SERVING THE QUANTITATIVE FINANCE COMMUNITY

gatarek
Topic Author
Posts: 345
Joined: July 14th, 2002, 3:00 am

Strange kind of philanthropy

Last edited by gatarek on February 24th, 2004, 11:00 pm, edited 1 time in total.

Nonius
Posts: 5115
Joined: January 22nd, 2003, 6:48 am

Strange kind of philanthropy

QuoteOriginally posted by: gatarekkapital,I would be happy with rules similar to scientific ones, i.e. public domain with some respected rules, including hierarchy. I don't mean censoring stoch vol BGM and wouldn't like stoch vol contributor to censor it. On the other hand - I'm really afraid that one day all models will be patented and that will be the end of derivatives research. If suffices to patent one simple model - you will have domino effect after. Creating a mechanism preventing something like that would be of particular value.Background of my observation is (obviously!) personal. We offered implementation of BGM model to one of big European banks and our competitor was chosen. OK, I will have many other projects, but in my view a hierarchy was broken.don't worry Gatarek, the patents don't really work.....

gatarek
Topic Author
Posts: 345
Joined: July 14th, 2002, 3:00 am

Strange kind of philanthropy

Last edited by gatarek on February 24th, 2004, 11:00 pm, edited 1 time in total.

kr
Posts: 1885
Joined: September 27th, 2002, 1:19 pm

Strange kind of philanthropy

Sometimes one should think good and hard about what certain words mean:a _businessperson_ takes ideas and tries to produce some kind of economicsa _researcher_ takes a dozen cups of coffee and turns them into ideaswhat is protected in general is usually: research for business purposes. I guess this is because research by itself operates on some kind of socialist economics. Probably there was a good reason why things evolved the way they did. I don't think anybody will ever take the JPEG thing seriously because I am not getting rich by fiddling with my wilmott icon (though maybe there is a case for that other mega source of internet traffic, where they must be making lots of money).To say that you're subsidizing somebody else... maybe that's a bad word choice, but the reason we have businesspeople is that somebody needs to generate the economics. That they used gold from your mineshaft is mostly irrelevant, because on its own, gold does not sell itself. That money does not appear without the gold is also important, but one needs to control the gold. Ideas can only be controlled to the extent that you control their raw dissemination. And an idea that is not disseminated may as well not be an idea - otherwise it's just a tree falling in the forest.Anyway, history tells you that research is more likely to die when it is not widely disseminated, rather than the opposite as you suggest. Otherwise, we would be publishing research article after research article on nuclear fuel processing to get the Iranians and N Koreans out of the game. This just doesn't make any sense. I have a bit of an issue with people who try to make quantfinance a purely theoretical subject. To me it seems a bit artificial to develop a field that originally came from practical problems and then push it in the direction of pure theory. If you want to do stochastic research, call it that, not quantfinance. That said, I don't see much use for quantfinance unless a businessperson is involved. One could take this even one step further and say that if you really are contributing something major, you'd know about it (and apparently this is the case). Nonius said most of this already... the sweet spot in many ways is to be a cigar-chomping guy who knows his way around the academic literature.

kapital
Posts: 236
Joined: July 15th, 2002, 8:16 pm

Strange kind of philanthropy

I don't think anybody will ever take the JPEG thing seriously ...maybe not the JPEG thing, but IP is a really hot and serious topic right now. you must have read about the sco claims. for the last 2 years, microsoft has been trying to scare companies wanting to switch to linux with the "watch out for IP litigation" line. out comes SCO with a claim that 80 f*cking lines of code that have made it into the linux kernel (out of millions) and that these lines represent the key IP that keeps the Unix like OS alive. At first, people are pissed, but no one really takes the suit seriously. Guess what? along comes Microsoft, who "licenses" the code in question for an undisclosed amount. think anyone is laughing about SCO's attempt to bring linux down now? think the boys in legal&compliance have started to care about linux in the enterprise? yup. maybe the litigation wont amount to much, but a lot of damage is being done in the mean time.also a lot of smaller level IP related scams going on. check this out: youmaybenext.com. there is a company going around, waving some IP threatening small (but solvent) companies that they either pay up or go to court. being scared, poorly advised, whatever they pay up. then they use that bankroll to do the same to slightly larger companies.

kapital
Posts: 236
Joined: July 15th, 2002, 8:16 pm

Strange kind of philanthropy

Gatarek,OK, I will have many other projects, but in my view a hierarchy was broken.probably not broken as badly as you thought. imagine how many extra lap-dances, gladhanders, promises to hire the MD's junkie cousin to do documentation, etc. it took for the other company to out-compete "the other company who has that guy that invented the model"? talking about massive externalities, you get a major positive one by having your work in the public. having it peer reviewed and pointed to as 1st rate academic material, then cited in who knows how many extending papers or emperical analysis must be worth a fortune to your company in advertising alone.net-net, i bet you benefit a ton (intangibly perhaps) from having other people adopt and use your model. case in point of an intangible benefit: look how foolish i looked above. my point was lost and all anyone thought was - what an asshole. and all it took was: "hey, man, -> Gatarek as in BGM". think that would have happened if you were the creator of the GammaPi method*? i sure as hell doubt it.*incidentally, what happened to those posts? i tried to find one to link to it, but they don't seem to be around anymore
Last edited by kapital on June 11th, 2003, 10:00 pm, edited 1 time in total.

gatarek
Topic Author
Posts: 345
Joined: July 14th, 2002, 3:00 am

Strange kind of philanthropy

Last edited by gatarek on February 24th, 2004, 11:00 pm, edited 1 time in total.

sdw
Posts: 61
Joined: January 16th, 2003, 12:29 pm

Strange kind of philanthropy

I think it's true what has been said earlier in this thread, that if all academics had patented their ideas then that would significantly stifle the furthering of research. Fortunatley, it has only been very recently (in modern history) that mathematics has taken on such a corporate role, ie in quant finance. Correct me if I'm wrong, but is this not the first time math has been so directly applicable in a business sense, and so ideal for the ultimate end of generating revenue? The old adage about why should i learn algebra when I'm never even going to go there seems far less relevant these days.It's the first time such an intellectual subject has had to contend with idea protection and patent laws: for instance why would Gauss worry about censoring his Fundamental Theorem of Algebra when its dissemination would have no forseeable financial penalty, and in fact would be only likely to benefit him in the long run by stimulating other great minds?I think we're on unfallowed ground here, and it will be interesrting to see the direction things take.

Nonius
Posts: 5115
Joined: January 22nd, 2003, 6:48 am

Strange kind of philanthropy

QuoteOriginally posted by: gatarekkr,I am no more a university professor, so my bitter post was probably a typical reaction of a biznesman after losing a tender . Then I noticed that my winning competitor sold my product (to some extend) and found it unjust. Maybe I am wrong, but a financial model is not an abstract and remote background for a practical application - a pricing model is such an application itself. On the other hand I don't expect anything good from patenting of financial models (provided it was legal) - neither for the patent owner, nor for the industry. I wonder what would be the just and effective solution - current "wrestling with no rules " is not.Gaterek, what is your favorite beer? tell me what it is, and the next time you stop in Paris I'll hook you up with some beers, chicks, and we'll BS about martingales, manifolds, and geometry under a Parisian full moon on my terrace.....you game for that?

gatarek
Topic Author
Posts: 345
Joined: July 14th, 2002, 3:00 am

Strange kind of philanthropy

Last edited by gatarek on February 24th, 2004, 11:00 pm, edited 1 time in total.

kr
Posts: 1885
Joined: September 27th, 2002, 1:19 pm

Quotea financial model is not an abstract and remote background for a practical application - a pricing model is such an application itselfthis is where I disagree with you... the model does not generate revenue all on its own - you have to milk it, and that requires farmhandsanyhow, just because the drug companies cannot indefinitely patent these new wonderdrugs they are developing does not mean that new drugs are not being developedit even suggests a kind of 'speed' in the information mining process.. you don't dig all the gold out of the hill at once and just dump it on the market. Anybody who has upgraded their M$software knows how this stupid game works (and it does indeed work). I have a small angle on this as well because some of our CLO technology has recently gone through the US patent office... It was not my doing and I have always thought that this was a dumb idea. The point is that you can twiddle with small details and it would be difficult to say whether the new implementation is any different from the patented one. At the end of the day this just generates work for lawyers and costs us all money, because they are defending a very vague boundary (to which your contribution is at the margin). Maybe it's just my own personal philosophy, but I think if you're gonna make money, you should do it in a way that is not really contestible in this fashion. If you want to play legal games, become a lawyer and rip us all off. gatarek Topic Author Posts: 345 Joined: July 14th, 2002, 3:00 am Strange kind of philanthropy Last edited by gatarek on February 24th, 2004, 11:00 pm, edited 1 time in total. RowdyRoddyPiper Posts: 529 Joined: November 5th, 2001, 7:25 pm Strange kind of philanthropy QuoteYou've just given a strong arguments, why trucks, airplanes and other tools should be free. A truck does not generate revenue all on its own - you have to milk it, and that requires farmhands Your use of a truck, airplane or other tool excludes others from using that truck, airplane or other tool, the same argument does not apply to financial models . Someone using your model (or a variant of) does not prevent you from using your model. A competitor using your model to do business with a bank that you felt you had a relationship with, or were competing to have a relationship with does exclude you from doing business with said bank, imposing a cost on you, unfairly I might add. I'm sure that they have foregone the opportunity to be introduced to your refinements for the model, insights to other areas and your charming personality, much to their regret. The reason (and this is just based on the facts that I have in front of me) that they did not chose you is that they had no way of verifying that the model your competitor was hawking was yours with perhaps some minor modifications. The need for a heirarchy (such as their is in the scientific community) is one way to approach this signaling problem. The pitfall is that the people who are making the purchasing decisions almost always do not have the knowledge of the person providing the solution (otherwise why purchase) and they are often in a tangental field (making recognition of the contributor/contribution difficult).I would propose something along the lines of an industry group for model (idea??) certification. The idea being as follows: The applicant for certification pays a fee ($5,000, \$50,000, sliding scale???) for a review of their idea by a panel to ensure that it is indeed a novel solution to the problem. The applicant draws up which parts of the implementation they believe to be new and the panel checks them off as either a true or false claim. Once the idea is certified as a new solution for a problem (problem type) the idea holder is at least armed with the good housekeeping seal of approval so to speak. Whether or not this carries any weight when you go to pitch a solution will vary from client to client. I for one would view it as a large plus, not only do you have an idea that hopefully no one has been implementing, but you also have the information that whoever is coming to you can think creatively about problems and really develop new solutions. There are plenty of pitfalls (no one cares about the cert, inability to get competent people to certify, cost prohibitive to scholars, students, entrepeneurs, etc.)Patenting and idea will be ineffective as a means of signaling originality. I don't think that the patent office is capable of understanding what the differences are between different models. Not that they are incompetent, it's just too specialized of an area. Also collecting rents from the patent will be difficult because people will either use your model covertly, or use something else that they don't have to worry about ponying up for. Also the logistics of negotiating agreements and enforcing compliance are very tough for something that is not a physical good. In structured finance the cases of "models" are much more trivial, they are mostly cash flow engines that allow for situational analysis. The levers you can pull on and where they are located are the main differentiators between models, along with methods for speeding up the generating and distributing of cashflows. I've had a few cases where people have showed me their model and it bore an uncanny resemlance one that I completed and distributed with some deal info a few deals back. The first time it happened I got kind of upset, but thereafter I usually just focused that energy on trying to remember what I did wrong in that particular implementation, it makes up for the lack or recognition of the contribution, that's for sure.

N
Posts: 2808
Joined: May 9th, 2003, 8:26 pm

Strange kind of philanthropy

Gatarek,If you could rewind history, what would you do different? Patent the model?? Keep it confidential??Use it yourself in trading?? Write a paper or two??I see the Sobol sequence patent as ironic. Most everyone seems to see value in a technique whichlowers the Kolmogorov complexity of generated series. When they try to calibrate models, no informationin yields no information out.The math involved with obtaining the actual time-dependent 'covariance' IR surface is extremely hard. The usual string theory approaches are way too primative. Do you think that might suggest a confidentialityapproach since a patent becomes public domain??Regards,N

trc
Posts: 95
Joined: April 4th, 2002, 2:28 pm

Strange kind of philanthropy

Gatarek,I want to thank you for the Libor market model because it is so beautiful.I believe that your desire for control of its use is ill advised.Instead it is gratifying to see that its use is spreading.Obviously your competitors like it too and so will the users.You will be remembered for the scientific contribution and nobody cares which company bought which implementation from whom.Only the seller cares -- briefly.Money is secondary.Do you know who was the third richest man even only two years ago?I don't and I don't care.But I still like the proof that there are infinitely many prime numbers from 2000 years ago.