Author’s Note: The debate rages in the back offices around the effect of multigame slots. These are games that are so sophisticated that they can change from a penny video reel to a virtual reel slot at the press of a button. Furthermore, this press of a button can be made by the player. There seems to be two opposing schools of thought on the subject—the first is that players given more choice will spend more time on device; the second is that players will still move around. This is part one of a two-part subseries on multitheme games. In the second part, we will provide examples of real analysis of the questions we are describing in part one.
While we cannot predict the future behavior of customers, we have the good fortune to have access to real data that can be used to shed some light on this question. In this article, we define a series of behavioral patterns relating to play on multigames, and then we will lay out a statistical basis for various behaviors to determine if these behaviors exist and to what extent they do.
Fundamental Approaches to Analytics
When looking at a problem from an analytical approach, it is important to keep in mind that a problem that appears to be simple can require advanced statistical and mathematical tools to solve. The categorization of the analytical space provides guidance on the time of approach we should consider.1
We have good access to the data, and the structure and business rules of this data are well understood. Standard slot management questions such as a report on the revenue generated by the gaming floor is a good example of this category of analytical problem.
In this space we do not even know the right questions to ask; we are often overwhelmed with data. In many cases, multiple ideas exist regarding what the outcomes of various activities are. What question should I ask to find out if I need more of a theme in my configuration options?
In this analytical category, we do know the questions to ask, but we are forced to build models to provide the answers. A large percentage of the questions we face in the understanding of multitheme games involve modeling.
Quite simply, we do not have either the data nor do we know the questions to ask relating to that unknown data. Many operators find themselves in this position with new highly dynamic games, multiple games already on the floor and new configuration controls being added by excited systems providers; however, the base data that shows the player response is simply not collected.
Looking at Dr. Spielman’s quadrant diagram, shown in Table 1, it is apparent that the gaming industry is at best modeling, but mostly researching many of the impacts of the multithemed gaming devices. With the improvements in the system, it is likely that the data will be available2, but the 480 times estimated growth in data will bring its own set of challenges. The challenge will then lie in the high dimensionality of the data.
Slowly Changing Dimensions of Gaming Machines
In a recent article3 we showed how the same question being asked by different departments in a casino can have quite different answers. The challenge in the case of downloadable games is that the slowly changing dimension of game theme configuration is so dynamic that the rules associated with a configuration need to be used to build the appropriate configuration changes. To illustrate this, consider this question: Does adding a theme to the available options constitute a new configuration?
The answer to this question is, it depends. For example, consider adding a new theme to the game and that theme is never played. From the perspective of slot machine analysis, we would not say that this is an entirely new configuration; we would say the change did not affect the performance of the game and that the configuration is essentially unchanged. On the other hand, from the perspective of the theme-level analysis (for example, evaluating different themes offered by a manufacturer), the fact that this theme got no play is significant. The slowly changing nature of the configuration will be covered much more in-depth in later articles in this series where we will expand on the different options and how they can be used to answer different questions.
The Metrics of Multitheme Gates
As we have discussed in our previous article4, these gaming machines of the future introduce new measurements. The meaning, importance and kinds of questions that metrics shed light on are quite different when looking at downloadable games. After we have defined these metrics and discussed some of the different uses of each metric, we will then consider four big behavioral questions.
Time-Based Metrics Discussion
There are many metrics on a gaming device including actual win, theoretical win, handle pulls and time on device. In Table 2, we consider only time on device or theme. In many ways, this metric is the simplest metric on a gaming device, unlike the hotly contested variable definitions of theoretical win5 or questions like whether to count the bonus rounds on a game as part of a single game or multiple activities.
Four Big Behavioral Questions
1) Do players spend more time on multigame machines?
2) Do players treat changes in game on multigame machines like moving to a different game?
3) Do all players like multigame machines?
4) If players like more choices, will they play online games and not visit physical casinos?
Question 1) Do Players Spend More Time on Multigame Machines?
Players have lots of choices on a gaming floor, and from our experience we can say that the average rated session is about eight to 10 minutes. To a game analyst, this may seem very short. It seems like players like to sit down, play a few hands, and then move to the next game. This behavior generates an almost constant movement of players between machines, almost like the musical chairs of the gaming floor. Now with such a fundamental behavioral pattern, as gaming floor analysts we need to be very careful when making assumptions about the impacts of giving players more choices.
Machine-based analytical approaches to answering this question include:
1. Is the time on device longer for multiple themed games?
2. If so, is it because the multiple themed games cause players to play longer, or are players who play longer simply more likely to be attracted to multiple themed games, without any incremental time on device for those players?
3. Do the players who now play on the multiple themed games spend more time on these devices before physically moving to another device?
4. Do players change theme instead of changing device?
Question 2) Do Players Treat Changes in Games on Multigame Machines Like Moving to a Different Game?
Do players sit and choose different games at one location, or do they follow the current trend of moving around the floor endlessly?
Player-based analytical approaches to answer the question from the player perspective include:
1. Break players into two groups—those that play the default game only and those that keep changing the theme. Estimate the proportion of players who have migrated to multitheme devices.
2. Can the differences in the behavior be explained by player dimensions such as age or gender?
3. Can the differences in behavior be explained by other behavioral data, excluding machine selection (such as average bet or velocity of play)?
Question 3) Do All Players Like Multigame Machines?
Clearly it is unlikely that all players will immediately like the new offering. The question is, can we migrate the play to the new games? Furthermore, in migrating these players to the new games, should we encourage them to make selections from the device menu, or are we better to create an environment more like the existing single theme devices?
Player and machine-based analytical approaches to answer the question include:
1. Roll out two banks of machines—one bank of multithemed games and one bank of single themed games (with same themes available at both banks).
a. From the machine perspective, which bank performed better overall?
b. From the player perspective, what were the differences in the players on bank 1 vs. bank 2?
Question 4) If Players Like More Choices, Will They Play Online Games and Not Visit Physical Casinos?
To fully explore this question, we are intending to look to other industries such as horse racing and video games. This will be covered more in future articles, but the danger is if people think that any game can be offered at a single screen from the comfort of their own home, will players migrate their play to Internet gambling? Of course today this is not possible because of regulations, but rumors are that this may be changing.
The big question is, should we be offering products to that altering gaming behavior to encourage choices at one device, more like the Internet gambling experience?
Into the Land of Unknown
As we enter the doors of the analytically unknown, ”researching the data” and the questions that need to be answered to manage the downloadable games, it seems likely that the analytics are about to change radically. While it is exciting analytically to explore new data sources and find new questions, it seems to the authors that the industry is well into production of these new products without having the tools to either collect the data or measure the effects of this exciting new technology.
We remain convinced that the degree of control and the precision of offering is a powerful asset in the hands of the knowledgeable and we continue on our quest to expand this knowledge.
1 Based on Dr. Howard Spielman, TDWI Las Vegas 2005 workshop.
2 Dr. Ashok Singh and Andrew Cardno,”The Petabyte Era of Gaming Data,” Casino Enterprise Management, September 2008, pp. 20 - 22.
3 Dr. Ashok Singh and Andrew Cardno, “Perspectives of Data: Financial vs. Slots vs. Marketing,” Casino Enterprise Management, June 2010, pp. 20 - 23.
4 Andrew Cardno, Dr. Ashok Singh and Dr. Ralph Thomas, “Gaming Floors of the Future, Part 1: Downloadable Games,” Casino Enterprise Management, July 2010, pp. 22 - 28.
Andrew Cardno has more than 16 years of experience in business analytics, ranging from modeling health care drive times to casino gaming floor analytics. He often presents on the future of analytics across the world and has spent the last seven years living in the United States and working with corporations around the world. He can be reached at andrewcardno[at]yahoo.com.
Dr. A. K. Singh is a professor at UNLV. He has taught statistics, mathematics and operations research courses at New Mexico Tech and advanced statistics classes including Time Series Forecasting and Data Mining at Harrah Hotel College at UNLV. He has more than 80 publications in theoretical and applied statistics and can be reached at aksingh[at]unlv.nevada.edu.
Dr. Ralph Thomas is Vice President of Database Marketing for Seminole Hard Rock Gaming. During his years in the casino industry, Thomas has focused on maximizing profitability by applying statistical analysis to the company database. Previously, Thomas spent 15 years in academia, as both a student and a lecturer of mathematics. He can be reached at ralph.thomas[at]stofgaming.com.