[fazzzoola]

OK, good news: after wrestling with all the good comments and some research, I have come up with an alternate "plan" - split GaLwood into two parts, the first, a basic system for novices and intermediates, the second for solid intermediates and above. The web site has been updated for the basic system only, with some structure for the future more advanced system.

Please understand that this is like a first attempt at the new structure. Any constructive criticism will be appreciated.

Thanks

[straube]

I think it's very awful. A few of the reasons...

1) 1N as 12-15 or 18+? So partner bids something with 8 points and finds out partner has the weak NT or passes with 7 and finds out partner has the strong NT. Unplayable.
2) Opening at the 4-level with 18+ and a 7-cd suit? You will be too high to start a conversation with your partner. You will be too high sometimes to make your bid. You have bypassed 3N.
3) Opening 1S with AKxx x xx AKxxxx or AKxxxx x xx AKxx? Partner will have to guess what 1S-1N, 2C shows.
4) 1C as 18+ usually unbalanced? Way too infrequent for such an important opening. What is 1C-1D, 1N?
5) What do you do with balanced 16-17?

I think that the core of your Galwood idea is your desire to open at the 1-level with 2-suited hands and at the 2 or higher level with 1-suited hands. Is that right? If so, you are distorting your openings around this concept and doing it at the expense of your 1N (and now) 1C openings. You also are well-versed in the probabilities for 1-suited vs 2-suited etc patterns. I think you are missing something that matters much more.

I would recommend that you study how the Fibonacci sequence applies to bidding. I think you would also do well if you studied symmetric relay systems. These two things would give you a better idea for how much information ought to be carried in a particular bid. Basically, the lower the bid, the more ambiguous it should be.

I asked you before to map out the frequency of your openings. Have you done that for this latest? I would guess...

1C-7%
1D-10%
1H-12%
1S- 15%
1N-30%
etc

Do you see the problem here?

You are focused on describing opener's hand. You frequently let opener show too much preference or too much direction. Your 2-level bids are arguably ok, but your 3 and 4-level openings are much too preemptive. You prevent responder from describing his own hand and prevent opener from describing anything more about his hand.

Here's an example of opening bids that are designed with preference in mind. See how very different these openings are from what you've been coming up with...

1C-16+
1D-2+ diamonds
1H-4+ hearts
1S-5+ spades
1N-13-15 balanced
2C-6+ clubs
2D-3-suited, short diamonds

Each successive suit opening has a higher length requirement. This means that each higher suit opening is less frequent than its predecessor and shows a fewer number of possible hand patterns.

Here's our system which hopefully illustrates the same idea...

1C-16+
1D-0+ diamonds
1H-5 hearts
1S-5 spades
1N-14-16
2C-6 clubs, could have 4 diamonds or 4 spades
2D-6 diamonds, could have 4 clubs
2N-6 clubs, 4 hearts

The other major thing I think you should work on is how to show your balanced hands in useful (probably 3 pt) ranges. For Precision it's typically...

1D-1M, 1N 11-13
1N 14-16
1C-1D, 1N 17-19
2N 20-21
1C-1D, 2N 22-23

Balanced hands are very common and very easy to describe and very useful for partner to know about.

[fazzzoola]

straube, on 2011-March-27, 12:14, said:

I think it's very awful. A few of the reasons...

1) 1N as 12-15 or 18+? So partner bids something with 8 points and finds out partner has the weak NT or passes with 7 and finds out partner has the strong NT. Unplayable.
2) Opening at the 4-level with 18+ and a 7-cd suit? You will be too high to start a conversation with your partner. You will be too high sometimes to make your bid. You have bypassed 3N.
3) Opening 1S with AKxx x xx AKxxxx or AKxxxx x xx AKxx? Partner will have to guess what 1S-1N, 2C shows.
4) 1C as 18+ usually unbalanced? Way too infrequent for such an important opening. What is 1C-1D, 1N?
5) What do you do with balanced 16-17?

I think that the core of your Galwood idea is your desire to open at the 1-level with 2-suited hands and at the 2 or higher level with 1-suited hands. Is that right? If so, you are distorting your openings around this concept and doing it at the expense of your 1N (and now) 1C openings. You also are well-versed in the probabilities for 1-suited vs 2-suited etc patterns. I think you are missing something that matters much more.

I would recommend that you study how the Fibonacci sequence applies to bidding. I think you would also do well if you studied symmetric relay systems. These two things would give you a better idea for how much information ought to be carried in a particular bid. Basically, the lower the bid, the more ambiguous it should be.

I asked you before to map out the frequency of your openings. Have you done that for this latest? I would guess...

1C-7%
1D-10%
1H-12%
1S- 15%
1N-30%
etc

Do you see the problem here?

You are focused on describing opener's hand. You frequently let opener show too much preference or too much direction. Your 2-level bids are arguably ok, but your 3 and 4-level openings are much too preemptive. You prevent responder from describing his own hand and prevent opener from describing anything more about his hand.

Here's an example of opening bids that are designed with preference in mind. See how very different these openings are from what you've been coming up with...

1C-16+
1D-2+ diamonds
1H-4+ hearts
1S-5+ spades
1N-13-15 balanced
2C-6+ clubs
2D-3-suited, short diamonds

Each successive suit opening has a higher length requirement. This means that each higher suit opening is less frequent than its predecessor and shows a fewer number of possible hand patterns.

Here's our system which hopefully illustrates the same idea...

1C-16+
1D-0+ diamonds
1H-5 hearts
1S-5 spades
1N-14-16
2C-6 clubs, could have 4 diamonds or 4 spades
2D-6 diamonds, could have 4 clubs
2N-6 clubs, 4 hearts

The other major thing I think you should work on is how to show your balanced hands in useful (probably 3 pt) ranges. For Precision it's typically...

1D-1M, 1N 11-13
1N 14-16
1C-1D, 1N 17-19
2N 20-21
1C-1D, 2N 22-23

Balanced hands are very common and very easy to describe and very useful for partner to know about.

Thanks for the reply (I know it takes effort to do so.)

I'm going to tke this, copy it into a some form I can study. This, obviously needs careful study and cogitation.

Please give me a day or two (?). Thanks (very much)

[fazzzoola]

Thanks for the edification!

----------------

1) 1N as 12-15 or 18+? So partner bids something with 8 points and finds out partner has the weak NT or passes with 7 and finds out partner has the strong NT. Unplayable.

reply:
I see your point. Change made... TY.

------------------------

2) Opening at the 4-level with 18+ and a 7-cd suit? You will be too high to start a conversation with your partner. You will be too high sometimes to make your bid. You have bypassed 3N.

reply:
OK, deleted that option. I'm not really sure why I had that one in there. Brain fart perhaps? Those do happen to us "older" folk. LOL

-----------------


3) Opening 1S with AKxx x xx AKxxxx or AKxxxx x xx AKxx? Partner will have to guess what 1S-1N, 2C shows.

reply:
In the case of the club suit, that is a problem, however, Would not a rebid of spades, probably at the 2 level, show 5+ spades, and still show 4+ clubs? And a future rebid of clubs show 5-5 distribution or better?

I'm beginning to think of reverting back to 1C as 12-15(16?) and no 5CM.

However, I have come to the conclusion that two versions need to be created, one for novices and lower level intermediates, and one for advanced intermediates and above. Much of the material recommended,
suggested, is rather daunting. Not that I have any fears of tackling it; just that it requires a helluva lot of digesting! lol


----------------------


4) 1C as 18+ usually unbalanced? Way too infrequent for such an important opening. What is 1C-1D, 1N?

reply:
I would take that to mean even distribution, no 5 card suits, at least 16 HCP (in case of 4-4-4-1 type shape), and unable to open with 2N...

---------------


5) What do you do with balanced 16-17?

reply:
Actually, I think I can argue that one... Doesn't the same situation arrise with "regular" systems: 1N=16-18 or 15-17, and 2N=20-22?... What happened to 18 and 19? (And yes, I did note it was accounted for in your recommendations at the end.)

-------------

I think that the core of your Galwood idea is your desire to open at the 1-level with 2-suited hands and at the 2 or higher level with 1-suited hands. Is that right? If so, you are distorting your openings around this concept and doing it at the expense of your 1N (and now) 1C openings.

You also are well-versed in the probabilities for 1-suited vs 2-suited etc patterns. I think you are missing something that matters much more.


reply:
You are quite correct about my desire - which may well get quashed, as I am seeing and learning quickly. The Truscott structure is rather precise, but, I can't see a novice or regular intermediate actually using it. (I guess I'd better "improve" myself, huh? LOL

So, back to the drawing board. Back to more studying and experimenting.

-----------------

I would recommend that you study how the Fibonacci sequence applies to bidding. I think you would also do well if you studied symmetric relay systems. These two things would give you a better idea for how much information ought to be carried in a particular bid. Basically, the lower the bid, the more ambiguous it should be.

reply:
Now, there's a gem! Thanks. But now another question arises: if the more ambiguous an opening bid is, does that not violate one or more of "David's Bidding Principles", especially relating to giving the opponents too much room to interfere?

---------------

I asked you before to map out the frequency of your openings. Have you done that for this latest? I would guess...

1C-7%
1D-10%
1H-12%
1S- 15%
1N-30%
etc

reply:
In random distribution of hands:
7-11 HCP occurs 38.9% of the time;
12-15 HCP occurs 31.6% of the time;
16-17 HCP occurs 15.2% of the time;
18+ HCP occurs 8.8% of the time.


Now comes the "tricky" part, combining HCP distributions with suit distributions. I am not even going to try - brain slowly frying... lol

To answer your question about mapping the occurences of various bids:
1C (showing 18+ HCP) should occur roughly 40% of the time.
1D/H/S should occur 60% of the time, obviously w/ some overlap w/1C - apples and oranges.

So, now come the big questions to be resolved:
1. Open with how many points? Current practice seems to favor 12.
2. Should 1N opening be the big gun, 16+ HCP, no 5-card suit?
3. Should any 2 level opening remain preemptive and use opener's 2nd round bid be a jump shift, indicating 16+ HCP and a long suit?


--------------------------------

Do you see the problem here?


reply:
Yes, I think so. However, that answer comes from a brain-fried status. LOL

---------------

You are focused on describing opener's hand. You frequently let opener show too much preference or too much direction. Your 2-level bids are arguably ok, but your 3 and 4-level openings are much too preemptive. You prevent responder from describing his own hand and prevent opener from describing anything more about his hand.

reply:
This stems from my own thinking (strategy) that responder being second to bid, is the first one to find out about opener's hand, and, therefore, opener should strive to let responder know as much as possible.

What you seem to be suggesting is the opposite, that opener should provide as little information as possible, while responder begins to supply information him/herself. I suppose that this strategy gives the LHO little if anything to go on, and may well shut out the RHO, if responder has any good values.

Yet, in the opening bids you described, the LHO seems to have some pretty good idea about opener's hand, points and shape. So, where is the advantage?


-----------------

Here's an example of opening bids that are designed with preference in mind. See how very different these openings are from what you've been coming up with...

1C-16+
1D-2+ diamonds
1H-4+ hearts
1S-5+ spades
1N-13-15 balanced
2C-6+ clubs
2D-3-suited, short diamonds

Each successive suit opening has a higher length requirement. This means that each higher suit opening is less frequent than its predecessor and shows a fewer number of possible hand patterns.

Here's our system which hopefully illustrates the same idea...

1C-16+
1D-0+ diamonds
1H-5 hearts
1S-5 spades
1N-14-16
2C-6 clubs, could have 4 diamonds or 4 spades
2D-6 diamonds, could have 4 clubs
2N-6 clubs, 4 hearts

The other major thing I think you should work on is how to show your balanced hands in useful (probably 3 pt) ranges. For Precision it's typically...

1D-1M, 1N 11-13
1N 14-16
1C-1D, 1N 17-19
2N 20-21
1C-1D, 2N 22-23

Balanced hands are very common and very easy to describe and very useful for partner to know about.

reply:
Would it be OK with you to borrow off these ideas?

--------------------------

last comment:
Well, I "knew" I had a lot to learn... Just didn't know how much I didn't know. I'm going to have to review, digest, and "ruminate".

Again, my humble thanks for your help!

[Vampyr]

fazzzoola, on 2011-March-27, 19:00, said:

However, I have come to the conclusion that two versions need to be created, one for novices and lower level intermediates, and one for advanced intermediates and above. Much of the material recommended,
suggested, is rather daunting. Not that I have any fears of tackling it; just that it requires a helluva lot of digesting! lol

Not sure if this is worth the extra effort -- newer players do not tend to be drawn to home-grown systems.

[fazzzoola]

Vampyr, on 2011-March-28, 06:31, said:

Not sure if this is worth the extra effort -- newer players do not tend to be drawn to home-grown systems.

I'm not sure either, however, it'll still be a great learning experience and may yet lead to other "good" things. I think the two systems ought to be named GaLwood Simple Simon, and GaLwood Expert Simon.

[mtvesuvius]

I like how the acronym will be "GES", which sounds a lot like "Guess". Quite fitting for this system IMO.

[straube]

mtvesuvius, on 2011-March-28, 17:22, said:

I like how the acronym will be "GES", which sounds a lot like "Guess". Quite fitting for this system IMO.

Clever, but I wouldn't make fun. He's admittedly in the learning process.

[glen]

straube, on 2011-March-28, 21:26, said:

Clever, but I wouldn't make fun. He's admittedly in the learning process.

"Learning process" = good objective for person
Creating new bidding system for "novices and lower level intermediates" = not a good objective for the bridge community, though it make take some learning process, and some remarks (e.g. is a guess system at all good for novices?), to see the concerns

[fazzzoola]

glen, on 2011-March-29, 06:44, said:

"Learning process" = good objective for person
Creating new bidding system for "novices and lower level intermediates" = not a good objective for the bridge community, though it make take some learning process, and some remarks (e.g. is a guess system at all good for novices?), to see the concerns

OH you guys! (gals?) OK, you do deserve some "fun" at my expense. I am not blonde, just bald! (Brain cells evaporating through top of head due to lack of insulation (hair). Of course, if one wanted to be a sit-down comic, like myself, one could define the acronym B.A.L.D. as brain always (in) low drive. LOL

OK, definitely not "GES"!!!! but, pehaps GALES? Now, doesn't that sound a bit better?

[fazzzoola]

Just for the record, I've added the random distributions of shapes and HCP in 40,000,000 randomly generated hands.

You'll find them on galwood.com, random distribution button - left side, bottom of home page.

Virtually the same distribution data can be found at two other web sites: http://www.bridgehan...istribution.htm, and
http://crystalwebsit...nd_patterns.htm.

[PrecisionL]

Your hcp frequencies are inaccurate:

44.6 % 7 - 11 hcp
15.1 % 12 - 15 hcp
05.7 % 16 - 17 hcp
04.1 % 18+ hcp

Reference: bridge odds complete, F. H. Frost, 1971, 2nd edition

Edited: 4/2/11: My minor was math and statistics and I independently verified many of Frost's data before the age of electronic calculators (I used a mechanical calculator!). Something appears wrong with your 'random' generation. The Encyclopedia of Bridge also lists data that exactly matches Frost's publication.

[fazzzoola]

PrecisionL, on 2011-April-01, 04:35, said:

Your hcp frequencies are inaccurate:

44.6 % 7 - 11 hcp
15.1 % 12 - 15 hcp
05.7 % 16 - 17 hcp
04.1 % 18+ hcp

Reference: bridge odds complete, F. H. Frost, 1971, 2nd edition

Well, I just double checked all three of my sources.
1. the 40,000,000 hands shows 7-11 hcp = 44.6%; 12-15 hcp = 25%; 16-17 hcp = 5.7%; 18+ hcp = 4.1%
2. the second source shows: 7-11 hcp = 44.6%; 12-15 hcp = 25%; 16-17 hcp = 5.7%; 18+ hcp = 4.1
3. the third source shows: 7-11 hcp = 35.1%; 12-15 hcp = 32.2%; 16-17 hcp = 10.8%; 18+ hcp = 8.8%.

For the moment, I have to trust my sources, at least in general. There The 40,000,000 was an actual random generation of each hand. I can "accept" a flaw in the original source, if, and only if, I accept the possibility the the randomness was not truly random as one was led to believe. Yet, the data seem to be relatively consistent with only one or two "blips".

I have yet to check F.H. Frosts data, but, I will look for it.

[fazzzoola]

OK, with all the "help", comments, suggestions, etc., I've come up with a Simple Simon version, which should be plenty good enough for beginners and intermediates:


1C - strong, 17+ hcp, any shape
1D - 12+ HCP, and no 5+ major
1H - 12+ HCP, 5+ hearts
1S - 12+ HCP, 5+ spades
1N - 15-17 HCP, even dist
2X - 9-11 HCP, 6+ suit
2N - 20+ HCP, even dist, 5 aces/kings
3X - 8-11 HCP, 7+ suit

[matmat]

Something is wrong.

your data shows that 0% of hands have 35-37%?

10HCP should be the most likely outcome.



The HCP distributions are calculable (i wouldn't want to do this, as it is tedious, but I believe Richard Pavlichek has )

Let's take the simplest case, 0HCP. This # of possible hands is equivalent to being dealt 13 out of 36 non-face cards.

so 36 choose 13 = 2310789600
out of a possible 52 choose 13 = 635013559600

so you should be getting 0 HCP roughly 0.3639% of the time. Not 0.05% as your numbers state...

naively i would expect errors on the order of 0.02% in your model; since the values appear to differ by more than this, I'd guess there's something not right with your random number generator.

[straube]

fazzzoola, on 2011-April-01, 16:07, said:

OK, with all the "help", comments, suggestions, etc., I've come up with a Simple Simon version, which should be plenty good enough for beginners and intermediates:


1C - strong, 17+ hcp, any shape
1D - 12+ HCP, and no 5+ major
1H - 12+ HCP, 5+ hearts
1S - 12+ HCP, 5+ spades
1N - 15-17 HCP, even dist
2X - 9-11 HCP, 6+ suit
2N - 20+ HCP, even dist, 5 aces/kings
3X - 8-11 HCP, 7+ suit

That's a lot better. Still quibble with 2N and the 2m openings, but still a lot better. Hope it works for you.

[campboy]

matmat, on 2011-April-01, 17:05, said:

The HCP distributions are calculable (i wouldn't want to do this, as it is tedious, but I believe Richard Pavlichek has )

It's not that tedious if you just want a frequency table.

FOR a,k,q,j = 0 TO 4
   p = 4*a + 3*k + 2*q + j
   n = (4 CHOOSE a)*(4 CHOOSE k)*(4 CHOOSE q)*(4 CHOOSE j)*(39 CHOOSE (13-a-k-q-j))
   f[p] = f[p] + n
NEXT j,q,k,a

Add in a function to calculate x choose y (making sure that it returns 0 if y is out of range) and you're good.

[fazzzoola]

straube, on 2011-April-01, 21:48, said:

That's a lot better. Still quibble with 2N and the 2m openings, but still a lot better. Hope it works for you.

So far, so good. ... p and I had good results on bbo this a.m.
Now, just need to work out the rest of the simple simon version. Once that is done, then will DEFINITELY need help with the Expert Simon, as I am no expert... lol


Thanks for taking the time and effort. Hopefully we can collaborate some more?

[fazzzoola]

matmat, on 2011-April-01, 17:05, said:

Something is wrong.

your data shows that 0% of hands have 35-37%?

10HCP should be the most likely outcome.



The HCP distributions are calculable (i wouldn't want to do this, as it is tedious, but I believe Richard Pavlichek has )

Let's take the simplest case, 0HCP. This # of possible hands is equivalent to being dealt 13 out of 36 non-face cards.

so 36 choose 13 = 2310789600
out of a possible 52 choose 13 = 635013559600

so you should be getting 0 HCP roughly 0.3639% of the time. Not 0.05% as your numbers state...

naively i would expect errors on the order of 0.02% in your model; since the values appear to differ by more than this, I'd guess there's something not right with your random number generator.

I have 3 sources, 2 of which are noted on the web site, and all three differ: 0.05%, 0.00% and 0.36%... since my primary ssource was a 40,000,000 random generated effort, I'd have to stick with it, in that it was an actual set of deals and not mathematical calculations (estimates) of probability. However, having said that, I do recognize that some calculations are more "accurate" than others, and yet, for my general purposes here - since I'm not splitting hairs - it'lll do? (Less than 1% is rather rare... <smiles>)

[PrecisionL]

fazzzoola, on 2011-April-02, 08:00, said:

I have 3 sources, 2 of which are noted on the web site, and all three differ: 0.05%, 0.00% and 0.36%... since my primary source was a 40,000,000 random generated effort, I'd have to stick with it, in that it was an actual set of deals and not mathematical calculations (estimates) of probability. However, having said that, I do recognize that some calculations are more "accurate" than others, and yet, for my general purposes here - since I'm not splitting hairs - it'll do? (Less than 1% is rather rare... <smiles>)

Some calculations are more "accurate" than others? I am sorry but this statement does NOT apply to these probability calculations of HCP. They can be calculated precisely, so your 'random' generation simulation appears to have an error.

Not to split hairs, but 0.3639 % for zero hcp is correct. :<)