With all the talk about ChatGTP and other AI programs representing both an opportunity and a real and present danger to humanity, it might be interesting to understand the role of AI in our slot world.
But first, let’s examine the chicken-and-egg debate about participation and leased slot machines:
“The optimal percentage of leased machines that a casino should have on its floor can depend on several factors, such as the casino’s budget, the cost of leasing versus buying, the popularity and revenue potential of specific machines, and the flexibility needed in terms of changing out machines to suit customer preferences and market trends.
- Budget: Leasing can be a cost-effective way for casinos to offer a wide variety of games without having to make large upfront purchases. This is particularly true for smaller casinos or those just starting out.
- Cost of Leasing Versus Buying: Leasing can be more expensive over the long term, so a cost-benefit analysis is needed. However, the cost of buying can also include maintenance and repair costs, which are usually covered in leasing contracts.
- Popularity and Revenue Potential: Some machines might be more popular than others or might have a higher revenue potential. In such cases, it might make more sense to lease these machines to ensure that they can be replaced if their popularity wanes.
- Flexibility: Leasing allows for more flexibility in terms of switching out machines to cater to changing customer preferences or to keep up with the latest gaming technology.
Without specifics on these factors, it’s hard to give an optimal percentage that applies universally. Casino operators should conduct a thorough analysis taking into account the above considerations to determine the optimal mix of owned and leased machines for their specific circumstances. Additionally, industry benchmarks and the advice of industry consultants can provide useful guidance.
As a general guideline, one might consider having a higher percentage of leased machines if flexibility and cost control are primary concerns, while a higher percentage of owned machines might be more appropriate if long-term cost-effectiveness and stability are priorities. But again, there isn’t a one-size-fits-all answer to this question.”
The seven paragraphs above are a pretty good primer on the endlessly-debated topic of how many leased or WAP machines should be on your floor. It could have been written by a prof at UNLV, clipped from a casino book on Slot Ops or even penned by a CDC Gaming columnist. It may have even been written by a politician seeking contributions since it flip-flops both ways.
The actual scribe was my computer, which spit out these words from ChatGTP. I typed the query “What percentage of leased machines should a casino operator have on their floor?” In far less time than it took me to type the question, it cranked out the paragraphs above. This latest version (GPT 3.5/GPT 4.0) works only with data that is two to three years old. It is scouring web content from the beginning of time until late 2021. (Note: GPT was developed by “Open AI,” and it is an acronym for Generative Pre-Trained Transformer.)
Based on the non-specific output to my question, it looks like this latest version of GTP relative to the casino world is at the college freshman level with little or no real-world experience.
One thing that may delay machine learning from catching up quickly in our industry is that most of these programs, as mentioned above, rely on scouring the web for content and basic information, then performing sophisticated analytics to find the best fit to the questions proposed.
I’ve always bemoaned the fact that there is not more sharing and spreading of information in gaming due to several factors such as IP rights, patents, and just plain secrecy and competitive concerns. But maybe in the context of BI, that’s not a terrible thing. Our best and most advanced P&Ps are not widely available on the web.
On the plus side, many software and analytic tool providers are using some elements of AI to make their products better today. GTP and the like are very good at creating code which looks for solutions and combinations of data in new and novel ways. Today, it still takes someone to put those new ideas into comprehensive packages that are useful and productive for us. In other words, human beings are still way ahead in this race and seem to maintain a distinct advantage …for now.
For answers to the original slot machine lease question above, well-educated and experienced bipedal carbon-based units, like Quick Custom Intelligence’s co-founder Andrew Cardno and his team, look at additional factors like location, gravity modeling, spatial associations, market basket analytics, devotion indexes and other KPIs, along with producing colorful and insightful data visualizations.
Other sophisticated organic processors like Nick Hogan, CEO of ReelMetrics, use many of those same advanced KPIs and apply their own proprietary segmentation model, while also leveraging their national database of billions of players sessions to drive recommendations.
It seems strange that both these individuals and their firms, which are among the best-of-the-best, are also using advanced AI tools to increase the enjoyment of players and produce stronger bottom-line profits. The irony is that the same technology they are using within their products is working on its own to make both of them obsolete someday (hopefully they’ll make it to retirement or bail out on golden parachutes long before that happens…hopefully.)
Educators have already been hit with waves of term papers generated by CPUs rather than student research. Most journalists, striking script writers and copy editors are keenly aware that the next hire on the copydesk may be ChatGTP 5.0, instead of an intern from Columbia, Missouri or USC.
I asked the current version of ChatGPT several other topics related to slot management, and all were bland or too generic like the “leased” response above.
At present, the top analytic firms are way ahead of “Chat,” “Google Bard” and “Bing AI” programs. But using new AI methods and machine learning, these analytic providers are getting valuable input on some of the age-old casino questions about casino customer behavior, churn, pricing, marketing strategies and product selection that were once considered unsolvable. Their solutions are well beyond the capabilities of artificial intelligence alone.
But remember, ChatGTP-type apps are still in their single-digit versions as of this writing. Does anyone recall the crude first release of the “WordStar” word processor or the pixelated graphics of the original game “Pong[i]?” (Note: The “digitaltrends” website reported last month that GPT-4 just created a playable version of Pong in under 60 seconds. It took Nolan Bushnell and Allan Alcorn months and months to get theirs off the ground at Atari in 1972.)
To put AI’s growth potential in perspective, think about comparing the aforementioned “Pong” to today’s “Call of Duty: Warzone.” Software and hardware consistently, and exponentially, improve with time. I can imagine the recruitment ads a decade or two from now on Indeed or ZipRecruiter:
Director of Slot Operations – Non-human position. Remote or onsite server with minimum 100 Gbps secure wireless capability. GPT-14.0 or higher. Minimum requirements: liquid cooled, parallel interlaced Intel i35s, unlimited cloud storage, Nvidia GeForce RTX 17800 Series graphic capabilities, dedicated redundant UPS and HVAC required. Depending on specs: $18,500/month lease. RFPs on demand.
Computer Operator: Slots –On-site only (location tbd). Duties include cleaning, HVAC maintenance, logging, inspections and installing upgrades for GPT system. High school graduate or GED required. Depending on experience: $21/hour.
It could be even longer before this occurs. I can’t imagine how long it will take before Artificial Intelligence can determine which shade of nail polish on a cashier violates the dress code; why you should ignore the GM who wants to remove all the high limit machines because of one big jackpot; how to crush the gossip about who is cheating in the ‘tip pool;’ when to have a ‘body odor’ discussion; and what to do when your slot tech runs off with the CEO’s or tribal chairman’s daughter. (Good luck with those, AI!)
# # #
[i] Too young to know what Pong is? Take a look at this link: https://www.ponggame.org/
