bluecalm

I also think Namyats is pretty terrible when we are at somewhat standard conventions. 4m preempt is useful and benefits of namyats comparing to opening those hands with 1M dubious at best.

I am curious, what program are you using for generating those hands ? Here are my results: I decided to do it on Lauria-Versace hands as they open 2D with 18-19 hands so additional bonus will be to see if they get into trouble on "problem layouts" (my initial hypotheiss is that they do). I have 9172 hands of them. 2D openings occured in 143 of them which is 1.56%. That's less than my guess in this thread. The opening faced 0-6 range in 52 of cases out of which I am able to pull 49 (I have some strange bug in db and some hands I didn't remove comments on mix up my .lin building process). I attach link to .lin file with those 49 hands at the bottom of this post. My quick review of those hands tells me they suffer here and would suffer significantly at matchpoints. Some hands with 2NT opposite 2PC, many bad games which probably could be avoided in other methods. On the other hand standard system would be much better here because of possibility of passing 1C which comes in handy with say 4 clubs and very weak hand. Anyway, the problems layout only occurs once every 176 hands (my intitial estimation of 1 in 100 was too pessimistic) and standard system reduced this number even more (as you can pass 1C). It seems like it's not that bad after all and I overestimated this problem. Lin file for reference: https://www.dropbox.com/s/ovehyqncjw4o3v1/2Dlaver.linpart0.lin

I play very little bridge lately (posting here while grinding poker on the side is about all bridge I get) but you gave me great idea - I am going to filter all 18-19 balanced from some major events and see how often those hands come up and give trouble "in practice". I will get back here with the results.

I have no doubt this is true. Yes. System is very small part of it. Why do you feel this thread benefits from reminding us of this well known fact ?

Yes but this is for one hand and there are two in a partnership so it's 2.2% altogether. Admittedly they sometimes open before us when we have 18-19 balanced but that won't happen too often. 2.2% is once every 45.5 hands. It's possible that I made a mistake or run too small a sample but 0-6hcp range is 50% of partner's range when we have 18-19 balanced according to both dmpro and dealer. Well, in my country people will be in those 1N and 2M 99% of the time if the split is 18-5, 19-5 and 18/9-6 gives them a choice of passing or inviting. It hurts to lose in such a way once every two tournaments. I think it's a major problem.

Yeah but admittedly they open (or at least used to as I didn't import the newest hands) 1D with 18-19 and 4D. I am interested in 18-19 bal part. 1D on diamonds is another topic and I think it's boring. I was responding to han's 3+ comment when I mentioned elite partnership. 1D with 18-19 and 4 D is another matter altogether (I still think it's worse than 1C but I think it's close).

Yes, very good. Far from elite though (elite is by definition very narrow group -pairs which win major things regularly ). Only Gitelman-Moss qualify by recent results but they aren't full time players let alone ones spending much time on honing their system.

And what would he do with 15-16 ? I think it's obvious 4H now anyway. 3H is too small a target with hand that distributional.

Nice story and good read but 2C bid at matchpoints is very bad imo.

Most people here (where I live) like split range Michaels so 2nd is the middle range for them. I think 3rd would be interpreted as very strong 6-4 hand and 4th would be interpreted as not so strong 6-4 hand which wanted to compete nevertheless. I expect it to be almost universal understanding among Polish players who think about those things. I am not a fan of it (I think always overcalling with a 2 suiter bid showing two suits is better if we have given suits) but that's how things are here.

There is useful rule that double is always t/o to pass or correct bid (two way doubles are an alternative to that). Here 4C could (and often will be) passed so we need t/o double as we can't wait with that. Once we know dbl is t/o to clubs it's easy to infer what 5C is. Similar but lower sequence: 1NT - pass - pass - 2C* pass- 2D** - dbl *-capp **-pass or correct Double should be t/o here because we are stuck with very likely 4-4-2-3 kind of hand and we can't afford passing as 2D might become a final contract.

I don't agree it's marginal. Strong hands are one thing, being able to compete in diamonds opposite weak hand is another (and freeing 2NT for other purposes after 1D is yet another) I know it's not a proof but it's still strong argument: there isn't one elite pair playing 1D as 3+ despite variety of systems and approaches represented in that group. EDIT: apparently Levin - Weinstein used to do that but it seems they open 1C now or maybe they vary it according to some other factors

Mainly because in comp when you open 1m and then double you have either real suit or 18-19bal. This is bad but it's twice as bad if you have this problem after both 1C and 1D.

Let's say we design standardish system with either T-Walsh, Walsh or up-the-line bidding. Our task is to decide what to do with 18-19 balanced hands. Some possible approaches are: a)standard: put it in both 1C and 1D and then jump to 2N b)improved standard: put it in 1C and jump to 2N after 1/1 response c)T-Walsh std: put it into 1C but then jump to 2N after transfer anyway d)T-Walsh alternative: put it into 1C and then always accept transfer with 12-14 (regardless of fit degree) and bid 1N with 18-19. Jumping is required only after 1S response. e)negative 1D: like in Polish systems, use 1D negative response and bit 1N with 18-19 after that; then normal 1/1 is constructive and jumping to 2N is not a problem f)put it in 2D opening Now, the major concern in standard is that if it goes: 1m - 1M - 2N then you are often too high already and if you wish to have a way to get out in 3 level partial your game/slam system suffers. This is a problem especially at matchpoints. This hand comes up about one time in 48 and my estimation is that in about half the cases you will suffer by not being able to stop in comfortable 1N or 2M contract. Another concern is competitive bidding when it's often important if partner has strong hand with real suit or 18-19 balanced (this is why a) is very bad). I feel that f) is clearly superior to b) because of competitive stuff but also constructive bidding is much better when we implement f). I feel however that cost of f) is still high - we lose 1N and 2H partials (as well as garbage stayman) comparing to d) and e). I feel this is big at matchpoints. Suffering once every two 50 hand sessions because of your system comparing to precision or polish club guys sounds bad not to mention that having 2D opening for other things is valuable too. This is why I am leaning toward eitehr d) or e). I have a lot of experience with e) and it's decent but 1C - 1D auctions are clumsy and the whole system is very passive making it easy for opponents to get into the bidding at low level. I have 0 experience playing d) but it looks tempting. My friend is currently doing this in his partnership and he likes it. Thoughts ?

Most hand evaluation methods focus on a situation when you know one hand and try to estimate how many trick in given suit you make knowing something about possible partner's hands. What I think is crucial for developing good bidding algorithm is a method to estimate how often contract make when you know two hands exactly (because then you can iterate all possible hands and make the best decision). Is any of you aware of any work done in this direction ? I have some thoughts but I would like to avoid reinventing the wheel if possible. I feel that dd simulations are not the way to go because: a)they are very slow b)they bias results for many honor combinations (for example AJT - Kxx is always 3 tricks when it reality it's more than 2.5 but less than 3).

I used the following dealer script: produce 10000 predeal west SJ97, H964, DJ953, CAJ2 predeal east SAKT432, HAQ2, D642, C6 hcp(north) >= 9 && hcp(north) <= 12 && shape(north, any 4333 + any 5332 + any 4432 +any 5422) && clubs(south) >= 5 && clubs(south) <=6 && hascard(north, QC) && hascard(south, KC) action printcompact And then count spade holdings for N. The results: Qxx - 30.72% xx - 27.74% Qx - 26.29% xxx - 7.96% Qxxx - 7.29% Or 64.3% for Qs. I feel that those conditions favor playing for drop too as they don't contain that max range is more likely for N (as rhm pointed out), allow for AKD♦ in N hand and for 5-4-2-2's including hands with 5 hearts. Still 64.3% is more than point count alone suggests. Assuming I didn't make any mistake in the conditions, does anybody has an idea how to arrive at such estimation at the table ? Anyway it's much higher than I expected. I now think Helgemo blundered on this hand.

I would be more happy to play most things in this thread than: -capp -better minor -standard way of handling 16-21 hands (which is jumping or reversing on 2-3 card suits) Those are not only terrible but also very frequent so they are more costly than say 1S -p-2S-p-2N = RKCB :)

I like this point.I think you are right.

This Pszczola is a lucky guy to partner such a player.

Just in case you missed it, my calculations were wrong and yours were correct. Sorry for confusion. As to sensational deals. I have different way of assessing this. For example I have a script which pulls all cardplay misplays from vugraph (in double dummy sense) for given player and put them into .lin files. I then go through them all. Helgemo is one of those guys who impressed me to such extent I am always careful to call "mistake". I am sure he makes those I just want to make sure as too many times my intuition was wrong.

Maybe Helgemo thoughts so as he played for a drop and it was down 3. In all seriousness I think Helgemo is one of the very few people who are good at this game and every time he does something and I doubt his decision I am assuming my intuition is wrong until I see very convincing evidence to the contrary. I haven't so far.

I think he might be reluctant to compete seeing probable misunderstanding of the opponents (and 3S likely being better contract than 2N). Let's ignore this possibility for the purpose of analysis. My quick script shows that there are 27 such combinations and 14 of them contain Qs (so that confirms your calculations almost). Those are: (A, K, K, Qs, Q, J) (0, 0, 1, 1, 1, 0) hcp: 7 Qs HERE (0, 0, 1, 1, 1, 1) hcp: 8 Qs HERE (0, 1, 0, 1, 1, 0) hcp: 7 Qs HERE (0, 1, 0, 1, 1, 1) hcp: 8 Qs HERE (0, 1, 1, 0, 0, 1) hcp: 7 (0, 1, 1, 0, 1, 0) hcp: 8 (0, 1, 1, 0, 1, 1) hcp: 9 (0, 1, 1, 1, 0, 0) hcp: 8 Qs HERE (0, 1, 1, 1, 0, 1) hcp: 9 Qs HERE (0, 1, 1, 1, 1, 0) hcp: 10 Qs HERE (1, 0, 0, 0, 1, 1) hcp: 7 (1, 0, 0, 1, 0, 1) hcp: 7 Qs HERE (1, 0, 0, 1, 1, 0) hcp: 8 Qs HERE (1, 0, 0, 1, 1, 1) hcp: 9 Qs HERE (1, 0, 1, 0, 0, 0) hcp: 7 (1, 0, 1, 0, 0, 1) hcp: 8 (1, 0, 1, 0, 1, 0) hcp: 9 (1, 0, 1, 0, 1, 1) hcp: 10 (1, 0, 1, 1, 0, 0) hcp: 9 Qs HERE (1, 0, 1, 1, 0, 1) hcp: 10 Qs HERE (1, 1, 0, 0, 0, 0) hcp: 7 (1, 1, 0, 0, 0, 1) hcp: 8 (1, 1, 0, 0, 1, 0) hcp: 9 (1, 1, 0, 0, 1, 1) hcp: 10 (1, 1, 0, 1, 0, 0) hcp: 9 Qs HERE (1, 1, 0, 1, 0, 1) hcp: 10 Qs HERE (1, 1, 1, 0, 0, 0) hcp: 10 So it's almost 14/27 = 51.85% If we eliminate AKQ of diamonds (2 possibilities with or without side jack) it gives: 14/25 = 56% (I made a mistake before in this post, now corrected) I don't think that's the end of the story though as we need to consider how often opener has 3 or 4 spades opposite him having 2 and those doesn't seem that easy to estimate in independent way. I may end up just dealing those hands and counting occurrences but it would be nice to estimate it without help of the dealer program to get some intuition about what are major factors influencing the odds here.

How I wish one day be a part of a partnership willing to endure system which such detailed agreements. I like it though, very nice.

Some ballpark estimation would be nice. I am trying to arrive at it but I am unsure if I am doing it right.

Really ? I remember some examples from the book and those were real suits with fit. Why not play some art rais here instead of taking two useful natural bids to show similar thing ? One can play 2NT as fit showing good raise and 3C/3D as natural. If I were to play 3C/3D as fit showing then 3card raise and 4card raise is certainly more useful than suit+fit. That's the whole point. With support you can use 2NT for example and land on your feet as you just conveyed the most important information about your hand (the support) while loses from not being able to show yout 6-7 carder are much more painful. If a guy on your right (ie the one who bids just before you) bids a suit then you are leading ! Anywya, in my country opinions are split. I feel most people play it as fit showing which seems so ridiculous to me that I needed to make a thread here :)