Finding a starting point to any slot machine analysis can be difficult, much like writing the lead paragraph for a magazine article. Indeed, starting your first analysis when your dataset is the size of a slot operators’ performance database can feel like complete writer’s block. This becomes particularly true in the gaming industry which acts as though every data metric is a trade secret and offers little to no publications as a starting point for beginners. With this in mind, I would like to offer some advice on how to start a meaningful slot analysis.
My first piece of wisdom is that before you start any analysis, you must ensure that your data is clean. A well-known saying in computer programing is “garbage in, garbage out” or GIGO. Simply, GIGO means that any nonsense data input will produce senseless results. The first property I worked at as an analyst had a little over 1,000 slots—I spent the better part of my first month renaming slot game descriptions and cabinets by physically verifying each machine and then transferring the results into the casino management system. I ended the project with a garbage free dataset where sensible analysis could be created.
The first months with EILERS-FANTINI Central Game Performance Database (GPD) was essentially no different, other than adjusting the cleanse strategy to cover multiple properties and covering exponentially more slots… I scrubbed the data before producing our first Game Performance Report (GPR) and continue to do so regularly.
With the chore of taking out the rubbish completed, an analysis can be started. I always teach junior analysts to start by logically segmenting the dataset. For GPR, I start by segmenting the slot performance database into Fair Share analyses—a chart or table comparing units to theo net win generated by basic segments like supplier, denomination and game category. This simple organization allows the data to communicate about the floor mix and illustrates at a high level what segments are producing and, more importantly, which segments are under producing. Using a Fair Share analysis to make strategic decisions—like the correct mix of games to spend capital on as well as what to remove from the floor when the new games arrive.
Another principal benefit of the GPD is the use of a database exponentially larger than an individual casinos. For example, try running a segmentation of your database to only include: owned, $0.01 base denom, curve cabinets, with less than 180 days on floor… you will likely find yourself analyzing a mere fraction of your database with not enough datapoints to draw a conclusion. Whereas applying the same filters in the GPD will produce thousands of up-to-date datapoints in just the trailing few months.
Applying similar filters to only the month of April were featured in Game Performance Report (GPR) and are shown in the chart below. This segmentation of the data displays the TwinStar J43’s strong performance across multiple markets but with less than half the units of the Arc and a third of the first curve cabinet, the Scientific Games Wave. The J43 is the best performer but has a lot of ground to cover to gain a floor mix of the other curve cabinets, all while facing stiff competition from the Helix XT’s release.
Utilizing larger databases with this methodology allows for more advanced analytics.
Once you have insights into your segmented floor mix and make subsequent adjustments, gathering intel on your competition can be the next best step in slot analytics. With the GPD, you can now easily drill down to your region and “walk” your competitor’s floor with the added benefit of seeing indexing performance metrics, all without taking a step. Though I would still recommend participating GPD casinos to physically visit their competitor’s floors, why not take advantage of this added analytical advantage that comes at no cost.
As our GPD continues to grow, the analyses we deliver and the insights individual casinos gain grow with it.