A quant review of “The Quants” by Scott Patterson

There were giant mutant quants destroying every … Oh, sorry.  That was ants not quants, and it was a Japanese movie not a book.

Given my blog’s remit, it seems obligatory to review this book.  The full title is significant.  It is “The Quants: How A New Breed Of Math Whizzes Conquered Wall Street And Nearly Destroyed It”.

I see three possible reasons for reading this book:

  • Getting a biographical history of quant
  • Learning quant concepts
  • Arriving at a diagnosis of market ills

Let’s look at each of these in turn.

Quant History

If a biographical history of quant is what you want out of the book, then I think you’ve come to the right place.

Curiously this is the aspect that other reviewers complain about.  There is talk of cliches and not keeping up with the Wolfes.  The book didn’t have that effect on me, and it’s not as if I’m not critical of writing style.  One more thing I don’t understand.

Quant Concepts

Please, please do not read the book to learn quant concepts.  Almost all of the explanations are wrong. Furthermore, if an explanation is correct, you’re not going to unlock it unless you already know the combination.

Market Diagnosis

There were giant mutant quants destroying everything in their path.

That’s the diagnosis that I got, and it seems to be the primary message that is being taken away from the book.  I see three problems with this:

  • It was business as usual
  • It sort of had to be quants
  • It wasn’t really quants

Business as usual

Market crashes have happened even without the benefit of silicon chips providing millisecond latency.  In the previous crash — the dotcom crash — I believe the quants tended to be on the anti-bubble side.  There obviously has to be an explanation for crashes that doesn’t require quants.

Had to be quants

Computers are used in markets.  If there are computers, quants are likely to be nearby.  Pretty much all of the action in markets is now impacted to some extent by quant.

It seems to me that blaming the crash on quants is like blaming food poisoning on food.  It’s not food in general that’s the problem.  It is specific foods and how they are handled.

Wasn’t really quants

The author doesn’t seem to have noticed that the book is more about gambling than about quant.  This is billed as a quant review, which implies there should be some data.  Table 1 presents some data — the approximate number of occurrences of two quant words (“alpha” and “beta”) and two gambling words (“blackjack” and “poker”) in the book.  I’m certainly not of the opinion that “alpha” and “beta” are the most important quant concepts, but references to these comprise, I believe, the majority of references to quant concepts in the book.

Table 1: Some approximate word counts

word approximate number of occurrences
alpha 46
beta 12
blackjack 42
poker 96

With this in mind, I suggest that the title should have been “The Gamblers: How A New Breed Of Card-Counters Conquered Wall Street And Nearly Destroyed It”.  I admit that sales might not have been quite as good with this title.

The Quants

The implied prescription of the book is to ban quant methods.  Those of us who can count past 30 could be disqualified from “fit and proper” status.  That would solve the problem, surely.

I’m not saying that quants are completely innocent.  For instance, the book describes the quantmare of August 2007.  That was the result of a lot of smart people acting stupid.  But the arrow that cut the rope that dropped the curtain that had hidden that stupidity was shot from a non-quant bow.  Or maybe a really, really bad quant bow.

The quantmare seems to have been the result of a liquidation of a quant portfolio.  The speculation is that it was a multimanager that was getting margin calls and the quant fund was the logical thing to sell off because it was the most liquid.  As far as I know no one has held up their hand so this remains speculation.  But lots of quant funds were doing essentially the same thing and so they were all hit.  The leveraged ones at some point needed to close positions because they were getting margin calls.  There was positive feedback — a vicious circle in the vernacular.  The quants shot themselves in the foot, zero collateral damage.

But one of the key things that started the avalanche that knocked the quants in 2007 and knocked everyone in 2008 was the idea that house prices never go down.  Something that defies the data we had at the time.  This was a lack of quant — or at least a lack of reasonable quant — not too much quant.

I do have a diagnosis of market problems.  I searched for a tactful way of stating it.  I failed.  Humans are stupid.  And the bit of cleverness that we exhibit makes us ever so much more dangerous.

Subscribe to the Portfolio Probe blog by Email

This entry was posted in Book review, Quant finance and tagged , . Bookmark the permalink.

2 Responses to A quant review of “The Quants” by Scott Patterson

  1. Adrian Allardice says:

    You are bullshitting yourself when you try to deny that quants are fully involved in the latest ‘meltdown’. Look at this silly comment:

    “…one of the key things that started the avalanche that knocked the quants in 2007 and knocked everyone in 2008 was the idea that house prices never go down. Something that defies the data we had at the time. This was a lack of quant — or at least a lack of reasonable quant — not too much quant.”

    “The idea that house prices never go down” was not quant?? What utter rubbish. This was driven 150% by the idiot pure math quants at S&P and Moodies and every scumbag bank quant [driven by their own lack of morals and scumbag commision driven bosses] who utterly refused to include the data from the 1930s – when house prices fell dramatically. They didn’t include it because then they couldn’t tell lies about CDOs. The same type of quant cretins in 1987 and 1998 refused to put realitic historic data into share or arbitrage models [data such as the 1929 crash et al] and thus produced garbage Gaussian distribution which show that one day 40% market crashes can’t occur except every 10000 years rather than the true result of around every 30 years.

    “..something that defies the data we had at the time…” is the essense of the quant problem. Too much quant in place of some basic thinking – anyone with any tiny bit of brains would know that prices [including houses] can and have fallen dramatically over long periods of time and much more often than normal quant models predict. Look at any data for the last 500 years and only stupid quants would state that using only the last 40 years of data – the crap golden years of houseprice data series that Moodies used – was even vaguely likekly to cover 99.999% of events over the next 100 years.

    It is this kind of crap quant methodology of exact but rubbish science rather that inexact logic – extended from stupid analysis of house prices to CDOs etc then to quant funds – that was deeply involved with the latest meltdown. Each major bull and bear market has its own bunch of different stupid humans but with the same stupid human hubris. The previous one had the stupid dot com analysts who ‘believed’ that every dot com would produce billions of profits – impossible in total. This one had quants with over-confidence in arbitrage models. Without them the bull markets would not have been as high, and the bear markets not as sharp and nasty.

    I can state this with absolute certainty having questioned myopic pure math quants during the last 10 years why they did not study data on housing markets where major price declines occured, when every applied maths quant such as myself knew that a housing crash was very likely. Pure quants were distainful as they knew at their data was right. Last 40 years = next 40 years. Crap!

  2. Pingback: Blog year 2010 in review | Portfolio Probe | Generate random portfolios. Fund management software by Burns Statistics

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>