Playing the percentages when it comes to slot machine rankings
In a recently published EILERS-FANTINI Data Insights Report, we focused on including an additional layer to our ranking strategy of top performing slots. The new metric is a Percentile Rank, which further standardizes top games to the various diverse reporting operators.
Percentile Ranks are applied in several industries and areas of everyday life outside of gaming. This is undoubtedly why when the initial question arose, “other than report minimums and filtering, how else can you adjust for the size of casinos when looking at slot performance?” the wheels turning in my mind came to an un-abrupt stop via a notification on my phone. The solution literally dropped into my inbox, via an e-mail from American Express, which read, “your updated FICO score is ready for your review.”
Credit agencies are one of several non-gaming industries that use percentile rankings. Take for example a credit score of 799… by itself, an arbitrary number. However, when compared to the rest of the population, that score ranks you in the 80th percentile in which, 80 percent of the population has a score of 799 or below and only 20 percent have a score of 800 and above. These comparative rankings provide individuals and lenders with a context much like the ranking slot indexes.
The Central Game Performance Database (GPD) now covers over 90,000 slots across 24 jurisdictions, collecting data from 153 casinos. Though reporting from 153 casinos settles some statistics, it can also skew others disproportionality. One factor that causes these tilts is the size of the casino or number of slots on the gaming floor.
To date in the GPD, we have observed on average that smaller casinos have a wider indexing variance between the best games on the floor, indexing near 20 times zone average, while the worst games on the floor index a tenth of the zone average. But in the larger casinos, we see top indexing nears three times the average and bottom indexing remains near one tenth. This means that the smaller casinos more greatly influence the average index of a top performing slot. The individual casino percentile rank provides a solution to this occurrence.
How it works: Before aggregating, all indexes are ranked against the individual casino to determine its percentile rank, which is the percent of games above or below that specific slot at that casino.
The result: The individual casino percentile rank standardizes the indexes by casino—in which the top index of 20x at one casino is converted to 99.99 percent; the top index of 3x at another casino is also 99.99 percent since they are both the highest rank by individual casino.
The percentile ranks are then averaged and sorted in descending order determining the rank of the game. Below is a side-by-side comparison of how the top 10 low denom video reel games fared by percentile rank (left), featured in the Data Insights Report, compared to index rank (right), featured in the Game Performance Report.
Gaining insights into the percentile ranks is relatively simple. Looking at the number one ranked game, Dancing Drums, we calculated a percentile rank of 92 percent which means that on average the 242 slots reported perform better than 92 percent of the games in their same zone.
From the two tables we note most of the games remain in the Top 10 but a few do fall out of the ranks, being replaced by new games. Taking a closer look at the impact of percentile ranking, Fu Fu Fu went from not being included in the Top 10, utilizing indexing, to the fifth ranked game by percentile rank. We see Fu Fu Fu only boasts an average theo net win index of 1.7x from 12 units but when looking at the five casinos reporting those 12 slots, we found that Fu Fu Fu ranked above 87 percent of the other units being reported from those casinos. Highlighting the inverse of Fu Fu Fu, Cash Wizard featuring Quick Hit has percentile rank of 82 percent, causing it to fall well behind games with lower average indexes even though it reports a strong index of 2.0x previously in the Top 10.
Though percentile ranking provides a new perspective to performance, balancing the Percentile Rank along with Theo Net Win Index and the Confidence Interval provides a more insightful indicator of a game’s overall performance. As of now, I believe these are the best metrics to accurately scale performance in the GPD and determine how well a game will perform on any casino floor.
With the math and new ranking explanation established, we want to provide an update of the top performing For Sale core cabinets in the market today, being led by SG’s TwinStar J43.
In the chart above, you can see that the Helix Xt is barely behind the J43 in the most recent snapshot and has traded ranks in none of the last three months. While not on the chart, the unit counts where the newly released Xt trails the J43 by nearly 6 to 1. The full report also featured the top five games by cabinet, as previewed for the J43 & Xt below.
The EILERS-FANTINI Data Insights Report is a supplement to the Game Performance Report and is free to all participating operators.