[jillybean]

We have one game, one set of rules but as expected, players have their own, individual style of playing the game, some even say this is how it must be played.
Unfortunately I think club players are most guilty of using MUST far too often. I have found that there are very few MUSTs in this game. I must not break the laws.

One approach that is circulating among the newer players in this area is Just Bid More! The idea being that defence is hard and players who defend more hands score less than when they are declarer, 10% less.

I know this is just another rumour. Defence is hard, every aspect of this game is hard but when the hand belongs to the opponents, good defensive agreements will get you the extra MPs.

What I am interested in is, are there any actual statistics which show Declarer or Defence having any advantage on scoring?

[DavidKok]

Perhaps not the answer you are looking for but here are some related thoughts:

• Richard Pavlicek has extensively studied the gap between single dummy play and double dummy play. At the risk of oversimplifying: single dummy the declarer has an advantage. To me this suggests that defending is more difficult than declaring.
  
• The answer to your question is very senstive to exact phrasing, and difficult to answer regardless of choice. There is a strong correlation between playing the hand and having an above average strength hand, so of course we expect declarers to go positive on balance. More relevant, and complicated, is the question of whether bidding more on any given hand outperforms defending. I think the law of total tricks provides at least some indication regarding the answer to this - we want to play if we have sufficient trumps to compete at the current level, and defend otherwise.
  
• I think your example as cited is not 'just another rumour'. Defending is very difficult, and I fully expect weaker players to habitually drop more tricks on defence than as declarer. I would not be surprised if they can score better just by bidding more aggressively - above and beyond any merits of bidding aggressively in the first place, which I think are also often underrated. In the long run though I think it is better to get more practice in as a defender. 'Scoring better by bidding more' is close to 'scoring worse by defending the normal contract', and it is not an advantage to be in this position.

[akwoo]

I think you'll want statistics particular to a player population.

My experience is that less experienced players need to bid more. For some reason it is hard to understand that -100 is a better score than -110.

[mikeh]

As with David, I’m unaware of any good data on this. Anecdotally, back when I was even more of a nerd than I am today,I did once track how often I or partner declared…in club games where the field is generally weak. Over about a year, or maybe 60-75 sessions, we played about 55% of the hands, and were consistently well above average in scores. I don’t, for a moment, see ‘grabbing dummy’ as ‘the’ way to win. You do need to know a little about how to play dummy😀

But it is definitely true that most non-experts make more mistakes on defence than on offence so sometimes the defenders give back anything declarer gives to them. So on average it probably behooves non-experts to ‘bid a little more’…but the advice has to be nuanced. My view is that most players fail to re-evaluate during competitive auctions….they underestimate the power of distribution…and the risks of having flat hands in contested auctions. So they overbid hands without shape and underbid one’s with shape.

Btw, while it’s usually (much) more fun to declare than to defend, I’m rarely unhappy to be on defence…except against expert declarers. If an expert declares against me, I expect at best an average result.

Last weekend, playing a Swiss with my main partner in a weak sectional field, one of our opps (holding my cards, as it happens) was complaining that the cards were ‘running the other way’. So I checked…we’d gone plus on defence on 15 of the 35 boards we’d played….contributing to our 98/100 VPs to that point. So bidding too much is not always the best approach.

[barmar]

Besides the differences in difficulty between declaring and defending, declaring simply offers more opportunity for higher scores. Bidding and making game scores at least 400 points. Getting that much on defense is much harder -- first, the opponents must have overbid, and you almost always have to double to get that much.

This is why duplicate bridge is not based on total points. But it's an important difference in rubber bridge.

[thepossum]

Statistically you get to play more if you defend

Sometimes the lower expectation of defence can be pleasurable as can watching partner declare

I had a strange perception that I generally score better when defending but it's not backed up by statistics
Maybe depends on the field and if they think the aim is always to declare and score games

NB Everyone knows my view is biased by online, mostly robot, bridge. And I have a prefence for real Bridge rather than Best Hand
And I am a rarity where getting a new hand and bidding was my favourite bit - no maybe defeating a contract - even better being opening lead
But since I happily move on to the next hand what does it matter

[jillybean]

 thepossum, on 2024-August-16, 15:53, said:

Statistically you get to play more if you defend

Sometimes the lower expectation of defence can be pleasurable as can watching partner declare

You need to play Obvious Shift.

[AL78]

My experience is I frequently score worse when I am on the wrong end of a hand bias and end up declaring two or three times in a 24 board club session. It is all very well saying bid more but it is very difficult to come into the auction when you keep picking up flat single digit point counts and your partner is silent. If you get opponents who just bid everything that is available to them and all you can do is pass and follow suit, you may have to bust a gut to get to 50% over the round, and even that might be impossible depending on what the rest of the field do.

A recent example, a three table movement 25 board Howell five board rounds with an inexperienced partner, first round against the weakest pair. Board 1, opps bid to a cold 4♠, 0% for us when the other two pairs fail to find game. Board 2 and 4, opps bid to game, both 50% boards. Board 5, opps bid to game and go off, 100% to us. Board 3, partner opens 4♥ on a four loser monster and we miss slam, a 50% board when no-one else finds it. Opponent's bidding to these games is poor (supporting at the three level with three when partner has only shown a four card suit instead of making a clear TOX, reversing on a fragmented semi-balanced 14 count on which they could and probably should have opened a weak NT) but they weren't punished for it. If you cannot hammer the weak pairs then forget about a decent finish, they will be throwing MPs at everyone else. They finished one MP ahead of us at the end. Over the full session: final score 50%, my average HCP 8.8, declared three times, defended 16 times. A good example why I have fallen out of love with bridge.

[AL78]

barmar, on 2024-August-15, 15:25, said:

Besides the differences in difficulty between declaring and defending, declaring simply offers more opportunity for higher scores. Bidding and making game scores at least 400 points. Getting that much on defense is much harder -- first, the opponents must have overbid, and you almost always have to double to get that much.

This is why duplicate bridge is not based on total points. But it's an important difference in rubber bridge.

At duplicate it is not the absolute value of the score you get that matters, it is the score relative to all the pairs sitting your way on that board, so being able to get higher scores as declarer is irrelevant. You can theoretically get a 100% board scoring -2000.

[helene_t]

It would be very easy for someone with access to data to answer this question.

FWIW, I played 218 hands in the last month on BBO, not including declare-only tourneys which are obviously not relevant here. Two boards were passed out, I will ignore those. The 216 hands break down as:

MP declare: 80 hands, average 54.3%
MP defend: 77 hands, average 42.4%
IMP declare: 36 hands, average +0.33 IMP
IMP defend: 23 hands, average +0.99 IMP

For a serious analyse one might want to throw out best-hand boards and maybe all robot boards, or just analyse them seperately. I might try to see if I can do something with BridgeBrowser do get a larger data set. But there is also the vugraph archieve and many club sites are possible to scrape for data.

It might also be interesting to distinguish between defend undoubled and defend doubled.

[paulg]

During lockdown and before bridge really got back to f2f, I played on BBO a lot and kept all the records. My regular partnership played 2,939 hands. These were all IMP events and mostly against competent opponents (Flight A regulars). The events in the file will be 8-, 12- or 16-board sessions/matches.

There were four pass outs.

I declared 687 hands at +1.07 IMP/board (real score)
My partner declared 769 hands (0.48)
LHO declared 722 hands (we scored 0.60 IMP/board)
RHO declared 757 (0.49)

These figures are interesting as we are an aggressive pair but clearly don't win the contract as often as you might think.

I opened the bidding on 788 hands, my partner on 803, LHO on 639 and RHO on 705.

[helene_t]

I just downloaded my pianola records. They are mostly from club evenings in Leeds and Harrogate, April 2011- July 2015. All matchpoints.

I had a Q rating during most of this time. I played mostly with the same partner. So in a sense it's a sample size of 1

It's 48 sessions so a good 1000 boards. We averaged 52.99% when declaring and 50.93% when defending. The standard error on the difference based on session means is about 1.3% so it is not statistically significant.

[helene_t]

Something I think would be very cool would be to make a tool that could identify "errors" in human and robot play, i.e. "failure to find sacrifice", "bidding non-makeable slams", "failure to lead trumps" etc etc. Where an "error" is defined as any action that reduces the maximum achievable score, assuming that all players bid and play DD afterwards.

Sorry if this is a bit off-topic, it's just that this thread made me think that if we are going to make some statistics to see if people systematically compete too little, we could extend it to look at more specific types of mistakes.

[helene_t]

Assuming an SD of 5.5 IMPs as reported by Gordon Bower in https://bridgewinner...d-2-sgh67m0fze/

the SE of the effect size in Paul's data is about 0.2 IMPs/board. Given that his partnership declarer advantage is 0.23 it is not statistically significant.

If more people will supply data, I could make a pooled analysis. But of course it's best if someone with access to a large number of hands could answer Jillybean's question.

Note: An SD of 5.5 IMPs is for high-level teams matches. It agrees with what Hrothgar wrote somewhere on this forum that in his experience it's about 5 (but I'm not sure if Hrothgar's data are matches or XIMP). In a large XIMP field, a back-of-the-envelop argument would lead to an SD of 5.5 / sqrt(2) but of course XIMP events can be more heterogenous than matches which would probably increase the SD.

[helene_t]

It suddenly occurred to me that the extra IMPs or MPs gained by declaring don't fully capture the award for aggresive competitive bidding. You also gain when you push opps up to a non-makeable level. This latter issue may be more difficult to get data for. Even if we had data including the auction, such as from BBO sources, there could still be ambiguity:

1♥-(1♠)-2♥-(2♠)
3♥-a.p.

3♥ may have been taking the push, or it may have been a game try which they would have made without the 2♠ bid also.

[awm]

helene_t, on 2024-August-18, 10:51, said:

It suddenly occurred to me that the extra IMPs or MPs gained by declaring don't fully capture the award for aggresive competitive bidding. You also gain when you push opps up to a non-makeable level. This latter issue may be more difficult to get data for. Even if we had data including the auction, such as from BBO sources, there could still be ambiguity:

1♥-(1♠)-2♥-(2♠)
3♥-a.p.

3♥ may have been taking the push, or it may have been a game try which they would have made without the 2♠ bid also.

Against that, you might help them in the play since your bidding gives them some distributional information. Or you might push them into a making game that they otherwise wouldn't have bid. There are really a lot of factors here.

[pescetom]

helene_t, on 2024-August-18, 10:51, said:

It suddenly occurred to me that the extra IMPs or MPs gained by declaring don't fully capture the award for aggresive competitive bidding. You also gain when you push opps up to a non-makeable level. This latter issue may be more difficult to get data for. Even if we had data including the auction, such as from BBO sources, there could still be ambiguity:

1♥-(1♠)-2♥-(2♠)
3♥-a.p.

3♥ may have been taking the push, or it may have been a game try which they would have made without the 2♠ bid also.

Agree with your intuition, dubious about your example.
Without the 2♠ bid, 3♥ would have been a game try with poor hearts, but here it is taking the push: a game try would be 3♣/♦, or at least for us.

[helene_t]

I almost finished writing the script that scrabes the vugraph archieve, I expect it to give about 100,000 boards on which the same team declared at both tables (those that were passed out at one table are of course also of interest). It will take a few days to run, though.

If anyone here would like access to the data, just tell me what you want extracted

Also, maybe Barry or Diana could advice me of what I am (not) allowed to share? Can I make a publicly available zip file or should I limit it to those who show interest here?

[jillybean]

I don’t have any use of the data but would love to see the Declare vs. Defend numbers.
Thanks for doing this, Helene.

[pilowsky]

Maybe BBO can run a few hundred hands with robots at all seats MP and IMPS.