‘Quants’, Kelly, and the RK formula

George Rebane – 19 May 2010 (V5apr18)

[I originally posted this little blurb on some work I had done on the Kelly formula prior to 2010.  Well, after some more noodling and squiggly pushing I found that there was a conflict in the way odds and expected gain/loss were defined which required me to revise what is now the RK formula.  As a result, I took down the 19may10 version and am now replacing it, appropriately edited and modified where required, with the current corrected version which still reflects its original publishing date.]

Contrary to the rumors you’ve been hearing for the last year plus, GeorgeW didn’t cause the Great Recession.  This fiasco had three main progenitors - government idiots, corporate greedies, and some very clever, but not clever enough, Wall Street traders practicing the arcane art and some of the more dubious science of financial engineering.

Scott Patterson has been reporting on things financial for the WSJ, and recently wrote The Quants – How a New Breed of Math Whizzes Conquered Wall Street and Nearly Destroyed It (2010) which immediately shot up to the top of the charts and is still there.  I finished Quants before going on travel and was just going to add it to my collection of finance and economic history books.  But then I thought that there might be some little gems of broader interest there.

Patterson goes back to the beginnings of financial engineering in the 1960s, and takes us right up to the fall of 2009, the Great Recession and all.  In fact, the whole purpose of the book is to explicate the role of the quants in fostering the recession.  What’s a quant?  A quant is a person trained in many of the mathematical ideas and toolsets of the system sciences.  A bunch of them came from research universities like MIT, Caltech, UC, … .  They are generally at the PhD level, so they have a demonstrated ability to think out of the box and extend human knowledge.

Now these young graduates were snapped up by various ‘Wall Street investment houses’ and hedge funds.  Their job was to examine the price behavior of all kinds of securities around the world, and then apply their math wizardry to develop automated trading programs to lickety-split execute profitable arbitrage transactions that their algorithms discovered in the massive daily dataflows of the finance and securities industry.  Since I have some knowledge of these high-jinks (actually, I also claim to be a financial engineer), this kind of historical background fascinates me.  And doubly so since the political reverberations of this recession will be with us for a long time to come.

Financial engineering actually began with Modern Portfolio Theory credited mainly to Harry Markowitz and William Sharpe.  They both became Nobelists for their work.  MPT is used to allocate funds among a short list of competing investments like stocks.  Sharpe’s contribution was the Capital Asset Pricing Model or CAPM which introduced the notorious alpha and beta numbers to a stock.  Brilliant and elegant modeling of an investment world that turned out not to exist – well, not to exist quite like their equations required.

Today MPT and CAPM are used primarily by professionals for CYA purposes to manage other people’s money.  If the client’s portfolio blows up, you can always point to the Nobel prizes and claim you did the prudent thing.  And thereby hangs a tale, or should I say ‘tail’?.

Most of the applied financial phrenology is still based on the view that security prices move up and down sort of like a drunkard’s (or random) walk.  If you assume this, you can fit nifty bell curves to the magnitude of price movements, and off you go into the land of stochastics.  Stochastics is a fancy word for random processes that depend on other random processes.  The problem with the random walk model is that it doesn’t predict big catastrophic price swings very well – practically not at all.

(To continue reading, please Download Quants2_2018)

Comments

[George Boardman]

I'm always reminded of the really "smart" guys who think they've got the markets figured out, like the big-brains behind Long-Term Capital Management. Two Nobel laureates in economics were among the big-brains.

After some initial success, Long-Term managed to lose $4.6 billion in just four months! As they like to say on Wall Street, this time it's different.

[George Rebane]

GeorgeB 1119am - Indeed, and LTCM threw all caution to the wind, following neither Kelly nor RK (which didn't exist at the time). William Poundstone's 'Fortune's Formula' is a very entertaining look at the impact that people like Claude Shannon, John von Neumann, and John Kelly had on both casino gambling and securities markets. I recommend it to you.
https://smile.amazon.com/Fortunes-Formula-Scientific-Betting-Casinos/dp/0809046377/ref=tmm_hrd_swatch_0?_encoding=UTF8&qid=1522959128&sr=8-2

[George Boardman]

I read "Fortune's Formula" a couple of years ago, which brings me to Edward O. Thorp. A mathematician, Thorp is best known for figuring out how to win at blackjack, as explained in his book "Beat the Dealer."

Thorp is a big fan of the Kelly criterion and worked with Claude Shannon to develop the first wearable computer. He also figured out how to exploit pricing anomalies in securities, and he put his theories to work at Princeton/Newport Partners, where he got rich.

A couple of years ago, Thorp put up $1 million to endow a math chair at UC--Irvine. It almost makes me wish I had paid attention in my math classes.