System changes

[42]

Good morning all!

Yesterday I read here: http://forums.bridgebase.com/index.php?showtopic=14039&hl=, where Fred says that his partnership made a system change.

I know an expert who loves to change the system every now and then (which leads sometimes to confusions...), especially when he watched wc players on vugraph or read their latest system scripts, or sometimes after a bidding accident.

My question now is: what makes an experienced and longtime expert-/wc-partnership change their system? I think that there doesn't exist the perfect system, every convention has its price, and a change will be well considered.

[Gerben42]

Depends on what exactly you mean. There are refinements, which might come up after some hand, small system changes which may be inspired by what other pairs do, or a total new system.

 

I was very surprised when Muller - de Wijs changed to relay Precision, but they did. Why? You probably have to ask them...

[Finch]

IMO, the biggest factor of all is psychology. Some people like tinkering, others don't. I have a good friend I play with occasionally - who has just won a major English championship - and he's a system fiddler. We've been playing more or less the same methods for nearly 20 years (gulp), but every time we play there's some new idea or improvement that he wants to play.

 

Against that, I play a lot with my husband and we play quite detailed methods, and they almost never change - maybe once a year, at most. He has a great reluctance to change things on the basis that the improvement is usually marginal and the risk of forgetting things is higher. If we were pros, and played full-time, we would probably make changes more often - though still at the same frequency per number of boards played! (By the way, we add things quite often: not changes, but sequences, usually in competition, that hadn't come up before or simply clarifying things that were murky. Our style is gradually evolving over time as well. )

 

There are many people around who love making changes - I know a pair where one of them will send 'system file updates' to his partner every week or so. I think he would (under duress) admit that these frequent changes are not good for their results, but they enjoy playing around with artificial methods, and they play bridge for fun.

 

I think there are four good reasons for making changes:

 

i) You are a top class full-time pro partnership and the edge you have from playing the best (rather than a 95% solution) is worth working at.

 

ii) You are a long-term partnership, but when you first started playing you didn't work a lot of things out in detail. (For example, I used only to play with one partner in 'fun' events; we played very little system. We've recently started playing more seriously, so we've improved some of the stuff that we were only really playing by default because we hadn't got round to discussing anything more complex.)

 

iii) You are a new(ish) partnership and don't yet know what you want to play, or are working on one particular area. We went through a phase of fiddling with our 1NT response structure over the course of a few months until we were happy with the final result, for example.

 

iv) You become convinced over the course of a large number of hands that what you are currently playing is definitely a long way short of 'best', and the pain of change is outweighed by the benefits. This is the same reason as when people change their methods after one bad result, but I think you need to apply a more long-term approach: only when you have a string of bad results and few good ones is it worth changing.

[hotShot]

I agree with you that there is no perfect system. After a while you get to know the weaknesses of your own system, which causes a sort of awareness for alternative methods that avoid these weaknesses.

So if you see an intersting treatment, that seems to fit to your system and bidding philosophy, you try it.

 

The problem of each convention is that it's hard to to see the full effect it has on the whole system.

It is obious that makes no sence to combine Benjamin with precision.

Playing a natural 5M-system changing your NT bid from 15-17 to 12-14 seems a small change, but the average strength of the 1m bids grows by 1.5 HCP and the average length is about 0.3 longer. This will have an impact on your aswering structure over 1m bids as well.

Many strong NT players respond to 1m openings with less than 6 HCP, if they are short in this suit. They do it, because usually partner is weak and there is no fit. They know that there is a risk to go down for a number.

Player that are used to play a weak NT, expect partner to be stronger and longer in his suit. They have the experiance that it is no big risk to pass.

[hotShot]

There is another aspect for weaker partnerships, the importance of the bidding system is overrated. If there were a supirior treatment, the laws if evolution say that this treatment will get dominant.

The fact that an WC level different systems are used, implies that no system is supirior.

Judgement makes the difference. If you force WC-player to use a different system, they might feel uneasy, but because of better judgement, better declarer play, better leads and better defence they will still dominate the field.

It is easier to blame the system than to admit, that your declarer play, your leads or your signaling was flawed.

[Free]

Sometimes you just have bids without meaning, and you encounter a certain situation where you don't have a solution to with the current methods. Here's an example for MOSCITO: after a 2♣ opening (9-14HCP, 6+♣), partner sits with a 5-5M invite hand. Relaying is no good sollution. We had the 3♦ bid free, so used that to show this type of hand. Problem solved.

 

However, if you don't have bids left, you should analyse first if it's worth a change or not. Sometimes a system just doesn't perform on a certain hand, but do you have to change everything just for 1 hand in a million?

 

I used to update a system every week, or even several times a week, but I found that my partners usually couldn't keep up. It's quite obvious for me to remember the new stuff, since I've done the work on it, I know how it works and why. But for partner who gets it in his hands, it's a lot more difficult. Sometimes they don't see the logic in it, or it's too complex, or whatever reason. Working on it together is the best solution.

 

With my favorite f2f partner, we decided not to make any changes for a while. The system definetly works ok, it could have some improvements, but we decided to play the system like it is now for a while and analyse it after an unknown period of time. This way we both get to learn the system in complete detail, and we'll have some experience where the system 'usually' fails :) I think if we'll make some changes, we'll probably make a lot of them at once, a real update. Tweaking at one particular part now and then isn't a good solution imo.

[helene_t]

I was very surprised when Muller - de Wijs changed to relay Precision, but they did. Why? You probably have to ask them...

They (especially de Wijs, I think Muller was more agnostic) were for some reason dissatisfied with their old system and figured that instead of tingering with it they might consider starting from scratch. Then they agreed to try out a number of different systems, each for a few weeks, and see if they could settle somewhere.

[42]

Gerben and hotShot, with my question I did not mean the system changes that weaker or more unexperienced players make, I am interested why a pro- or expert+/ worldclass partnership does within the system :) And since there is a system, changes often influence other sequences as well.

Changing the whole system is another case, like changing from 2/1 to a Relays or HUM system, for which reason ever.

[Gerben42]

If there were a supirior treatment, the laws if evolution say that this treatment will get dominant.

 

You'd hope so, but for bidding systems the Law of Exposure counts. People will play what they are taught or hear about, not what works. A great example of this is Cappelletti, which has it completely backwards. 2♦ is the worst possible bid to show both majors of the 4 options (Dbl, 2♣♦♥).

 

For an example of Caren's question, I guess the answer lies in some situations where the pair has been unsatisfied with the result, for example Fantoni and Nunes opening weaker (12 - 13 HCP) hands with both majors with 1M, even though in the rest of the system 1M show 14+. They realized that opening these hands 2M was a long-term loser so they changed it.

 

Another example can be found in "I love this game", where Sabine & Dany changed their defence to a strong ♣. Also I bet that the longer the partnership exists, the rarer system changes will become.

 

A world class pair that would just be starting would I guess start by taking an "off the shelf" standard of what the individuals are used to, say for example WJ2005, and then make some changes to fit their personal style right away. Then later on they might only add changes where they see holes, or to try something new.

[hotShot]

If there were a supirior treatment, the laws if evolution say that this treatment will get dominant.

 

You'd hope so, but for bidding systems the Law of Exposure counts. People will play what they are taught or hear about, not what works. A great example of this is Cappelletti, which has it completely backwards. 2♦ is the worst possible bid to show both majors of the 4 options (Dbl, 2♣♦♥).

Who said evolution is a fast process?

[hotShot]

Gerben and hotShot, with my question I did not mean the system changes that weaker or more unexperienced players make, I am interested why a pro- or expert+/ worldclass partnership does within the system :) And since there is a system, changes often influence other sequences as well.

Changing the whole system is another case, like changing from 2/1 to a Relays or HUM system, for which reason ever.

Caren the reasons to change your system are always the same.

 

1) Something does not work as well as planed.

2) You found a bidding sequence you and your partner interpreted differently (usually with a bad result).

3) Analysing your results you find a system based weak spot.

4) You got into a bidding sequence where you could not show your hand properly.

 

They are just less frequent in established partnerships formed by expert+ player.

 

Every decent bridge player will play a board as good as a world class player once in a while, world class player just do it 95+ of 100 boards.

[glen]

The four reasons given above do not cover why Greco-Hampson changed their system.

 

See: http://forums.bridgebase.com/index.php?showtopic=8270&st=26

 

Other reasons for moving include:

1) Using methods less error-prone (perhaps this is a superset of 2 above);

2) Dropping methods that never seem to come up;

3) Dropping methods that do not seem to produce over time;

4) Dropping methods to reduce the memory strain on the partnership (Larry Cohen is a fan of this);

5) Dropping methods that work against the partnership style;

6) Adopting methods that fit the partnerships style.

[hotShot]

The four reasons given above do not cover why Greco-Hampson changed their system

Since you seem to know a fith reason, why don't you share it with us?

At least that would be more interesting, than to know who did it.

 

Edit: Thank you officeglen for adding reasons.

 

Maybe they just wanted to do something completly different. :)

[paulg]

The four reasons given above do not cover why Greco-Hampson changed their system.

 

See: http://forums.bridgebase.com/index.php?showtopic=8270&st=26

 

Other reasons for moving include:

1) Using methods less error-prone (perhaps this is a superset of 2 above);

2) Dropping methods that never seem to come up;

3) Dropping methods that do not seem to produce over time;

4) Dropping methods to reduce the memory strain on the partnership (Larry Cohen is a fan of this);

5) Dropping methods that work against the partnership style;

6) Adopting methods that fit the partnerships style.

Summarises my view too.

 

When changing our system now I make a conscious assessment on the 'return on investment' - probable gains, learning effort, remembering investment, impact assessment on the rest of the system.

 

The only large change to get through recently has been the transfer responses to 1♣: gains include a more aggressive approach, a fair amount to learn but it comes up all the time so memory effort is relatively small, positive impact on the rest of the system.

 

Paul

[mike777]

change system?

 

One gets the impression there is almost no one out there who really understands in full what they are playing now. :).

 

I do not really accept the premise of the question, you are assuming these pairs really have a system that they know and understand.

 

Even Meckwell, Hamway and the Italians do not fully comprehend their partnership agreements under pressure. These changes are really just a process of trying to understand just what the heck their partnership agreements really are.

[dogsbreath]

dear Hotshot..

 

Evolution is wonderful .. in nature it created the SeaHorse and in bridge it created irish keyCard Gerber . Partnerships evolve or stagnate.

Rgds Dog :)

[awm]

For me, changing methods basically goes through the following set of steps:

 

(1) Identify a problem. This is normally a set of hands where a bad result was obtained. The result needs to be primarily attributable to bidding (i.e. not a bad opening lead or mistake in the play or defense). Note that a bad result is relative to the result you think you should obtain, not necessarily relative to the result obtained by the field (especially a weak field). Thus it's possible to identify problems even in bidding practice.

 

(2) Examine the solutions used by other partnerships in similar situations. Try to weigh costs versus benefits, and suitability for your partnership's general style. Keep in mind that memory is important here! For example some of the solutions used by Garozzo are probably very effective in partnerships that have no issues remembering them, but are perhaps not suitable for casual pairs.

If nothing seems like a good fit, then (and only then) consider "making something up."

 

(3) Trial and error. Usually this involves a lot of time in the BBO bidding practice room. Try to determine whether the new method is helping or hurting. If it seems good, then try it out in some real games. What are the ripple effects on the rest of the system? Are you and partner able to remember the new method? During this time system often changes a lot (almost day to day) as little tweaks are made to the new method.

 

Perhaps an easy example was that Elianna and I started off playing negative free bids. There were some immediate issues with determining which sequences were forcing after the "positive" double and suit rebid. We first considered making double followed by new suit forcing to game, with the direct free bid being "up to invitational." After trying this for a while, we decided that this gave the negative free bid too wide a range (it was too tempting to bid over it on marginal extras and this seemed to be wrong fairly frequently in practice). Eventually we gave up negative free bids and went back to forcing free bids, and have been happier since.

 

For an example involving weirder methods, Sam and I recently overhauled our strong club response scheme. Before, we had been playing a relay system in the "symmetric transfer oriented" style. The problem we noticed was that in a "control based" relay scheme locating key queens is extremely difficult. We often came across hands in bidding practice (and a few times in actual play) where a particular queen was the difference between a good slam and a bad one, but our methods couldn't discern whether this queen was there before the six level. After looking through several strong club response methods, including a fairly natural scheme, symmetric, moscito-style responses, and ultimate club, we eventually decided that basing our methods on "relay points" where A=3,K=2,Q=1 would be better than straight control showing. Our current method is something of a hybrid between symmetric and moscito.

[Elianna]

For an example involving weirder methods, Sam and I recently overhauled our strong club response scheme. Before, we had been playing a relay system in the "symmetric transfer oriented" style. The problem we noticed was that in a "control based" relay scheme locating key queens is extremely difficult. We often came across hands in bidding practice (and a few times in actual play) where a particular queen was the difference between a good slam and a bad one, but our methods couldn't discern whether this queen was there before the six level. After looking through several strong club response methods, including a fairly natural scheme, symmetric, moscito-style responses, and ultimate club, we eventually decided that basing our methods on "relay points" where A=3,K=2,Q=1 would be better than straight control showing. Our current method is something of a hybrid between symmetric and moscito.

You and Sam have actually tried out very different systems, if I recall. I seem to remember a month or so of a Polish-like club, when it was proclaimed as so fun and great before it was dropped. :)

 

I don't remember what came before that.

 

And did you play Recursive Diamond together? :)

[foo]

...And did you play Recursive Diamond together?  :)

WITW is "Recursive Diamond"?

 

"1D" "Alert!"

 

*curious voice* "yes?"

 

"A D bid commands me to make the cheapest D bid available after each of GOP subsequent calls until a later bid commands me to stop recursing or we have bid past 7D {this is obviously known as 'a D stack overflow' or 'blowing the stack'}"

 

*very confused* "huh? Is that legal?"

*both opponents, pitifully, in unison* "Director! help? our head hurts..."

[jdonn]

...And did you play Recursive Diamond together?  :)

WITW is "Recursive Diamond"?

 

"1D" "Alert!"

 

*curious voice* "yes?"

 

"A D bid commands me to make the cheapest D bid available after each of GOP subsequent calls until a later bid commands me to stop recursing or we have bid past 7D {this is obviously known as 'a D stack overflow' or 'blowing the stack'}"

 

*very confused* "huh? Is that legal?"

*both opponents, pitifully, in unison* "Director! help? our head hurts..."

You don't want to know...

 

It's a strong diamond system, with a 1♣ opener that can be a variety of different hands such as a certain balanced range, 3 suited, and another thing or two. Adam knows a lot more about it than I do, but I don't recommend it...

[awm]

Sam and I have gone through a lot of system changes because we both like to tinker, we have relatively good memories for complex stuff, and we're a strong enough partnership and playing against high enough level opposition that we think it matters. Basically:

 

We started off playing Recursive Diamond because we both knew it (because of partners in common) and it's goofy and fun. The system picks up a lot of good results on shapely/light/limited major suit openings and natural/intermediate 2-minor openings, but the sequences after the strong artificial 1♦ opening are nothing special. We decided there must be a way to improve our slam bidding after strong hands without sacrificing the rest of the opening structure. Looking at what many of our friends were playing (and methods again that we were both familiar with from other partnerships) we switched to a strong club system with transfer-oriented symmetric relay responses while retaining our 1M/2m opening structure.

 

The big problem at this point seemed to be that people always bid over our strong club. All this relay stuff we spent so much time perfecting basically never came up in practice, and we ended up in a lot of "guess" auctions at high levels. At this point we considered switching to a two-way club style opening structure, but we didn't want to give up on the relays since they seemed to work so beautifully in bidding practice (and we do occasionally have an unobstructed auction). After a bunch of work to try to hybridize polish club with symmetric relay we came up with something halfway decent and tried it for a while. The problem was, getting to the right spot after people preempt a polish club seemed even more difficult because responder can't bid freely on nine counts in case opener has a weak notrump. Rather than adopt negative free bids (which seems to be what many two-way club players do) which would take us well out of our comfort zone, we decided to give up on the experiment and revert to strong club. After all, plenty of top pairs play a big club, it can't be all bad... :)

 

More recently we discovered three issues with symmetric-style relays. One is the previously mentioned "finding the queen" problem. A second was the sequence 1♣(strong)-1♦(neg)-1♥(really strong) which burns a lot of space and also makes it harder to describe the minimum club openers that include hearts. The third was that sometimes the wrong hand seems to be doing the describing and it would be nice to give responder a chance to relay on some hands. After a fair amount of experimenting we decided to switch to AKQ points with direct bids showing "semi-positive up to min GF" and 1♦ being either desperately weak or a slammish GF. This also helps a bit with the "what happens when opponents interfere" issues because responder makes the ambiguous 1♦ call less often (and when he does and then opponents bid, opener usually wants to pass).

 

Another interesting situation arose when we noticed that in general our results over our (natural intermediate) 2♦ opening seemed to be better than over our (natural intermediate) 2♣. This is kind of weird because we obviously have more space over 2♣. We were playing basically natural bidding over the diamonds and something close to Wei precision over 2♣. We tried a bunch of different relay-based structures over the 2♣ with marginal improvement. After a lot of work, we devised a transfer-based structure that seemed to be excellent, and spent a lot of time in bidding practice perfecting it. Then we played a bunch of real hands, and realized that this structure was so complex (and totally different from anything else in our system) that we couldn't remember it, and scrapped it. Now we play 2M forcing and 2♦ relay (we can remember relays because we use the same relay structure throughout pretty much).

[foo]

<Advertisement mode on>

 

Adam, your post is basically an advetisement for an KS-like 2/1 GF, 1N= 12-14, 5cM, sound minor opening system with asking bids and relays in certain auctions.

 

Diving into a well developed KS system's 1m auction is =much= more dangerous than diving into a Strong C or Strong D auction.

 

...and it is much more flexible than any of the Forcing 1C or 1D systems.

 

<Advertisement mode off>

[MickyB]

<Advertisement mode on>

 

Adam, your post is basically an advetisement for an KS-like 2/1 GF, 1N= 12-14, 5cM, sound minor opening system with asking bids and relays in certain auctions.

Nono, you misread - he was advertising Siege.

[awm]

Well it's easy to read what I'm saying as advertising anything I suppose. Sam and my experience has been that controlled, light, natural openings are a big win and fit well with our general style. Of course, your experience may differ, but we find that we usually win by opening 1M with hands like:

 

AQxxx

x

Kxxx

xxx

 

We don't feel particularly comfortable with styles that advocate 1M with a range of 8-21 hcp because we find it difficult to control the subsequent auction. We don't like "four-card major maybe canape" styles because we feel that a big part of the advantage of "getting in early" is knowing opener's longest suit which kind of goes by the wayside if you open 1♠ with four small spades and a longer minor. We also feel that light major openings interact poorly with styles where responder must immediately decide whether to game force (i.e. 2/1 GF or GF relay) because the degree of strength necessary makes such a GF too infrequent.

 

We're also not big fans of the weak NT opening at vulnerable, having gone for -200 at MPs opposite no game way too many times.

 

Of course, there's no reason our experience has to match anyone else's. Some players are more comfortable with certain styles than others.

[foo]

We find that we usually win by opening 1M with hands like:

 

AQxxx.x.Kxxx.xxx

 

We don't feel particularly comfortable with styles that advocate 1M with a range of 8-21 hcp because we find it difficult to control the subsequent auction. We don't like "four-card major maybe canape" styles because we feel that a big part of the advantage of "getting in early" is knowing opener's longest suit which kind of goes by the wayside if you open 1♠ with four small spades and a longer minor. We also feel that light major openings interact poorly with styles where responder must immediately decide whether to game force (i.e. 2/1 GF or GF relay) because the degree of strength necessary makes such a GF too infrequent.

...and there's no problem opening disciplined light 1M openings w/ 2+ Quick Tricks and an easy rebid like the one above in a longest suit first based 2/1 GF style. You just have to make sure you and GOP are on the same wavelength.

 

If you are going to open hands like the above in a 2/1 GF context, then Responder has to make sure they do not shade their 2/1's in terms of controls and trick taking power.

...and your Declarer play has to be appropriate to how aggressively you are bidding.

 

 

We're also not big fans of the weak NT opening at vulnerable, having gone for -200 at MPs opposite no game way too many times.

YMMV. IME, _disciplined_ 1N= 12-14 openings are a win over the long run at all vulnerabilities in any form of scoring.

 

Disciplined 1N=12-14 openings, a decent 1N response structure, and good judgement will find good games (and avoid bad ones) that are more difficult to make the correct decision on in Strong NT systems.

 

When there is no game for Us on the board, disciplined 1N= 12-14 openings give Us an advantage in winning the partscore battle. TBF, this is negated to some extent by the fact that Weak(er) NT's mean We will play 1N on some hands where We should be in a suit contract that Strong(er) NT pairs will not.

 

IME, the pairs who most tend to get in trouble are the ones who use 1N= 11-14 as an undisciplined semi-preempt.