Risk Taking and Gender: The Case of Bridge

[tenoblak]

Hello

 

Here is a small research that I did, that speaks for itself, and I thought it might be of interest to the membership of BBO. I am very interested in your reaction.

 

Risk Taking and Gender

 

I look forward to hearing from you.

 

Thank you,

Dr. Arnon Perry

[Antrax]

Hey,

 

You have some minor English errors (raw instead of row, "full game" is in Hebrew but they just call it "game" in English, etc).

More importantly, when you select boards at random or rounds at random, you're opening yourself to accusations of cherry picking. Why not perform a full analysis of the results?

Finally, why not try and correlate the level of the player (roughly correlated to their number of matchpoints, perhaps weighed by length of membership) to eliminate that factor?

[Vampyr]

Exhausted instead of exhaustive, researches instead of researchers, and so-called requires a hyphen. Also numbering the contracts from 1 to 35 is an odd thing to do, and your scale from 18-35 would seem to include two non-game four-of-a-minor contracts.

 

But anyway the level of the contract is not always the full picture. Perhaps a pair bid a sacrifice, or a phantom. Perhaps a pair's methods allowed them to locate extra values, so they explored for slam but stopped short of it. Perhaps a pair's methods didn't allow them to stop short of a poor slam. And state-of-the-match considerations cannot be ignored.

 

Also, though it pains me to say it, the skill level of the Open field and the Women's in a world championship is not equal.

[barmar]

Also, though it pains me to say it, the skill level of the Open field and the Women's in a world championship is not equal.

That's not necessarily a criticism of women players. It's likely that the best women players will choose (or be selected) to play in the open events rather than the women's events. This dilutes the strength of the field in the women's events.

[microcap]

Just in case you didn't know, they have done studies in the investment management world [my field] that show that in fact, women DO take less risk in investing than men do.

 

However, because of this, women are actually BETTER portfolio managers--the men take stupid risks that over time hurt their results.

 

Women are ridiculously underrepresented in the field of banking and investing, we would be better off with more of them....

[Cromlyn]

I would imagine that women take just as much risk as men do where there are no safety or money issues at stake

[Maurice]

By eliminating doubled and redoubled contracts you remove the most significant risk factor!!

 

For instance, doubling a 2d bid and leaving it in for penalty is a far greater risk than NOT doubling a 2d contract and letting it play.

 

For me, that makes the study next to useless.

 

Also, I do not think the bidding is the biggest risk - it is in the play of the game. Will you risk going down in solid contract to take that extra trick in match points? Will you double those "interfering ops" in a part score or take your solid part score? (The "youngsters" in the recent USBF championships showed how risky this behaviour is, and of course how rewarding.)

 

I'm not even sure that risk taking is a meaningful thing - what counts more often than not, is getting to the right contract, and not booting the easy ones.

[xeno123]

Hello

 

Here is a small research that I did, that speaks for itself, and I thought it might be of interest to the membership of BBO. I am very interested in your reaction.

 

Dr. Arnon Perry

 

I assume you are familiar with a more detailed study on risk-taking that showed some sex-differences:

 

http://dash.harvard.edu/bitstream/handle/1/4405460/RWP10-034_Zeckhauser_alia.pdf?sequence=1&origin=publication_detail

 

One of the authors was economist Richard Zeckhauser, who was a Blue Ribbon pairs winner.

 

Peter

[jogs]

Any study of less than 10,000 boards is pretty much useless. Also the 5% significance level is two std dev(or more precisely 1.96 std dev) not just one.

[Hanoi5]

I agree with most of what has been said. The sample is too small and it hasn't even been studied in the right way. Risk is not only getting to a higher level.

[Vampyr]

That's not necessarily a criticism of women players. It's likely that the best women players will choose (or be selected) to play in the open events rather than the women's events. This dilutes the strength of the field in the women's events.

 

There is also the fact that a lot more men play bridge at all (at least competitively -- women may outnumber men when it comes to playing at home or at the golf club, but I don't know anything about that).

[SimonFa]

It's well known that girls are more likely to choose maths, science and engineering subjects in single sex schools than they are in mixed schools where they are likely to revert to stereotypes. What this study shows is that women playing against women take the same risks as men playing against men.

 

It would be far more interesting to know help if there is a similar effect at bridge and whether it is men who are showing off when playing with or against women rather than women changing their behaviour.

 

I'm not convinced that this study has added anything to the general body of knowledge in this debate.

 

 

http://www.standard.co.uk/news/girls-75-more-likely-to-take-maths-if-they-go-to-a-singlesex-private-school-8951710.htmlscience

 

Its a newspaper article for easy reading, do your own Google for scientific papers.

 

PS for a good read about women's bridge and risk taking try this: http://www.amazon.co.uk/Love-This-Game-Sabine-Auken/dp/1897106068

[brettnj]

Although I found the results interesting, I consider them extremely speculative. One must first

take into consideration the players'level: i.e., top male professionals compared to female pros,

and so forth at various levels of play. The results could be significantly skewed at lower levels.

In my experience, at the highest levels, male professionals will be less risk-averse than female pros.

However, at lower levels, my experience is that female players tend to be more aggressive.

 

Brett Kunin (West Orange, NJ)

[Zelandakh]

I think you misunderstand risk at bridge. To a large extent it is often riskier not to bid game than to bid it. How a game or slam is bid can often represent different levels of risk, even when the contract is identical. In addition, 8 boards is a laughably small sample. In other words, I think your research is scientifically worthless and I would certainly reject it in peer review. It might just fly as a Final Year thesis if fleshed out with an analysis the rest of the boards for these events. Sorry to be blunt but I am a mathematician and it pains me to see such an incomplete paper presented as a basis for discussion.

[helene_t]

As a statistician who has done lots of peer reviews of applied statistics papers I will have to agree with Zel that this study is very far from meeting the requirements of a scientific journal.

 

Obviously you would need a much larger sample size, and you shouldn't present your findings as saying anything about risk taking. If you measure the levels of the contracts then just present it as such. Speculating about whether this says anything about risk taking or not is not very interesting IMO.

 

You might want to consider whether you really need paired data. Scrabing hundreds of thousands of hands from club and tournament records and classifying them by first names that could be recognised as male or female will not only give you a larger data set but also allow you to investigate whether there are gender differences that are general as opposed to specific for specific levels or countries.

 

If you really want to address the issue of risk taking you will have to think carefully about how that can be measured. It is not easy. It might be impossible. It might be possible if based on a combination of player interviews and statistical analysis. A sacrifice is often a risk-averse strategy but players at less than expert level may not recognise it as such. Some may feel they are being risk averse when they decide to defend undoubled, even if they really made a high-risk choice.

[barmar]

I think you misunderstand risk at bridge. To a large extent it is often riskier not to bid game than to bid it.

If you're thinking totally rationally, that may be true. But human psychology is not rational. We worry more about taking a loss than missing a gain, so risk aversion will usually cause you to bid lower (to avoid going down) than higher (to avoid missing the game bonus).

 

Game players may be better at overcoming this in the game than in real life, but I'm sure it still biases them. One of the aspects of becoming an expert player probably includes better control of this.

[Zelandakh]

If I score +170 against +620 then I take a bigger loss than -100 against +140. We can have a good discussion about the psychology of how we view a loss here, and it is true that some very weak players consider any plus score a gain and any minus score a loss. But the data points in play are experts for which that is certainly not the case. There are also such classic cases as taking out insurance or bidding game as a two-way shot. Another classic that you see at the local club but not so much by good players is the 1♥ - (1♠) - 2♥ - (2♠); 3♥ - (3♠) - 4♥ - (4♠) auction type. I consider this kind of auction much more risky than a simple 1♥ - (1♠) - 4♥ or whatever but this analysis suggests the former is less risky as it leaves open the possibility of stopping lower. Similarly, 3NT can easily be a huge risk (in terms of making the contract) versus 4M or 5m but worth it for the potential MP upside; on the flipside not bidding 3NT when you know the field will mostly be in it is also a risk. Much of good bidding judgement in bridge is weighing these risks with imperfect information. And quite a lot of the time the heavier risks come from potentially underbidding than potentially overbidding. And this is (just one of the reasons) why the simple 1-35 scale is simply a poor proxy for level of risk.

[diana_eva]

I think you misunderstand risk at bridge. To a large extent it is often riskier not to bid game than to bid it. How a game or slam is bid can often represent different levels of risk, even when the contract is identical. In addition, 8 boards is a laughably small sample. In other words, I think your research is scientifically worthless and I would certainly reject it in peer review. It might just fly as a Final Year thesis if fleshed out with an analysis the rest of the boards for these events. Sorry to be blunt but I am a mathematician and it pains me to see such an incomplete paper presented as a basis for discussion.

 

Actually I encouraged Arnon to post here precisely because I felt the forums are a great place to get constructive replies. It is just a v small attempt to evaluate risk taking, and all the comments should be a good starting point for a "serious" research if anyone feels that would be interesting.

 

I think it is a good discussion and IMO it's an interesting subject.

[gszes]

In order to "study" anything (like bridge for example:)) the person performing the study should have

 

a much better than minimal understanding of how the game works. The author apparently has no clue

 

about game/slam bonuses and how they affect bidding. Basic scoring is also ignored since diamonds

 

and clubs are not treated equally and hearts and spades are not treated equally though both pairs

 

score exactly the same. A team/pair that quickly realizes their side has little/no chance for game will try to

 

stop the bidding as low as possible and appear to be underacheivers when their expertise actually

 

makes them more likely to succeed than those pairs that bid higher with the same values.

 

 

 

 

The bidding is probably the single most significant factor in determining risk behavior and it is completely

 

ignored in favor of seeing only a finished product. For example a pair bid to say 2h and were willing to play

 

there when their opponent in the pass out seat reopens the bidding and the pair somehow winds up in a

 

making game that everyone else bid easily. The pair that was willing to play 2h was acutely risk averse but

 

would appear to be risk neutral when peeing solely at the final bidding and not the entire auction.

 

 

 

 

I am not a huge fan of these types of studies because they make the erroneous attempt to try and apply

 

the statistics of the whole to individuals and generally using numbers like a "statistically significant" 5%.

 

Statistics can indeed be useful but there is no substitute for studying the work of an individual and

 

deciding if that person is risk averse or not. The study also has to make sense which the basic premise

 

of this study fails to do. I know this last paragraph should have come first but I felt it might be more

 

useful to the person conducting the study if they read why this particular study is worthless and apply the

 

lessons to future endeavors. I sure hope this was not a government sponsored study sighhhhhhhhhh.

 

 

 

 

 

 

 

 

[Vampyr]

The bidding is probably the single most significant factor in determining risk behavior and it is completely

ignored in favor of seeing only a finished product.

 

But maybe the play could yield more fruitful analysis since an exact risk/reward ratio can be determined. Comparing how often different groups will try for an overtrick with percentage O when this line of play offers D odds of going down, as opposed to taking the 100% line to make exactly seems to involve far fewer variables plus the chance of quantifying the risk involved.

[monikrazy]

Interesting concept for a study.

 

Unfortunately, the way the study defines risk isn't accurate: "Our general assumption is that the higher the winning contract the greater the risk of failure."

 

Risk would be better defined by the chance a partnership fails to maximize its score with a given set of actions.

[barmar]

Interesting concept for a study.

 

Unfortunately, the way the study defines risk isn't accurate: "Our general assumption is that the higher the winning contract the greater the risk of failure."

 

Risk would be better defined by the chance a partnership fails to maximize its score with a given set of actions.

Risk should be defined relative to the information that has been exchanged during the auction.

 

Making a major or NT game usually requires about 25 combined points (HCP + distribution). If the auction has only shown that the partnership has at least 20 points, and they bid game anyway, they're taking a big risk. But if it has shown at least 28 points, then making game is very likely, and NOT bidding game is taking a big risk.

[barmar]

Also, risk can be taken during the auction. If you invite game when you only have a minimum, you're taking a risk that partner has nothing extra and you'll get too high. Conversely, if you have extras and fail to invite, you're risking missing a game if partner has a hand that would accept the game try.

[Vampyr]

Sorry barmar, deleted the post you replied to because I thought it didn't add much to what Zel's said above.

[barmar]

Is 7♥ harder to make than 1 ♦? Well, there are thousands of hand combinations that can make the former but not the latter.

True, but effectively irrelevant. No competent players, with hands that can make the former, is going to bid the latter, no matter how risk-averse they are. So you can ignore that possibiilty.

 

The 3NT versus 4/5minor example is more on point.

 

These are all just variations on the theme I described: You have to compare what they did bid against what they should bid given the information they have.