Paul Wilmott: most quants are stupid

Yike Lu

Finder of biased coins.
Having a high IQ in no way guarantees that you will not do stupid things. I've witnessed this firsthand countless times, both as an observer and as a participant (though not in a quants context). Paul's statement gets a thumbs up from me.
 
More from the same source:

Let me explain. Many quant books contain vast volumes of unrealistic mathematics. Some people get carried away with the beauty of this mathematics with no corresponding understanding of finance and, more importantly, of human nature. These people are dangerous, as you cannot talk to them about the real world. If you tell them their models do not work, they’ll talk of all sorts of abstract notions, proving themselves right in their heads. Unfortunately, all this is without any reference to the real world. And in finance—which is as much about people as mathematics—if you can’t grasp that, then that is dangerous.

I have always advocated the mathematics “sweet spot,” that fine balance between a sufficiently advanced knowledge of mathematics to do the job in the real world, while not being so abstract as to lose your head in the clouds. You must not dumb down quantitative math, else you cannot understand more complicated derivative products. But, equally, you do not want to stray into the even more dangerous area of really high-level math, where people get carried away by the subject’s beauty. There are some 5,000–10,000 Masters in Financial Engineering graduates churned out each year, and I would not employ a single one of them myself, as they are so hopelessly out of touch.
 
Dr Wilmott is the proprietor of wilmott.com, the popular quantitative finance community website, the quant magazine Wilmott and is the Course Director for the world’s largest quant education programme the Certificate in Quantitative Finance (CQF)
I train more quants than anyone else on this planet
There are some 5,000–10,000 Masters in Financial Engineering graduates churned out each year, and I would not employ a single one of them myself, as they are so hopelessly out of touch.
These points bring interesting questions

Since his CQF program trains quants supposedly differently, would he employ any of his students? That's called put the money where your mouth is.

Also, Dominic Connor (Wilmott's business partner) or any quant recruiter shouldn't have a problem finding the best quants. They just have to hire the CQF graduates instead of the 5-10K hopelessly out-of-touch MFE graduates.

Somehow, I don't see any of this happening.
 
I cant help but feel a bit on edge when a leader in the field such as Wilmott is willing to make such a provocative claim. I understand that he intended his statement as more of a generality rather than a fact, but its another blow to the legitimacy of the MFE brand and similar studies in a time when the quant community is already vulnerable to criticism. I'm sure that I'm overreacting, but I wonder if there's a revolution brewing... Who knows in 10 years the Financial markets may be focusing more on the "qualitative" aspects and then we'll all be called "QUALTS"
 
most quants are stupid..unless you read my book...or take my course.

While he right for the most part, his 'sell' is terrible.
 
most quants are stupid..unless you read my book...or take my course.

While he right for the most part, his 'sell' is terrible.

His "sell" is a bit too crude. But his book on quant finance ("Paul Wilmott introduces Quantitative Finance") stresses heuristic insights using math, and I agree with him that that is the correct direction. The definition-lemma-theorem-corollary style doesn't work that well in applied math, let alone finance.
 

bob

Faculty (Undercover)
Totally agree. From where I sit, modeling is about coping in a world of incomplete information, and about doing the best you can to capture how the world does work, rather than how it ought to.

It's not that the pretty math with a solid theoretical foundation has no place, but for every useful tool out there you will find many others that, despite the hard work and intelligence that went into creating them, are showy nonsense.

To my mind, this is precisely why solid training as an FE is important. It takes market knowledge, a good grasp of mathematics, and a solid understanding of the practicalities of implementation in order to know and evaluate the available options. The fundamental work of an FE is to exercise informed judgment.
 
Totally agree. From where I sit, modeling is about coping in a world of incomplete information, and about doing the best you can to capture how the world does work, rather than how it ought to.

It's not that the pretty math with a solid theoretical foundation has no place, but for every useful tool out there you will find many others that, despite the hard work and intelligence that went into creating them, are showy nonsense.

To my mind, this is precisely why solid training as an FE is important. It takes market knowledge, a good grasp of mathematics, and a solid understanding of the practicalities of implementation in order to know and evaluate the available options. The fundamental work of an FE is to exercise informed judgment.

I couldn't put it so eloquently myself. :)
 
Quants are not stupid

I think some people are really stupid!

Quants by nature needs a good level of Mathematical understanding. The so called high power "abstract"
mathematics such as concentration (Levy) functions and other measure theories only help to develop the subject
further. Yes models do have limitations and those who base any trading decisions on them 100% are asking for trouble and yes
indeed stupid. Wilmott thinks all quant maths should be noddy PDE's well it's not. Only by applying deep abstract pure mathematical ideas can any mathmatical based subject grow, look at theoretical physics! So for all you applied mathematicians out there understand that your subject is only advanced by understanding the underlying pure. So I suggest Wilmott and others go and learn some real mathematics before passing comment.
 
I think some people are really stupid!

Quants by nature needs a good level of Mathematical understanding. The so called high power "abstract"
mathematics such as concentration (Levy) functions and other measure theories only help to develop the subject
further. Yes models do have limitations and those who base any trading decisions on them 100% are asking for trouble and yes
indeed stupid. Wilmott thinks all quant maths should be noddy PDE's well it's not. Only by applying deep abstract pure mathematical ideas can any mathmatical based subject grow, look at theoretical physics! So for all you applied mathematicians out there understand that your subject is only advanced by understanding the underlying pure. So I suggest Wilmott and others go and learn some real mathematics before passing comment.


I think your point "only by applying deep abstract pure mathematical ideas can any mathmatical based subject grow, look at theoretical physics!" is true, but "growing a subject" is primarily the job of professors, not quants.

Paul Wilmott perhaps incorrectly chose the word "stupid" to express his opinion, but his basic point is clear and sound (i.e., the gap between theory and practice, and the danger of being obsessed with maths).
 

DominiConnor

Quant Headhunter
First up, although Paul is my business partner, and I teach part of his course, I am in no way his spokesman. He is entirely capable of doing that for himself.

I suspect he was using "stupid" in the sense of "idiot savant", ie people who over-specialise, and because they are excellent in one area try to apply that blindly to all problems. That's not unique to quants, to an extent we all do that. Recently I was explaining to a friend my most successful technique for picking up women in bars which was wholly based upon surface catalysis of organic chemical reactions.

I speak to at least as many quants as Paul, and too many lack what I call "awareness", in that they blindly accept models as reality without drilling down to the deeper issues that warn us of boundaries where this model might fail horribly.
There is a delicate balance here, so delicate that I am not as harsh as PW when people get it wrong. To get good in the use of a particular set of models you have to immerse yourself in them, and it is thus very easy to start treating them as truth, not useful approximations to real things. Indeed the nature of "real" in this line of work is quite slippery. For instance the famous mathematician / programmer Dijkstra wrote a series of parodies of a future math department who sold equations. The idea was meant as a joke, even now to most 'real' people the idea that you can make a good living buying and selling standard deviations is really not obvious. Many of us here will go to even less intuitive areas of finance.

I have identified a type of 'nearly good' quant I label as an algegraist. All quants can do algebra, and a good % can do symbol manipulation vastly better than I can. For some that is a means to a useful end, but for others I am left with the impression they can find solutions to PDEs (or whatever) that I can't or at least much faster than I ever could, but they have no model in their head of what it means.
"Meaning" is tricky to define here, and if it were easy the skill would be cheap, and the price paid for quants tells us that it is not.

Here's one interview question that might illuminate what I'm trying to express here, and it's simply asked, but perhaps not simply answered.

Why do stock returns approximate to a log normal distribution.

Forget the fact that they don't exactly do that, but focus on why the pattern is so common.
 
I suspect he was using "stupid" in the sense of "idiot savant", ie people who over-specialise, and because they are excellent in one area try to apply that blindly to all problems....

I speak to at least as many quants as Paul, and too many lack what I call "awareness", in that they blindly accept models as reality without drilling down to the deeper issues that warn us of boundaries where this model might fail horribly.
There is a delicate balance here, so delicate that I am not as harsh as PW when people get it wrong. To get good in the use of a particular set of models you have to immerse yourself in them, and it is thus very easy to start treating them as truth, not useful approximations to real things. Indeed the nature of "real" in this line of work is quite slippery.

I have identified a type of 'nearly good' quant I label as an algegraist. All quants can do algebra, and a good % can do symbol manipulation vastly better than I can. For some that is a means to a useful end, but for others I am left with the impression they can find solutions to PDEs (or whatever) that I can't or at least much faster than I ever could, but they have no model in their head of what it means.

To have these qualities a quant would need a broader perspective on the world -- some idea of modern finance, how it evolved, its role in the world. Along with some idea of political economy. For instance, the political economic conditions that allow quant finance to exist -- a fiat dollar in a world where dollar is still global reserve -- are fragile and have existed for less than forty years. There are indications it is coming to an end. It will have a profound impact on the world of finance, and by implication on quant work. Yet such matters are never discussed even on this forum. The distinct impression one gets is of science and engineering graduates entering quant study and quant work with tunnel vision, focusing exclusively on applied mathematical techniques and computer programming -- but having not the foggiest of the broader world in which these arts have their present limited applicability.

How many people here realise the financial crisis of the last three years is not conjunctural, not some fluke, but symptomatic of deep-seated structural problems that are not going away and that these will have -- are having -- a profound impact on quant employment, and on the nature of quant work?

Yet to be fair such insight cannot be expected of someone in his mid- or even late-20s.
 

bob

Faculty (Undercover)
How many people here realise the financial crisis of the last three years is not conjunctural, not some fluke, but symptomatic of deep-seated structural problems that are not going away and that these will have -- are having -- a profound impact on quant employment, and on the nature of quant work?

More than you think, I would wager. The same could most likely be said of traders, salesmen, and other practitioners in the business.

Still, in the heat and light of actually doing the work one has (perhaps) chosen, I find that while it may not be unusual for people to consider broader implications, it is always unclear what bearing this has on the mundane decisions that are the stuff of life for mortals like us. This fact--that we need not "[squeeze] the universe into a ball / To roll it toward some overwhelming question"--is actually one of the mercies of existence. We know we can and will die, yet it is for precisely this reason that today we must set about feeding and clothing and sheltering ourselves, our families, whatever the weather. And, yes, whatever the consequences, if it comes to that.

I agree that we are now experiencing a potent change of some sort, but the mistake is to believe that this is somehow a new, unexpected, or overwhelming condition. It is and has long been a defining characteristic of the modern (note: an English word whose current usage and meaning dates from long before the Industrial Revolution) that you never truly know "what rough beast, its hour come round at last, / Slouches toward Bethlehem to be born."

Somehow the change's particular character manages to seem both incomprehensible at the time and inevitable in retrospect, which to me is one of the delights of seeing it all unfold.
 
I agree that we are now experiencing a potent change of some sort, but the mistake is to believe that this is somehow a new, unexpected, or overwhelming condition. It is and has long been a defining characteristic of the modern (note: an English word whose current usage and meaning dates from long before the Industrial Revolution) that you never truly know "what rough beast, its hour come round at last, / Slouches toward Bethlehem to be born."

Somehow the change's particular character manages to seem both incomprehensible at the time and inevitable in retrospect, which to me is one of the delights of seeing it all unfold.

You are quite right. Though if I may so -- with some diffidence -- the writing has been on the wall for several years. What is happening now is part of longer-term trends, tectonic shifts in power, whose outline can be discerned without knowing exactly how the dynamic will play out. Still, as you point out, food has to be put on the table. And the day-to-day exigencies of life may leave scant time to reflect on broader changes.
 

DominiConnor

Quant Headhunter
@Bigbadwolf I agree quants need to build a view of the world outside their screens. That's why in the Quant Careers Guide, we recommend not only a some non-math finance books, but a subscription to The Economist newspaper.
 
@Bigbadwolf I agree quants need to build a view of the world outside their screens. That's why in the Quant Careers Guide, we recommend not only a some non-math finance books, but a subscription to The Economist newspaper.

I prefer the FT myself. The Economist Intelligence Unit does, however, produce quarterly country reports that show perspicacity.

If I had to recommend general reading offhand, it would include any book by Michael Hudson, F.W. Engdahl's recent Gods of Money (for the history of US finance capitalism), Jack Rasmus's Epic Recession (historically flavored analysis), and The Great Credit Crash, edited by Martijn Konings. I haven't yet seen Chossudovsky and Marshall's The Great Economic Crisis, but I suspect it should be added to the list. Books like these would shed some insight into how finance fits into -- or rather dominates -- this modern world of ours, and how this came to be.
 
"Many quant books contain vast volumes of unrealistic mathematics. Some people get carried away with the beauty of this mathematics with no corresponding understanding of finance and, more importantly, of human nature."

I could not agree more. I see many colleagues that live literally in another world.

I think every preofessional should draw a line between academia and business. An abstract-maths loving quant is more suitable for an academic career because in the busines environment if you dont add value to the company you are useless !!!

I totally agree with Paul and Dominic. At the end of the day they can train as many quants as possible but they cannot really CREATE or SHAPE them. it means they give them the tools but they cannot really change their way of thinking.

I think that the meaning of the sentence written by Dominic should be very clear to anyone who seriously wants a job in the financial services industry:

"I speak to at least as many quants as Paul, and too many lack what I call "awareness", in that they blindly accept models as reality without drilling down to the deeper issues that warn us of boundaries where this model might fail horribly."

In other words, get real because more difficult does not mean more profitable !!!!

PS I dont know Dominic , I never met Pual and I never attended the CQF programme this is simply an unbiased and sincere analysis
 

Yike Lu

Finder of biased coins.
I wonder what kind of training quants get when they start at a big bank.

Often times the recruiters go for the smartest math/physics/engineering undergrads/grads/PhDs who often have no understanding of finance. They then brush off the objection that "I don't know a lick of finance!" with "You're smart enough, we'll teach you."

Well, what do the banks teach these guys? It appears to have been woefully inadequate, at least in the past, and this has come back to haunt them and therefore lead to Wilmott's statement.
 
Top