Not signed in (Sign In)

Vanilla 1.1.9 is a product of Lussumo. More Information: Documentation, Community Support.

Is this question acceptable?: Mathematics of gun control

Bottom of Page

1 to 46 of 46

  1.  
  1.  
    There are already three votes to reopen the following question : http://mathoverflow.net/questions/51656/mathematics-of-gun-control-closed

    I'm rather surprised -- I thought that this was pretty straightforward (it's not clear to me what the specifically mathematical question is). I'm also not comfortable using MO as a tool for political campaigning, no matter what the issue is. On other forums, that tends to strongly corollate with a decreased signal to noise ratio.

    What does everyone else think?
    • CommentAuthorYemon Choi
    • CommentTimeJan 10th 2011
     

    I agree with Andy. It seems to me that the question presupposes many things which are outside the purview of a purely statistical analysis, and while I always feel I should catch up with 2000 years of political philosophy I'm not sure it could or should be done on MO.

    • CommentAuthorWillieWong
    • CommentTimeJan 10th 2011 edited
     

    I would vote to close it if I see it open. It has relatively little to do with mathematics, and its connection to the profession is tangential at best. As far as I can tell the "statistical" part of the question is actually asking about raw data, and not statistical analysis. That falls in the realm of social or political science, not mathematics.

    • CommentAuthorYemon Choi
    • CommentTimeJan 10th 2011 edited
     

    To recap: the first sentence of the original post is

    What is the most convincing and comprehensible statistical/mathematical argument for strict gun-control laws?

    Right off the bat: where? when? with respect to what social, moral or political aims or norms? Without these clarifications the question isn't even well-posed for a politics or philosophy discussion, let alone a statistical one.

    • CommentAuthorHJRW
    • CommentTimeJan 10th 2011
     

    I'm firmly in agreement with everyone here, and will certainly vote to close if the question is reopened.

  6.  

    I really don't think this question has much chance of generating anything but heat and light. I'm happy to lock it, but will wait a bit in case anyone wants to shout me down.

  7.  
    I support locking it. I'd vote to close if it were ever opened.
  8.  
    Please lock the question. I'd vote to close if reopened.
    • CommentAuthorgrp
    • CommentTimeJan 10th 2011
     
    Scott, I am willing to shout you down.

    I am not in favor of having the question open. I think however, that we have not heard sufficient from those who wish the question to be reopened. Also, it is hard for me to see why preventing this question from being reopened is so desired. Locking the question down now seems to me to stifle a voice that needs to speak, and will be perceived by many as (Mathoverflow subcommunity-) politically motivated. Indeed, the question as stands it seems inappropriate to me, but some time should be given to allow opportunity for improvement and discussion. I feel not enough time has passed. Also, the question can be shut down again by a moderator if it has not significantly improved.

    Gerhard "Ask Me About System Design" Paseman, 2011.01.10
  10.  
    @grp : I don't think that it is fair to interpret closing/locking this question as "politically motivated". I suspect that at least some of the people who object to it agree with its political sentiment (for instance, I mostly do), but think that it is unambiguously inappropriate for MO.

    In fact, wondering whether there are political motivations behind closing it is precisely why it should be closed and locked. I think it is best for MO to remain strictly apolitical, both for practical reasons (I don't want anyone to feel unwelcome to post here!) and for philosophical reasons (I think math is far too important to be politicized!).
    • CommentAuthorgrp
    • CommentTimeJan 10th 2011
     
    Andy, I don't think such an interpretation is fair either. I still think it will be so interpreted. In any case, I am advocating for a little more time to hear from those who wish the question reopened. I feel some part of the community is rushing to stifle something, rather than just having faith in the process that most closed questions endure.

    Gerhard Paseman, 2011.01.10
    • CommentAuthorYemon Choi
    • CommentTimeJan 10th 2011
     

    I have some sympathy with Gerhard's point of view - or rather, the point of view which he is drawing attention to. On the other hand, on a few questions I remember seeing recently, people who vote to re-open didn't take up the invitation to argue their case on the corresponding meta thread. (This is one example I have in mind.) I still think the original question is either badly worded or begging the question, and that's independent of my own views on firearms policies.

  13.  

    I agree with Gerhard. I don't support the question, but I don't think it should be locked.

    What benefit comes from locking the question? Lots of people here have said that they'd vote to close if the question is reopened and have given arguments for why the question doesn't belong on MO. We haven't seen much in the way of argument for keeping the question open. So there isn't much worry that the question will remain open.

    What ill comes from the question being locked? Aside from the problem that the locking may appear politically motivated, I think it undermines the social mechanics of MO. Three reputable members of the community have voted to reopen. If some others want to vote to reopen, I think they should. The question would be open for a bit, but we'd get a benefit out of it: the names of the people who voted to reopen would be publicly posted. We'd be able to say to them, "Hey X, please explain why you voted to reopen this." We may get a good explanation. If we get a bad explanation (or no explanation), the reopeners will (hopefully) feel that voting to close/reopen is something they should do more thoughtfully. I think imposing this kind of cost for thoughtless or unjustified votes is much better than having a moderator lock the post. It cultivates the sort of behavior that keeps the quality of MO high.

    I can think of two main reasons to lock a post: (1) there's an edit war or repeated edits, gaming the fact that edits bump the question, and (2) the comment thread is getting out of hand. Neither of these is the case here (yet?). I'm unlocking the question now, but if you have a good argument for locking it, feel free to convince me (or another moderator) to lock it again.

  14.  
    Hello. The question was intentionally provocative, I admit. I appreciate the discussion here on meta. I will not argue here for reopening the question, but rather simply set a context for discussion. Two issues frame the debate, in my view.

    First, MO is a sociological entity, the largest forum for discussing mathematics. It cannot divorce itself from its existence in the real world (the way math can). It has a narrow *intended* purpose, but as the center of mathematical discussion around the world, there is also power in MO beyond its intended use. Super-users may not have invited that power and may not wish to wield it, but it is there. The meta-question of how to deal with an MO question with real-world/political implications should be confronted.

    Second, there are certain issues, usually related to morality, which some people believe trump the ideal (?) of the insulated academic. The best examples may be in letters among academicians in and around WWII and surrounding the issue of nuclear energy. (Back then, the power surrounding academic discourse was perhaps more concentrated.) Personally, I'll admit, I have been struggling with this question as it relates to the prevalence of gun violence in the US (including on university campuses).

    I think that writing this within the body of the question is unnecessary. The context is the context. Discussions about the context are appropriate to meta.

    As for the question iteself, it was too short and vague, just a stab in the dark (excuse the metaphor). I should perhaps have amassed all the evidence that I wanted analyzed ahead of time and done some regressions just for fun. As some pointed out, that is a job which would take a lot of work, whereas the answer would be more immediate to an expert (getting such answers quickly is part of the beauty of MO). I found the following article in the Journal of Quantitative Criminology (http://www.springerlink.com/content/h1r258q79648h836/), but I also found other articles and books which seemed to be written by cranks. I am not expert enough to give an assessment of these articles or their conclusions, and I don't know the academic discipline of quantitative criminology well enough to know who is out there right now, studying what (just as someone new to symplectic topology might not know what is happening right now in symplectic Khovanov homology).

    I disclosed my intended use for the response, which, while not part of the practice of mathematics, is part of one attempt to uphold the responsibility of the profession of mathematics. The question was too difficult and big for me to answer, I admit, and I hoped to tap the energy, expertise and interest of the MO community.
  15.  
    @Eric I personally agree that we need gun control. If I had the power I would outlaw the sale of guns entirely right now - even guns used for hunting, sport shooting, etc. My neighborHOOD has a big problem with gun violence, and I know that if guns were harder to obtain a lot of that would disappear.

    Your question is not a mathematical one though. Maybe if you asked for statistics that strict gun control laws lower crime it would be closer to a mathematical question. I think a lot of people who oppose gun control laws do so because they generally oppose the government limiting their personal freedom. No amount of statistics will convince such a person that gun control laws are good - it is a matter of principle. Your question seems to imply that any reasonable person should be convinced by some statistical argument, which I think is potentially offensive.

    So maybe a question asking for a statistical analysis of crime reduction correlated with strict gun control laws is what you are after. I don't think this question would be appropriate for MO either, because it could easily be someones master's or Ph.D. thesis. I see no reason why you should expect a snappy MO answer - all you could get was a reference. I don't even think MO is a great place for making such a reference request.
    • CommentAuthorAndy Putman
    • CommentTimeJan 11th 2011 edited
     
    @Eric : I don't think we're arguing whether or not mathematicians (or other academics) should be involved in politics or whether or what role that involvement should play in their professional lives.

    The point is that MO is intended for technical mathematical discussion. My feeling (and I think this is widely held, though we haven't had that many discussions of it on meta) is that we welcome questions about mathematical statistics, but not about "applied statistics" in the sense of what statistical evidence exists about specific problems. I don't think those kinds of discussions make sense here -- the problems are not really mathematical, but philosophical and sociological. For all the major issues (gun control, abortion, war, etc), each side has any number of studies. While some of those studies can be critiqued on technical grounds (eg bias in the design of their experiments), in the end mathematicians AS MATHEMATICIANS have only a small role to play. This is not to say that mathematicians should not be involved -- it's just that when they are involved, they are not really doing mathematics in these sense of proving theorems, etc.
  17.  

    I disclosed my intended use for the response, which, while not part of the practice of mathematics, is part of one attempt to uphold the responsibility of the profession of mathematics.

    This is all well and good, but I strongly disagree that any such attempts should take place on MO for reasons that have already been stated by others. As with many things, this is a topic much better suited to a blog.

  18.  

    I agree that the locking of the question seems premature.

    But here is one point: the question currently has four votes to reopen. We have already had a somewhat extensive discussion here, but I can't see anyone except Prof. Zaslow who has clearly supported the question, and site mechanics imply that none of the reopen votes can be his. So it seems that no one who has voted to reopen has explained themselves here. Even when the question finally gets reopened (as is starting to seem likely), the site mechanics are such that those who voted to reopen will not be listed. On the other hand, when the question is closed again (as is inevitable from what people have said here), five people will go on record as being, somehow, against the "mathematics of gun control". This seems a little unfortunate.

    Regarding the necessity of bringing social and political issues to MO: I certainly respect the sentiment behind it, but I don't agree that it is necessary to do so. MO is not the only major internet forum for mathematics. There is also nLab, sci.math, sci.math.research, math.stackexchange.com, the mathjobs wiki, and so forth. MO has become popular relatively quickly in large part because of a well-chosen and well-adhered to specific purpose: our concentration on research level questions and answers is especially appealing to research mathematicians. Deviating from that goal has the potential to undermine the site.

    Moreover, arguing that MO must be open to certain kinds of questions and discussions by virtue of its success and popularity strikes me as very strange. Anton, Scott and a few other young mathematicians had a vision of a website with a specific purpose, and they built it. Anton in particular has put a phenomenal amount of work into this project. If someone else wants a website with a different worthy purpose -- say, "mathematics and social responsibility" -- go for it, build it, sign me up as a member. But don't piggyback onto this website: that seems neither fair nor productive.

  19.  
    @Andy: You are right that the intellectual merit of the question is one of applied statistics, and if that has been deemed outside the purview of MO then it is not appropriate on those grounds. You are also right that mathematicians qua mathematicians have only a small role to play in political discourse, and when they are involved (as mathematicians) they are not really doing mathematics but engaging in the business of the profession of mathematics. This, as you point out, is outside the *intended* usage of MO. I concede this as well.

    Suppose there were a political movement to put our national treasury up for grabs (against, say the EU) with a spin of a roulette wheel: spin 17 and we win theirs, anything else and they win ours. American mathematicians would have a professional, societal interest in organizing a comprehensible statement proving the expected futility of this arrangement. How would they arrange for this? What if some tried to construct this argument collectively on MO? (And would the same question, but without societal implications, be treated differently? Should it be?) Is the current situation of gun violence analogous, or have I simply constructed a glib example? I don't know. As I confessed, I am struggling with this question myself.
    • CommentAuthorEmerton
    • CommentTimeJan 11th 2011
     

    Dear All,

    I voted to reopen the question. I think between the question and the comments from Eric, the intent was made clear enough, and Eric's example of the Freakonomics analysis of abortion laws suggets to me that people are dismissing the applicability of mathematics and statistics to the analysis of gun control laws too reflexively. It may be that the question won't get answers from the current MO readership, but that (to my mind) has never been a ground for closing questions.

    Regards,

    Matthew

  21.  

    American mathematicians would have a professional, societal interest in organizing a comprehensible statement proving the expected futility of this arrangement. How would they arrange for this?

    On blogs. It would be more than enough, for example, to get Terence Tao to mention the issue on his blog. I disagree that the current situation is analogous because arguing against a new policy, especially one for which you have not provided a convincing reason, is much easier than arguing against an existing policy with well-established supporters.

    Eric's example of the Freakonomics analysis of abortion laws suggets to me that people are dismissing the applicability of mathematics and statistics to the analysis of gun control laws too reflexively.

    I am not in any way dismissing the applicability of mathematics and statistics to the analysis of gun control laws. I still disagree that that analysis should take place on MO. I would much prefer that the question be reworded to fall more clearly under the purview of MO, e.g. "what are good references for existing literature on this subject" or something similar.

  22.  

    Dear Matt,

    Thanks very much for coming forward and explaining yourself.

  23.  
    Hi Matthew. I agree with Qiaochu on this. There's nothing wrong with discussing details of specific mathematical models here, which appears to be what you're talking about. I don't interpret the original question as having that shape. I interpret the original question as asking for a rationalization of a pre-determined outcome, in particular one that supports a particular political orthodoxy. IMO this isn't appropriate for MO.
  24.  
    Dear Matt,

    I also want to thank you for your explanation.

    Let me say a little more about why I am opposed to having this question open. My usual standard for deciding whether or not to leave a question open is "Is there some professional mathematician who is interested in this"? Of course, I am violating that standard here, so I should probably explain myself. One feature of MO that distinguishes it from many other places on the internet is its high standard of civility and professionalism. I fear that if this question is left open, then we risk decreasing this level of civility. The internet has a bad track record for polite discussion of explosive issues like gun control.

    In other words, not only would there be a question here that is not on-topic (which I don't view as that big of a deal), but there is a risk of damaging what I view as important core values of this site.

    I think that one difficulty here is that there is a tendency for us mathematicians to assume that if someone disagrees with us, then that person must be missing some piece of factual information (or be wicked, but let's put that aside). However, this is not true -- there are intelligent people of good will on both sides of all the major political issues I can think of, and both sides have plenty of studies supporting their view of the world and are well aware of the evidence of the other side (this maybe isn't true of the "man on the street", but that's not who we are really talking about here). To resolve this, one must not discuss "naked facts", but values and moral philosophy. This is hard work, and I don't think that MO is well-suited to it.
  25.  
    I wouldn't be comfortable with MO supporting any sort of political or sociological viewpoint at all, no matter how non-controversial (and this one is controversial). If it reopens, I'm voting to close.
    Personally, I would very much prefer a world in which people were to reflexively dismiss the applicability of mathematics to any and all political and sociological questions. The converse- (mis)use of mathematics to forward a political agenda- seems far more dangerous (and widespread).
  26.  

    I think that one difficulty here is that there is a tendency for us mathematicians to assume that if someone disagrees with us, then that person must be missing some piece of factual information (or be wicked, but let's put that aside).

    Especially if that mathematician is aware of Aumann's agreement theorem...

    • CommentAuthorgrp
    • CommentTimeJan 11th 2011
     
    I have heard in this thread from some who want the question reopened. While I agree that certain social issues should be considered and debated by the academic community, and the mathematical community in particular, I think that MathOverflow is not the forum for such a debate. I also agree that the treatment of the question seems (to me) to be partly a knee-jerk reaction. More time has passed, and perhaps another day is warranted. After that day, if no essential change has occurred, it seems to me reasonable that the question remain
    closed and possibly be locked. Here is why:

    The question as currently phrased makes an assumption which I think is untenable. The stated motivation for the question is to create a document which would receive a certain consensual agreement from the MathOverflow community, and presumably would receive favor from the mathematics and possibly also the scientific and academic communities. I suspect there is enough diversity in the MathOverflow readership alone to make it unlikely that such a document exists. At best, one can have a rationalization, with some basis in
    reality, for people to argue. The consensus that is required is that of society as a whole; such a document will help a small part of society, but I suspect it will not be a new item in the gun control debate. Similarly, a statistical argument on a social issue will not be decided by mathematicians on a purely mathematical basis; feelings and other items which are not mathematical will determine whether one signs it. I am not seeing that
    the question has an answer.

    The commentary on the question itself is appropriately restrained. However, if this becomes an open-close tug-of-war, then locking the question makes sense to me. I will suggest this thread be closed if no more light is forthcoming.

    Here are questions that I think would be appropriate for MathOverflow.

    "I am crafting a document in favor of stricter gun control laws. I want to include some statistical analysis in support of this. I am aware of references A , B and C, but need to supplement them. Can you tell me of more recent references?"

    "I am looking at applying statistical argument to public policy, specifically in the area of gun control. I have two arguments and would like to know which seems more compelling. In summary A. Alternatively B. Which would be more persuasive (in the sense of easy to convey to both mathematicians and nonmathematicians)? Also
    is there yet another argument C that is even more convincing (and less technical, so more accessible to a wider audience)?"

    "I am applying the following statistical argument towards stricter gun control laws: D . Are there any weaknesses in this argument, or ways to improve it? I ask primarily for a mathematical analysis, but other criticisms are also welcome."

    The primary benefit these questions have over the original is they are more specific, and they make no assumptions about the politicial make up of the MathOverflow community. They (to me) seem more amenable to reasoned discussion as well. In particular, they try to use MathOverflow for what it is best at: giving experts a chance to answer quickly specific questions in their specialized field. These questions might also be closed, but I think with some helpful response, if only in comment form.

    Gerhard "Ask Me About System Design" Paseman, 2011.01.11
  28.  

    Pete L. Clark:
    [T]he question currently has four votes to reopen. We have already had a somewhat extensive discussion here, but I can't see anyone except Prof. Zaslow who has clearly supported the question, and site mechanics imply that none of the reopen votes can be his. So it seems that no one who has voted to reopen has explained themselves here. Even when the question finally gets reopened (as is starting to seem likely), the site mechanics are such that those who voted to reopen will not be listed. On the other hand, when the question is closed again (as is inevitable from what people have said here), five people will go on record as being, somehow, against the "mathematics of gun control". This seems a little unfortunate.

    At 250 reputation, you can vote to close/reopen your own questions, so Eric can vote to reopen. When (if) the question is reopened, the people who reopened it are listed in the revision history (see this example). When the question is reclosed (as it probably would be), five people will willingly go on record as saying that the question is off topic at MO--a statement they believe--just as five people have already done. I don't feel this is unfortunate.

    Regarding the necessity of bringing social and political issues to MO, I can imagine (not really, but in principle) a social or political situation so dire that mathematicians have a clear moral obligation to take a united stand and using MO to do it is essential (for some reason, blogs and other more appropriate forums are inoperable or insufficient). In this far-fetched situation, a population of thoughtful voters would end up keeping the question open.

    Regarding the specific question, I think it's off topic on MO, and it seems to me that mathematicians are not in a privileged position to see how to solve the problem of gun violence. Gun proponents claim (I think) that less gun control can actually reduce gun violence. To me, this seems like somebody proposing we reduce the number of car crashes by making cars completely self-destruct upon impact, thereby encouraging careful driving. This sounds crazy to me, but I don't have any reason to believe that mathematicians are any more likely to agree with me than anybody else, so rallying under the banner of mathematics is going to be no more effective than rallying for the issue in general. Good data and analysis would certainly make for a stronger position (it's important to find out the right answer, not just find support for your gut reaction), but generating it again feels off topic on MO. Maybe stats.stackexchange.com (or the forthcoming economics SE site) would be better.

  29.  

    Daniel Moskovich: I wouldn't be comfortable with MO supporting any sort of political or sociological viewpoint at all, no matter how non-controversial (and this one is controversial).

    I agree that MO should mostly be neutral on such issues, but it's impossible to avoid effectively supporting viewpoints the community holds. By it's very existence, MO supports a sociological viewpoint that collaboration and transparency are good. There's nothing wrong with that.

    if this becomes an open-close tug-of-war, then locking the question makes sense to me. I will suggest this thread be closed if no more light is forthcoming.

    You can only vote to close (or reopen) once on any given question, so even then I don't see the benefit of locking the question. I agree that a high-heat/low-light discussion is worth closing.

  30.  

    @Anton: after more than a year of active membership, I thought I finally had mastered the site mechanics. Your response showed me two places in which my understanding was incomplete. (It now seems unlikely that these are the last two....) Thanks.

  31.  
    Update: I had a meeting with a professor of statistics, who explained what is typically done in, e.g., applying data analysis to medicine, other aspects of law, issues of workplace discrimination, etc. The analyses all seemed case-by-case. I said, "isn't this the universal problem in the subject? Are there good, readable treatments?" None came to (his) mind. Causation, of course, is what a rigorous researcher would like to establish. He did, however, forward me a link to the following publically-available program by a professor of politics at Princeton: http://imai.princeton.edu/research/mediationP.html.

    Also, though I did vote to re-open (as you might see for yourselves), I did not create any voting avatars of myself, in case you wondered. If the problem does re-open, I can see from the comments it will be permanently closed shortly thereafter.
  32.  

    I think that one difficulty here is that there is a tendency for us mathematicians to assume that if someone disagrees with us, then that person must be missing some piece of factual information (or be wicked, but let's put that aside).

    Especially if that mathematician is aware of Aumann's agreement theorem...

    I only learned about Aumann's theorem a few days ago, so I may have misunderstood. I thought the point is that even if the agents do not have the same factual information available to them, they still have to agree. They only need to have each other's opinions available.

  33.  

    I may have misunderstood as well. I'm taking "common priors" as a stand-in for "same factual information," but it may not actually be.

    • CommentAuthorgilkalai
    • CommentTimeJan 12th 2011
     
    The question does not seem to be appropriate for mathoverflow. A related question about statistics referring to gun control issues are better asked in the statistics Q/A site. One may consider mathematical modeling to study the issue but again this is not a question where we can a priori expect that this will be of special interest to mathematicians. Also the "call for action" aspect of the question is not appropriate.

    Regarding the "aggreeing to disagree" theorem of Aumann. Indeed the whole point is that the agents need not have the same factual information, and they need not reveal the private information they have. The assumption of common proprs implies that they draw the same conclusion from the same information. To say that "they still have to agree" is a bit of an overstatement. The theorem only says that they cannot "agree to disagree". This is a nice and important theorem. However, on a few occasions Aumann and I agreed to disagree.
    • CommentAuthorDL
    • CommentTimeJan 12th 2011 edited
     
    @Qiaochu: I believe "common priors" is meant in the Bayesian sense. Also "common knowledge" has a rather technical definition--see http://en.wikipedia.org/wiki/Common_knowledge_(logic) .
  36.  
    Hello. I wanted to return to a point I raised in this discussion. I wrote, "Suppose there were a political movement to put our national treasury up for grabs (against, say the EU) with a spin of a roulette wheel: spin 17 and we win theirs, anything else and they win ours. American mathematicians would have a professional, societal interest in organizing a comprehensible statement proving the expected futility of this arrangement. How would they arrange for this?" Some suggested certain prominent blogs, but I would advocate that MO is far more prominent and powerful tool for addressing such a catastrophe. I also think that we are not very far from this hypothetical scenario.

    The stock slide happening this very second is erasing hundreds of billions of dollars out of US stock evaluations. The decline is widely believed to be related to S&P's downgrading of the US credit evaluation. Many have long argued (as did Nate Silver on the NYTimes Politics blog most recently), using specific evidence, that these ratings are seriously flawed. Some believe that poor quantitative skills lie behind this ratings failures.

    Question: Would it be acceptable in these circumstances to use MO to propose a polymath project to create a free, open-source credit ratings algorithm to compete with Moody's and S&P?
    • CommentAuthorvoloch
    • CommentTimeAug 8th 2011
     
    @Zaslow: no.
    • CommentAuthorStorkle
    • CommentTimeAug 8th 2011
     

    @Zaslow: yes.

    • CommentAuthorquid
    • CommentTimeAug 8th 2011
     

    No. This is off-topic. And, I would find it very up-setting if this "question" was asked.

    • CommentAuthorfedja
    • CommentTimeAug 8th 2011
     
    @Zaslow OK, but expect that the output will be B- (at best).

    I'm afraid that the best advice a mathematician can give to the modern politicians and economists is "Erase the blackboard and let's start this computation all over". I doubt they'll listen to such words though.

    The interesting question here is whether the mathematicians, as a community, should take more active role in the society. My answer would be "yes, if we are 100% sure about the advice we give and no otherwise". Yes, the people who are at controls are mostly totally incompetent and no, we are no more competent than they on many issues.

    However I am in favor of open MO discussions of any question that has non-trivial mathematical content. So, if the OP shows how exactly he wants to approach the issue from the mathematical standpoint and asks specific questions about data analysis and such instead of asking how to convince other people, I'll vote to reopen. As is, the question has no mathematical content.
  41.  
    @fedja: Thanks. My last question was about a potential question to be posed, not the original one (which I agreed was flawed on several counts -- see above). From what I understand about the way ratings agencies developed their algorithms to analyze CDO's, they were flawed in several statistical assumptions which an open-source algorithm could expose and correct. I dunno, but it is frustrating watching an economic collapse while thinking that our mathematical community actually could do something to help. (In an ironic twist, this closed question just got a "Nice Question" badge!?)
  42.  

    Eric, I find the notion that better mathematics would have over-ridden greed, ignorance, political expedience, and the general ratchets of human nature, rather over-optimistic.

    it is frustrating watching an economic collapse while thinking that our mathematical community actually could do something to help

    I would replace "frustrating" with "hubris". If we can help, we can help as people. I sadly lack either the genius or the chutzpah to apply cohomology of Banach algebras (or my experience of teaching 1st year calculus) to the political problems which are exacerbating the economic woes of that place south of Canada and west of Ireland.

    • CommentAuthorWill Jagy
    • CommentTimeAug 8th 2011
     
    I have a memory that is relevant. During a large conference at msri, they had a panel discussion about opportunities for Ph.D.'s outside academia, the panelists having made such moves themselves. Two women had both worked for IBM at some earlier stage. They agreed on one point, which was that they would invent a new algorithm for something, then go down the hall to tell the engineers, who would refuse to use it. Engineers, I suppose, do their inventing in other directions, and take a large subset of mathematics as fixed in stone. The mathematicians then had a lengthy process of selling their research to the engineers, probably successful on some occasions, not others. Both were happy to have left IBM behind them. I bet it would be much harder to sell an economic model to a government or government agency.

    Yemon, have you ever really tried hard to apply cohomology of Banach algebras to the problems of the U.S.? You can't know it's impossible until you've tried.
  44.  
    Thanks for the conversation, folks. I personally know a number of academic mathematicians and physicists who have had careers in finance, sometimes preceeding and sometimes following faculty appointments. One of the world's most successful hedge funds was created by a renowned mathematician. I don't think that analyzing financial data is beyond the ken of a *community* of mathematicians. As to the other points regarding politics, the idea of creating an open-source algorithm means that the ratings could be published outside of control of a political body. (Ratings of sovereign bonds are dependent on publicly available data.)
    • CommentAuthorNilima
    • CommentTimeAug 8th 2011
     
    There are three assumptions here:
    (a) that a mathematical community has something important to say about a specific economic or political issue
    (b) that this mathematical community will be able to communicate the results of their deliberations to a truly wide audience
    (c) that everyone else interested in the issue will listen to the mathematicians

    The probability of all three assumptions holding in a given instance are small. Case in point: designing safer structures. Mathematicians know how to properly discretize the relevant PDE, and compute approximate solutions. Most commercial software used in practice is provably bad at some of these problems. There are many open source academic codes (by mathematicians) which do way better. Does anyone care? Probably not until something bad happens, and usually not even then. http://www.ima.umn.edu/~arnold/disasters/sleipner.html
    • CommentAuthorfedja
    • CommentTimeAug 9th 2011
     
    There is an issue of taking risks here. You know, like in that funny story about software developer:

    Dad, why does the Sun rise on the East in the morning?
    Every day?
    Yes!
    No deviations, delays, or other strange behavior?
    No!
    Very good, son. Just don't touch or change anything now!

    To accept a new code, one needs to make sure that it, indeed, gives better results, is compatible with the other software, etc. What is in the public domain is a good *core* code. What is often needed is an application-oriented package. I noticed that engineers are quite happy to discuss and implement solutions to the problems that give them a real headache (though not before an extensive testing) and are rather reluctant to change things that work quite (though, maybe, not completely) satisfactorily and I can understand them there.

1 to 46 of 46

  1.  
Back to Discussions Top of Page