Dark Pools: The rise of A.I. trading machines and the looming threat to Wall Street

I have just finished reading Dark Pools by Scott Patterson. This book compliments Patterson's earlier book The Quants, which I have previously reviewed. Again, not the kind of book you would want to read if you are only interested in sports gambling. However, if like me, you are interested in automated trading of sports betting exchanges then this is a useful read. Plenty of names are familiar to me, having worked for Reuters, the former owner of Instinet, a "pool" discussed in the book.

Dark Pools discusses the rise of automated trading on Wall Street, beginning in the 1980s. Initially, there was resistance from NASDAQ and NYSE who wanted to protect the human market makers through whom stock trading was solely performed. Many algorithmic traders believed (correctly) that they were being ripped off by the market makers who would agree amongst themselves as to the appropriate spread for a particular stock so as to milk the most profit for themselves. If the trader was a big bank then market makers would also "front run" their own orders to get in first and push the stock price up further, again for profit.

The growth of personal computing combined with networking allowed for the creation of the first independent exchange (Island) through which buyers and sellers could electronically transact stock sales quickly and cheaply with narrower spreads. This form of trading developed into algorithmic trading (routing an order to the exchange with the most favourable price) and then to high-frequency trading as technolgies improved. Liquidity and volume increased and spreads narrowed. Even traditional exchanges like NYSE went electronic. It looked as though everyone was happy.

As ever, when money is involved, the bottom line corrupted the system as trading houses and exchanges colluded to make profits. High frequency traders and electronic exchanges split commissions between them using the maker-taker system of trading. The maker who put up the price sharing the commission with the exchange, whilst the taker of the price paid the commission. These rebates (as the maker commissions were called) became a raison d'ĂȘtre for HFT who now placed orders simply to hoover up these rebates. Special orders were created by the pools that favoured the high-frequency traders front running the market and that were not available to other investors.

With decimalisation of the exchanges, spreads were so tight that the only way high-frequency traders could make any money was through the maker-taker rebates. The more trades the more rebates, which made up for the market making trading losses. The scrapping of the up-tick rule for short selling allowed the bots to move the market in any direction they so chose. All of this resulted in an average holding time for a stock of 22 seconds and fewer IPOs for new companies. Hardly the ideal way of investing in your country's economy. Regular investors were duped through dark pools where institutional investors could get the best prices. Pension funds and day traders in their homes ripped off with the worst prices in the lit pools.

Dark Pools is a fascinating read that will be of use to the automated sports trader for a little insight into what it takes to reach the top as a programmer in financial markets. You may even get a few ideas that you can use on the sports exchanges.

Amazon - Dark Pools

Secrets of Successful Betting

Secrets of Successful Betting just looks like a stocking filler put out by Raceform. However, it is packed with 23 chapters, full of useful information.

The first two chapters give a very complete history of the origins of bookmaking, followed by the mechanics of the modern betting market.

Chapter three covers hedging, be it photo finishes, ante-post, price coupling or a backer's book.

Staking systems are covered in chapter four with confirmations of my belief in upping your bets when you are ahead and pulling back when you are losing. You won't see mention of Martingale or Fibonacci. Progressive, proportional and Kelly staking are all covered.

Your bankroll is covered in chapter five, covering psychological aspects too. Multiple and each-way bets are covered in chapters six and seven.

All aspects of handicapping are covered in chapters eight to thirteen. As my regular readers will know, this is not a subject that I care for, greatly, so I am not qualified enough to comment on these chapters. Needless to say a thorough account of conditions and selection systems is given. Also there is mention of losing run probabilities. Even a winning system has its bad days.

Chapter fourteen covers estimation of win and place chances with an easy to read treatise of the Harville place pricing model.

British and Irish race courses are covered in chapter fifteen. Chapters sixteen to twenty-three cover much reference material that can yield new ideas.

This is not a book for those looking for an Ainslie or Beyer style handicapping book. However, the book covers the mechanics amd mathematics of bookmaking more precisely than any other book that I have read. For the beginner it is a must read and for the expert it can yield some gaps in your knowledge.

The Compleat Horseplayer

This is a review of Dr. David Edelman's book The Compleat Horseplayer, which is not to be consued with Tom Ainslie's book of the same name. I have never read Ainslie's book as I am not a fan of "handicapping" techniques.

Edelman is part of a long line of economists who have used horse race betting markets as part of their research. Some of his research papers can be found on this website.

The Compleat Horseplayer consists of nine chapters that are worth every penny of this slim volume. The first chapter is an introduction to the elements of horse race betting; odds, markets, value and information. Odds, markets and value don't really need mentioning to readers of this blog.

Information is the most important aspect of horse race betting. The successful bettor is one who has the best information. And the best informed bettors are usually insiders with information that is not privy to members of the public. Edelman gives you a formula for evaluating information and determining its financial value.

The second chapter is titled Money Management and it is no surprise that the chapter concentrates on Kelly Criterion. Towards the end of the chapter we are also given formulae for value, volatility and the Sharpe's Ratio, which is something that I have used for evaluating systems.

Chapter three deals with Technical Analysis of prices, including arbitrage, hidden information and exotic betting.

Fundamental Analysis is covered in chapter four. Class, speed, weight, fitness and some red herrings are discussed. At the end of the chapter a method for evaluating multpile fundamental variables is given with the Multinomial Logit Model and mention of Bolton and Chapman's paper, which can be found on this website.

Chapter five deals with exotic betting and place pricing models. Edelman gives an interesting alternative to Harville, Henery and Stern's place pricing model. After some discussion with Edelman via email I managed to produce a program that generates Edelman place prices to compare with existing models.

If you do wish to contact Dr. Edelman then insure that your query is of the taxing academic kind and not that of the inquisitive amateur otherwise you will soon outstay your welcome. A consulting academic does not come cheap and time is money.

In chapter six we are given a treatise on placing bets. As this book was published in 2001 there is no mention of betting exchanges so this chapter is the least useful to us.

Evaluation of an ongoing system is given in chapter seven with discussion of Regret, Deconstruction and Information Theory's method of scoring the value of information.

Chapter eight gives a little time to automation and chapter nine gives some concluding remarks.

For a small book there is a lot of useful information packed inside it that will send you scurrying for research papers. A valuable addition to the intelligent bettor's library.