This is the second and final extract from James Quinn’s best-selling book The Best of Thoroughbred Handicapping, Leading Ideas and Methods, which we have chosen as one of the 10 best racing books of the last 20 years. Go to for more details.

Beyond publishing the books and ?magazine articles that established him as the high priest of handicapping in the 1930s and the 1940s, Robert Saunders Dowst went public in 1936 with perhaps the only fundamentally sound system capable of continual seasonal profits.

Dowst’s system had been born a year earlier in 1935 when a St Louis betting commissioner, no less, a gentleman sporting the handle of “Liberal Tom” Kearney, burst into print with the unerring observation that the only way to beat the races was to play winners.

Other than indicating that the horse Adobe Post in 1934 had won 20 of 50 starts and that a bet on each would have netted profits, the commissioner provided his followers with no directions as to how to select the winners in advance of the races.

But Dowst did.

In Profits on Horses, the Dowst Consistency System came to life. In its simplicity and scope it was indeed a system for all times. Dowst postulated two verifiable assumptions, from which he derived one principle of selection and 11 rules of exclusion.

Following “Liberal Tom’s” magnificent insight, the Dowst system was based on the precept that good horses can beat bad horses. Good horses, said Dowst, were those that beat members of a specific class consistently.

Thus, Dowst put forth his two premises:

(a) all thoroughbreds are divisible into fixed classes; and

(b) when a horse wins a race or runs close he normally does so by virtue of his own speed, gameness and quality, and not by the advantage of a jockey, post position or a clever trainer.

The chief difficulty was establishing the operational definition of consistency. After tinkering with diverse formulae, which did not work, Dowst hit on the one that did. To wit, the Dowst principle of selection:

  1. Play to be limited to horses that had won at least a third of their starts when finishing in the money at least half of the time, provided any qualifying horse is the only one of its kind in the race.
  2. A horse with 10 starts this year can be rated on this season’s record alone. If starts number fewer than 10, rate the horse on this year’s and last year’s records cumulatively, regardless of total races.

When the system appeared in book form and in Esquire, the national stampede began. As Dowst had predicted, the system rang up munificent profits during the whole of 1936. It repeated the feat in 1937, by which time Dowst had prepared a list of the horses that qualified as a play.

For a time the sweet smell of success permeated the air. The secret of beating the races was out, and it worked. But it came to pass that Dowst was wrong on one important point. Dowst himself had argued the system would remain failproof unless so many of the public bet on consistent horses that their prices bottomed out under the weight of the money.

Knowing the contrariness of the horseplayer, Dowst dismissed that dismal possibility. On that he erred. The public bet Dowst’s consistent horses off the board, and the Dowst Consistency System stopped working.

Five decades later, in the context of contemporary handicapping literature, what is the legacy of Robert Saunders Dowst? Does the Dowst Consistency System deserve a revival?

Dowst on racing and handicapping is not so out of date, as the leading authority of the century’s first half left a rich and pungent body of work.

He has left handicappers as a first contribution the theoretical definition of class not yet improved. Good horses still beat bad horses, and a horse’s class is arguably best assessed by identifying the specific class of horses it can beat consistently.

The Dowst system worked, not because he eventually discovered a working definition of consistency, but because he had precisely comprehended the true nature of thoroughbred class.

Dowst repeatedly contended that class held the key. Indeed the Dowst Consistency System would have been better named the Dowst Class Consistency System, as the essential ingredient was demonstrated ability against a specific class of horse.

Dowst’s exclusion rules honoured his high regard for class repeatedly, and Rule 8 prevented play on system horses when “definitely stepped up”. Others forbade playing fillies against colts, horses aged seven or older, chronic quitters, and claiming horses valued at $1,500 or less.

When the Dowst system falls flat today is on the centre point of consistency. The probability studies of William Quirin concluded that consistency was overrated. Horses that have won three or four of their past 10 starts do win more than their share of races, but:

(a)    the public over-bets that kind; and

(b)    inconsistent horses win enough.

Quinn showed that horses that won just one of their 10 previous starts won almost a fair share of their starts. Fred Davis’s probability studies supported Quirin on the matter of inconsistency, concluding this was insufficient reason to regard horses as non-contenders, particularly in claiming races.

Yet the Davis data did suggest a modern variation of Dowst. Davis found that recent consistency outperformed consistency. Recent consistency was defined as winning two of the latest six starts.

Moreover, recent consistency proved more important among better horses, and in studies of allowance races, recently consistent horses performed significantly better than inconsistent ones (won one or none of their latest six). Davis did not report whether recently consistent horses in allowance races returned profits.

A modified application of the Dowst Consistency System is tenable, at small investment until profit margins are determined. The principle of selection now reads:

  1. Play to be limited to horses in allowance races that have won at least two of their latest six starts, provided any qualifying horse is the only one of its kind in the race.

Before presenting the rules of exclusion, which apply without exception, it’s instructive to consider why Dowst’s carefully calculated operational definition of consistency no longer applies. To be sure, it’s a matter of modern thoroughbred form the methods trainers use to regulate form, the demands on form of the modern racing calendars, and the resulting variations in horses’ form cycles. Dowst did not take the form factor seriously.

Excepting downright unsoundness his system ignored it. In Dowst’s time approximately 8,000 horses were in training to compete on a limited calendar. Relatively sound and able to begin with, when racing began the 8,000 were relatively fit and ready to race.

Trainers did not leave enough time to race horses into top condition, and fewer horses became severely overworked during the shorter season. On these points times have changed. During 2001, no less than 72,205 horses competed on 7,661 racing days.

Untalented, unfit, and overworked horses do not easily become consistent horses. Nowadays handicappers best find the class of the field, not so much by identifying the kinds of horses a horse has whipped consistently, but by closely evaluating form cycles, to determine whether one horse is ready to run at its authentic best today.

Click here to read Part 1.

By James Quinn