[hotShot]

I done some experimenting comparing card play against DD play. The result was that experts lose about 1 trick in 4-5 boards to the DD-Solver while beginner lose about 1 trick a board.

If 2 beginner pairs play, the declarer will often get 1 more trick than the DD-solver, because when 2 opps lose a trick each and the declarer loses his average trick there is still an overtrick left.
The subjective impression is that they underbid and have to work on their bidding.

If an expert declarer plays 2 beginner opponents, his average gain will be about 2 tricks/ board, this is enough to compensate for being in the wrong partscore or game. Bidding is almost irrelevant.

[nigel_k]

When all players are world class, no doubt more IMPs change hands due to bidding than play. But this doesn't necessarily mean that differences in bidding skill have a larger effect than differences in card play skill. There is more luck in bidding.

[JLOGIC]

Even strong experts vs world class, or world class vs super world class, the difference is not going to be the bidding. I have played in the cavendish 5 or 6 times. I would say that you would be surprised at how many imps you win for making normal games on a good line. Many people even in that field just miss the play and go down. Why do Levin/Weinstein and Helgemo/Helness type pairs have a big edge in that field? It's because they drop very few tricks, it is not because they are bidding so much better IMO. Even in a field like that, it is mainly their cardplay winning them those events.

Also, I am typically about the 20 seed in the spingold/vanderbilt. Sometimes I get to meet a very high seed, and I usually lose. Why do I lose? I am confident it is because they play and defend consistently better than myself and my team. As Adam said, there might be more swings generated by bidding, but those are largely random/system swings. If our team gives up say 40 imps in cardplay, this is just insurmountable unless we get very lucky on slam hands/system swings. Giving up 40 imps may sound like a lot, but that is like 3 games and 2 partscores in 64 boards against one of the best teams in the world by 6 people combined, it is easy for even the 20 seed in the spingold/vanderbilt to give up this much.

I feel like people are understating the level at which bidding really becomes a bigger factor than cardplay, it is only between a handful of the best teams in the entire world.

[hrothgar]

JLOGIC, on 2011-February-24, 14:29, said:

I feel like people are understating the level at which bidding really becomes a bigger factor than cardplay, it is only between a handful of the best teams in the entire world.

Consider the following experiment:

Run two team matches using 8 copies of Jack, all running on identical hardware.

In the first match, all eight copies of Jack would play the same system.
The purpose of this match is to establish a baseline and determine how much variance gets introduced by various random decisions that Jack makes.

In the seocnd match, one team would play 2/1 Game Force
The second team would play something radically different (MOSCITO, Precision, what have you)

We could then examine whether the change in methods introduced a statistically significant increase in the variance of the board results...
If there was a significant difference, we could also categorize the magnitude.

[JLOGIC]

Then aren't we assuming completely equal cardplay? Obviously there will be much more variance with different systems, if they played the same system it would be a tie! (Unless jack is like gib in that it simulates and sometimes does different things in the same positions).

I think these are 2 different things though:

1) The edge to be gained in cardplay at pretty much every level is much more than the edge to be had in bidding, and increasing your cardplay skills a little bit would make you win more often in every event than increasing your bidding skills by a lot.

2) More swings are created in bidding because of system swings/luck.

Obviously #2 is just a function of what systems are being played. I totally agree that most swings will be CREATED by system swings if 2 teams play very different systems. I do not think this matters though, this evens out really quickly. I understand the high variance argument, but even in a 24 board match if you are getting out declared/defended badly, you are still very unlikely to win. And imo people who use that argument do too many wayy - EV things in the bidding that are just flat out bad, which hurts their chances to win. But that is a different argument.

I have been discussing #1, because I feel like that is what people are talking about. Being good in cardplay is just way more important, and is what wins events.

[MrAce]

JLOGIC, on 2011-February-24, 14:29, said:

I feel like people are understating the level at which bidding really becomes a bigger factor than cardplay, it is only between a handful of the best teams in the entire world.

I % 100 agree with this statement Justin, as well as all u said before about the difficulty of teaching/learning the card play/defense skills and of course finally the social amusement of bidding compared to play/defend stories.

JLOGIC, on 2011-February-24, 15:25, said:

I have been discussing #1, because I feel like that is what people are talking about. Being good in cardplay is just way more important, and is what wins events.

This is where i am confused. Depends on which event are u talking about to win, but i assume you are refering to biggest main events. Can we say "Being good in card play gets u to eliminate the % 90 of the field, but for winning an important event u need to outbid opponents " ? ( since the opponents you will face will be one of those handful teams and eventhough they can make mistakes in card play, it still wont be as much as mistakes that were made in bidding ? ) What do you think ?

EDIT : I am big time M.Rosenberg fan for example, and i believe he is the best card player ever. And he proved that in par contest individual. But it seems to me that, just like a lot of excellent card players, he makes his way all the way upto the line that seperates the winner from 2nd or 3rd, and fails there losing to those who has done much more work (or more accurate agreements) with their pd in the bidding. Not that he is unsuccesful, i just think he doesnt win as much as i think he deserves. (talking about teams )

[awm]

Perhaps bluecalm's thread gives some hard data on this.

He measured the difference in IMPs versus double dummy play in the same contract on a declarer-by-declarer basis. The difference between the best and worst declarers (but note that these are hands from top-flight events in the vugraph archives, so the "worst" declarers are still extremely good) was about 0.72 IMPs/board after accounting for the opening lead. The difference between the best and worst defenders was about 0.93 MPs/board after accounting for the opening lead.

He then tried to measure the best bidders, by looking only at the contract reached and computing what would happen on double-dummy play and defense (after the lead). This seems like a good estimate of IMPs picked up due to the contract reached (or contract the opponents reached) rather than due to play, which seems like a good measure of bidding acumen. The difference here between the best and worst bidders was about 1.06 IMPs/board.

Now obviously all of this is somewhat effected by small sample sizes... but it does seem to suggest that at the elite level of these vugraph events, bidding and play/defense are factors of roughly comparable significance. Bidding appears to be a little more important than play, but it's not anything like an 80%-20% breakdown.

Of course, adding in large numbers of somewhat weaker players might substantially change the results. Perhaps it'd be interesting to look through BBO tourney hands and perform the same experiments.

[hrothgar]

JLOGIC, on 2011-February-24, 15:25, said:

Then aren't we assuming completely equal cardplay? Obviously there will be much more variance with different systems, if they played the same system it would be a tie! (Unless jack is like gib in that it simulates and sometimes does different things in the same positions).

My understanding is that Jack's card play is non deterministic...
The primary reason that I suggested this type of test was to have some hard numbers to measure the variance that adopting a significantly different approach is able to impart.

As I recall, Gerben did some simulations a few years back describing the trade off between variance and expect value.
Don't suppose that you have that handy?

[helene_t]

hrothgar, on 2011-February-25, 08:39, said:

My understanding is that Jack's card play is non deterministic...

correct.

[cherdano]

hrothgar, on 2011-February-25, 08:39, said:

My understanding is that Jack's card play is non deterministic...
The primary reason that I suggested this type of test was to have some hard numbers to measure the variance that adopting a significantly different approach is able to impart.

As I recall, Gerben did some simulations a few years back describing the trade off between variance and expect value.
Don't suppose that you have that handy?

But even if two human teams adopt exactly the same system, there will be a lot of variance due to 50-50 judgment decisions or style decisions/opinions going one way or another. Bridge between humans always has more variance than the variance introduced by Jack's non-deterministic behavior in an all-Jack match. I think all your experiment would do is give an upper bound on the additional variance coming using different systems.

Anyway, I still think the original question was about the difference in IMP expectancy between in a match between a top team and a less good team - not about the expected variance.

[zasanya]

I suggest a simpler experiment .From the forum posters Lets select 4 excellent card players who have never played as partners and let them start with sayc without any further discussion .Let us select 2 pairs of volunteers who have often played together as partners and who believe they have a very good bidding system but whose card play is not in the same league as our excellent card players.Let them play a 32 deal match.The result could be interesting.

[mfa1010]

JLOGIC, on 2011-February-24, 14:29, said:

Also, I am typically about the 20 seed in the spingold/vanderbilt. Sometimes I get to meet a very high seed, and I usually lose. Why do I lose? I am confident it is because they play and defend consistently better than myself and my team. As Adam said, there might be more swings generated by bidding, but those are largely random/system swings. If our team gives up say 40 imps in cardplay, this is just insurmountable unless we get very lucky on slam hands/system swings. Giving up 40 imps may sound like a lot, but that is like 3 games and 2 partscores in 64 boards against one of the best teams in the world by 6 people combined, it is easy for even the 20 seed in the spingold/vanderbilt to give up this much.

Fortunately even the super strong teams are not flawless either. Imps do come the other way also. Especially if one succedes in putting up a reasonably tight game and thereby make sure that they are under pressure too.

JLOGIC, on 2011-February-24, 15:25, said:

I think these are 2 different things though:

1) The edge to be gained in cardplay at pretty much every level is much more than the edge to be had in bidding, and increasing your cardplay skills a little bit would make you win more often in every event than increasing your bidding skills by a lot.

2) More swings are created in bidding because of system swings/luck.

More imps are exchanged in the bidding that in the play. Even though luck/variance is a much bigger factor in bidding, it is very far from all just about luck. I think 'normal' edges in cardplay and in bidding are comparable in size. But I'm very uncomfortable about assigning percentages.

cherdano, on 2011-February-25, 10:21, said:

I still think the original question was about the difference in IMP expectancy between in a match between a top team and a less good team - not about the expected variance.

Yes. In principle we take two pairs and examine the edge of the better pair in cardplay and in bidding over a large sample of boards. Say they win 0.6 imps/bd in cardplay and 0.4 imps/bd in bidding. If this ratio is representative for comparisons between any two pair, then we could say that cardplay is 60% and bidding 40%.

[hrothgar]

zasanya, on 2011-February-25, 12:25, said:

I suggest a simpler experiment .From the forum posters Lets select 4 excellent card players who have never played as partners and let them start with sayc without any further discussion .Let us select 2 pairs of volunteers who have often played together as partners and who believe they have a very good bidding system but whose card play is not in the same league as our excellent card players.Let them play a 32 deal match.The result could be interesting.

My card play and defense are often considered to be epic (and not in a good way)...

TimG or Free might be willing to partner with me playing MOSCITO...
(Assuming that anyone had enough time to get back in practice)

[FrancesHinden]

hrothgar, on 2011-February-23, 09:57, said:

One might argue that the rules of the Spingold - and most any other ACBL event - are rigged to eliminate high variance systems...

It's easily possible to play a very natural system in a high variance way. Don't confuse the system with the way it is played.
Try strict 4-card majors (possibly with a longer minor), very light openings, totally random pre-empts, frequent* psyches and 1NT on all balanced/semi-balanced hand in range, for a non-standard range. Play some unusual overcall structure. You will rarely get to the same contract by the same declarer with the same information exchange.

*not so frequent or always of the same type as to come to a partnership understanding

The currently popular high-variance treatment in England (at least near me) is to play 2C and 2D openings as 4-4 in the suit opened and a second suit, 0-8 HCP. On the evidence so far, it's such a bad method it loses many more imps than it gains.

[FrancesHinden]

JLOGIC, on 2011-February-24, 14:29, said:

Even strong experts vs world class, or world class vs super world class, the difference is not going to be the bidding. I have played in the cavendish 5 or 6 times. I would say that you would be surprised at how many imps you win for making normal games on a good line. Many people even in that field just miss the play and go down. Why do Levin/Weinstein and Helgemo/Helness type pairs have a big edge in that field? It's because they drop very few tricks, it is not because they are bidding so much better IMO. Even in a field like that, it is mainly their cardplay winning them those events.

What's so frustrating is that when these hands are presented as play problems here on the forums (where a "good line" is necessary) I get them right about 95% of the time. The difference is getting them right 95% of the time at the table.

My feeling is that when I play a weaker (but not massively weaker) team, we pick up most imps from their defence. It's also so much easier playing against weak defenders, because they are predictable.

[JLOGIC]

Well imo good cardplay is not mainly about hands that are posted on forum type hands, it's routine hands where you deceive the opps or at least make their life as hard as possible, and routine hands where you have to figure out what the opps hold in their hands. Those types of hands are rarely posted here, and I think are the biggest difference in a Meckstroth type and some other type that is very strong technically but not as good of a declarer as Meck.

Quote

Even though luck/variance is a much bigger factor in bidding, it is very far from all just about luck. I think 'normal' edges in cardplay and in bidding are comparable in size.

I agree with the first part (obviously), but not at all with the second. Of course these are all just our opinions since we cannot quantify any of this.

Quote

Fortunately even the super strong teams are not flawless either. Imps do come the other way also. Especially if one succedes in putting up a reasonably tight game and thereby make sure that they are under pressure too.

Also agree with this, and that is why upsets are possible (especially when they aren't huge upsets). I think the best way to upset a team is to just have a day where you guys play really well. Most 20 seeds are capable of on a good day, outplaying a 3 seed if they're having a bad day imo. High variance low EV things are not really the way to go, also imo.

[Vampyr]

zasanya, on 2011-February-25, 12:25, said:

From the forum posters Lets select 4 excellent card players who have never played as partners and let them start with sayc without any further discussion.

I think that "bidding" in this thread (at least in the OP) was more about judgment and skill than system. Good bidders will get to more of the right contracts playing any (non-ridiculous) system.

[jogs]

gnasher, on 2011-February-21, 15:31, said:

Bidding is 80% of forum postings.

The points are won and lost from the play. The bidding is more interesting for post mortems.

PrecisionL, on 2011-February-21, 16:20, said:

Benito Garozzo (World Class by Marc Smith, 1999):

"In pairs competition, you can effectively forget about slam bidding. You need to concentrate on declarer play and defense - that is where most of the points are lost." Added 2/23/11 @ 9 am

I believe that statement.

[tolvyrj]

Is bidding 80% of bridge, its hard to say; but everything starts with it. If u r in a wrong contract it dosent really matter how good is your declarer play. And that goes with defence too, if u get into the bidding sequence and oppos still wins the contract to them selves, u r better prepared to the defence.
Jon Baldursson said that bridge is a bidders game and if u look the for example MeckWell in action the do bid a lot with all kinds of hands. They r exellent in both constructive and destructive bidding. What makes bidding difficult is that requiers good understanding between partners to function properly and lot of practise.
But its also true that u cant be a good player if u r just a good bidder and mediocre card player and vice versa. Those who r really good master both fields; like Rosenberg for example .

[hotShot]

Bidding is the most overrated part of bridge.
I think that is a consequence from the fact that most books and teachers lessons are about bidding.
I think that is because it is incredibly hard to teach declarer play and defense.

A good card player, playing on BBO will usually make1 trick more declaring, than the average player on BBO.
If his partner is a good card player too, than they will also average 1 trick more defending. (Mostly because a large number of players, declare and defend really bad.)
Even if their bidding would be below average, they would score at the partscores, they will punish overly optimistic game bids and even sub-standard bidder won't miss every game.
Such a pair should usually rank in the upper third of a BBO tourney result and the same is true for many club games.

Additionally you should think about this,
if someone is bad at card play, how can he judge the full potential of his hand and that of the combined hands?
If someones judgment is bad, how can he be a good bidder, if he bases his decisions on his lower trick taking ability?
Finally is it possible to be a good bidder, if you are a bad card player?

Based on this I am pretty sure that card play outranks bidding.