I think 1 point.

I would suggest that if you are going to proclaim that anybody who disagrees with the brilliancy of your analysis is ‘a very weak player’, you need a better defence than your ‘simple not not-finely-honed’ valuation method. Or better, just use less inflammatory language.
Dec. 13, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
On the other hand, both South hands perform almost identically when run through a double dummy engine. Hand #1 has an additional ace, good heart texture, and an AQx likely in good position. I gave North 4+ hearts and 5-7 HCP, East a hand that he might double with. With that minimum dummy, both South declarers made 2 about 45% of the time.
Dec. 13, 2017
Doug Bennion edited this comment Dec. 13, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
I don’t know that it adds a whole lot to the discussion, but if the leader was a double dummy engine:

On hand #1, the double dummy guy thinks the lead of a spade stands out, followed (considerably) by heart, club, diamond which trails badly …

On hand #2, he thinks best lead, by a fair bit, is a spade, followed by diamond, club, heart.
Dec. 12, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
The project sweeps all Kxxx holdings into a single ‘container’, regardless of spot ‘quality’. Sure, KT98 is a considerably stronger holding than K432, but the project is concerned with ‘relative’ lead value between holdings, not some absolute measure. The pot consisting of all kinds of variable-spot Kxxx will be compared to, say, the pot consisting of all kinds of Qxxx or xxx or AJxxx or whatever, so given a large enough sample, the good spots and bad spots will level out.

The project reports on 24 different holdings. In the 100K sample, Kxxx arises 20.7K times and is a ‘best lead’ 12.5K times. On the other hand, Jxxx arises 18.3K times, and is a best lead 12.8K of those. Which looks like a stronger lead?

Samples for all holdings are a good size. The most frequent holding is xx, at about 40K (many 4432 and 5422 shapes). Even the ‘rarest’ 5-card holdings are in the 4K range.

Incidentally there is a 4-card holding that gets fewer bests than Kxxx …
Dec. 12, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
And on average, about 2.7 suits (of 4, natch) tie for best lead.
Dec. 10, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
For what it’s worth. I’m in the middle of a little project on leads, and I was surprised to see that for about 15% of all deals (100K sample of hands playing in notrump), the winning lead is any card. For about 35%, at least one card from every suit is represented in the winning lead category. The difference is due to holdings like Kxxx, where any spot is right, but leading the king isn’t.

By the way, the project keeps begging me not to lead from Kxxx :-)
Dec. 10, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
They surely do, especially if you never open 1M with 5332.
Dec. 9, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
Kery, easy for you to say.
Dec. 9, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
A quick sim gets the same results, and also points out that partner will have 5+ hearts about 18% of the time. Don't know if Jim's methods can uncover those fits.

About half the hands with heart fits will have neither a spade nor diamond fit.
Dec. 8, 2017
Doug Bennion edited this comment Dec. 8, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
You don't need an American passport to get into Canada, any photo ID will usually do. You need it to get back into the USA.
Nov. 3, 2017
You are ignoring the author of this comment. Click to temporarily show the comment.
4333 with all suits stopped, is actually a great holding because it avoids ‘xx’ which is the weak spot with 4432 and 5332 hands.
Nov. 1, 2017
.

Bottom Home Top