Case Study $7$: Decision Rules for Racehorse Wagering

Problem Description

People have bet on the outcome of horse races for a very long time. They try to earn large profits from

their bets by finding ways of outwitting other betters. A large number of books have been published on

horse races, and each author claims that following his$/$her methods will lead to profits. So far you have

not been able to make big profits using any of the existing methods. Therefore, you decide to build a

system that will help you and others place sophisticated bets.

Horse racing is one of the most documented sports. There are thousands of pieces of data published on

each upcoming race. This data is available to you, and you need to build a database to keep all this

information. Once you build your betting support system, you need to test it on some sample data and

report your observations.

Database Design

The following are the main entity types of the racehorse wagering database. For each entity type, we

provide some of the corresponding attributes.

$1.$ Horse: The main attributes are identification number, name, birthday, height, weight, breed

$($Arabian$,$ English, etc.$),$ races attended, races won, injuries, etc.

$2.$ Owner $($of the horse$)$: The main attributes are identification number, name, address, etc.

$3.$ Jockey: The main attributes are social security number, name, address, weight, height, birthday,

years of experience, race history, major achievements, etc.

$4.$ Racetrack: The main attributes are name, location $($address$),$ type $($dirt$,$ turf, etc.$),$ distance, etc.

$5.$ Race: The main attributes are name, date, distance, number of times it is organized in a year,

location, number of horses participating, number of jockeys participating, etc.

Access Development

The following are some of the queries one can create to increase the functionality of the database.

Queries:

$1.$ It has been shown that there is a high probability that the horses that have won the race at least

once will win the race again. Therefore, there is a tendency in the public to bet on the horses that

won at least once. To help with the decision about which horse to bet on$,$ run a query that presents

for each horse the following information:

a$.$ Name, breed, age, height and weight.

b$.$ Percentage of time the horse won a race.

c$.$ The average number of times the horse won.

d$.$ Percentage of time a horse won the races that were held on types of tracks such as today$$s

track.

e$.$ Percentage of time a horse won a big race $($races with attendance more than $30$ horses$).$

Rank this information by percentage and average number of wins.

$2.$ In fact, percentage and average number of wins are not very good indicators of how a horse will

perform in the current race. However, this information combined with the information about the

time of the last race, changes in weight carried, current speed rating, and previous class rating are

a better indicator. Run a query that puts together this information.

$3.$ People believe that the following are indicators of a good $($forthcoming$)$ performance:

a$.$ The horse has not raced within one month of today$$s race.

b$.$ The horse had won at least two of its last five races.

c$.$ The horse finished no worse than second in at least five of the last ten races.

d$.$ The horse and its jockey have raced together in the last two years.

e$.$ The horse has either raced and finished second or better in the last ten days or shown an

exceptional workout in the last four days.

Run a query that lists the horses in each of the categories listed above.

$4.$ The following are other $($secondary$)$ factors to be considered in choosing a horse to bet for:

a$.$ Find the horse that was first at the last stretch and won the race by at least one and a half

lengths.

b$.$ Find the horses that in their last race did not finish in the money.

c$.$ Find the horses that in their last race lost more than $$ lengths in the stretch.

d$.$ Find the horses that have a speed rating at today$$s distance within five points of the

highest speed rating of all the competing horses in the past four races.

e$.$ Find the horses that have participated in at least ten races during the last year.

$5.$ Some people demonstrate interest in$,$ or tend to be loyal to$,$ a particular horse. Create a query that

prompts for the identification number of a horse and returns the following:

a$.$ The date and time of the last race the horse ran.

b$.$ The time the horse made on the last race run that was at a distance close to today$$s

distance.

$6.$ People believe that there is a high probability a horse will not win a race given the following

conditions:

a$.$ The horse never won a race.

b$.$ The horse did not win during the last $($say$,10)$ races.

c$.$ The horse never had a race at today$$s track.

d$.$ The horse never ran today$$s distance at today$$s track.