Home Where’s the Money? Part 12: Magnet Games and Paradise Fishing

Where’s the Money? Part 12: Magnet Games and Paradise Fishing

Authors’ Note: This is part 12 of our 18-part “Where’s the Money?” series. In this article, magnet games will be examined and their role in mini casino design explored. Magnet games often dominate the character of both the area and the players in an area. The examination will continue with a deep dive into the performance characteristics of Aruze’s Paradise Fishing, using preference filters and market basket analysis.

Magnet games are leaders in slot performance. They are highly correlated to their surrounding games, and they also outperform the surrounding product (see our discussion of losers, leaders, loners and laggards in the December 2012 issue of CEM). These games are wonderful additions to a gaming floor, as they both attract players to a game, and they might also create “spill,” driving additional play on surrounding games. If a game has a high spill effect, then we have truly found a driver that can be used to optimize the gaming floor. Furthermore, we can measure this spill effect as a “lift” on surrounding games by looking at reverse cannibalization. To do this, we need to understand these two new game metrics of spill and lift, which we can measure by performing a directional analysis.

Spill is the amount of additional revenue that a game contributes to an area due to overflow from its high demand. As additional games are added, the spill will typically reduce. Spill can be measured in dollars of theoretical win. Lift is the percentage of increase in revenue in the surrounding games.

Chart 1: Player Preference and Loyalty
Chart 1: Player Preference and Loyalty
Directional analysis is a probabilistic view of the migration of gaming floor play that shows the likelihood of a player moving from the magnet game to another game. Think of this as a kind of chaining process where the customer’s current play pattern has an increased probability of driving the player to a surrounding game. This directional analysis provides a very insightful view of the gaming preference. Quite simply, the questions are, “From this game, which game are customers likely to move to next?” and “Where did customers who are playing this game likely play previously?” (An in-depth review of directional analysis will be left for a follow-up article.)

To answer these questions, we need to look at player preference. There are two ways of doing this: The first is game preference, which is behavioral and observable, and the second is an actual loyalty for the game. During each visit that a new player exhibits a preference for, a game gives us clues that the player has game loyalty. (See Chart 1.)

­Understanding if a player has game loyalty provides predictive knowledge about many aspects of game optimization. If we are intending to move a game, players who have game loyalty need special consideration, and possibly specific communication, about the movement of the game. Also, games with loyalty may present an opportunity to become the centerpiece of a mini casino strategy, as we’ve discussed in previous CEM articles about Penny Alley and Jackpot Wharf.1

One of the major differences between loyalty to a game and preference for a game is the amount of time it takes to measure the two. Preference for a game can be measured on a single visit—with the understanding that this preference can change over time. Loyalty to a game cannot be measured on a single visit, or even a handful of visits. However, loyalty is no doubt measurable. For example, we can look at the effects of physical movement of the game and measure the number of players who are prepared to seek out the game in its new location. This brings us to the very interesting challenge of calculating the probability that a player is truly loyal to a game if they exhibit game preference.

Chart 2
Chart 2
Bayes Theorem
We will now show how game preference can be used to calculate the likelihood of game loyalty using the Bayes Theorem. The Bayes Theorem is a wonderful probabilistic model that gains knowledge from clues or hints. Let’s calculate the probability of game loyalty (A) given the evidence of preference play (B) on a player’s visit. The probability of A given B is equal to the probability of B given A times the probability of A divided by the probability of B:

P(A|B) = P(B|A) * P(A)/P(B). (See Chart 2.)

Typically the event that we are interested in is not observable, but the evidence of these events is observable.

In other words, if a new player shows preference play for a game, in our illustrative example, they are more likely to have game loyalty. Understanding that this new player has game loyalty means we can say that there is a 60 percent chance they will play on the game during future visits. Further visits provide more clues about the player’s feelings toward the game, increasing the accuracy of the inference model.

Paradise Fishing
In our series of articles about Jackpot Wharf2, we described how this mini casino was recently completely re-merchandised and reconfigured to draw in more visitors from the adjacent Bass Pro Shops at Silverton Casino. The Jackpot Wharf initiative continues to maintain its revenue increase at around 50 percent above pre-change numbers.

Figure 1: Paradise Fishing without Preference Filter
Figure 1: Paradise Fishing without Preference Filter
Paradise Fishing is a popular penny video slot game that features an impressive large screen community bonus. Paradise Fishing has extremely high occupancy and is a strong financial performer. As such, it is even more important to merchandise the immediate area correctly to take advantage of the spill opportunity it provides.

The following section analyzes the preference play on Paradise Fishing to enable changes that will improve the spill effect onto surrounding games. To do this, we need to start with a preference filter.

A preference filter is used to screen out the noise from your data. It makes sure that only the gaming activity that your customers spend on games they like shows up in your analysis. It ignores the trial transactions—customers trying a game and not liking it, customers spending a few minutes waiting for the actual magnet game to open up, etc. A game can show good performance because of heavy trial activity, but this can quickly fall off when it fails to retain customers. A preference filter is one simple way of developing a model for preference play, and it does not filter out players; it filters out some of the play of all players, in essence leaving only preference play.

Figure 2: Paradise Fishing with Preference Filter
Figure 2: Paradise Fishing with Preference Filter
The inspatial graphic in Figure 1 shows the gaming activity for Jackpot Wharf players who showed moderate interest in Paradise Fishing. We wanted to look at this group of players to test our merchandising strategy—we were looking for fairly even play patterns inside Jackpot Wharf. Without preference filters, it appears we have succeeded at our goal: Paradise Fishing players are playing throughout the entire area, and we might assume that it is properly merchandised.

But see what happens when we apply a preference filter in Figure 2.

Suddenly, entire banks are gone! Specifically, see Bank A, which in the original graphic, looked very popular. After applying the preference filter, not a single Paradise Fishing player shows preference play on these games—and that’s our target group for this merchandising strategy! However, as Figure 1 shows, many of them did play on the bank. Possibly the players were merely biding time while waiting for Paradise Fishing to open up. This is a potential high performing location because of its proximity to the Paradise Fishing magnet bank.

Figure 3: Top Preferred Games
Figure 3: Top Preferred Games
Also look at Bank B. In Figure 1, without the preference filter, the bottom two games were the weakest on the bank. But once we apply the preference filter, we see that this group of customers clearly prefers the bottom games. As with Bank A, many customers played the top games, but the majority of the group showed preference play on the bottom. Preference is an extremely important distinction, as product in a high-traffic area may perform well, especially if it is transient traffic.

We can now apply the same preference filter to the entire floor to locate products that this target group prefers to play, and then bring it into the area. Further, we can isolate players who actually prefer Bank A and identify an appropriate area to move it to. (See Figure 3.)

Applying the preference filter floor wide also shows a handful of other preferred products that the target group plays heavily. The majority of their other preferred products are actually on the opposite side of the casino floor. It would make sense to pick up one of these banks and simply move it right next to the magnet games.

Applying preference filters, we can find a new home for Bank A. Figures 4 and 5 show the preference filtered market basket numbers. Figure 4: Game Preference Paradise Fishing
Figure 4: Game Preference Paradise Fishing

In Figures 4 and 5, we are showing “tree map” data visualization. This visualization is very good at handling multiple hierarchical dimensions of data. In this case, each small box represents one slot machine and the size and the color both represent the preference filtered revenue numbers. These small boxes are grouped into larger boxes based on their attributes, for example in Figure 4, we see that Video Slot is a large box that encompasses a large number of small boxes (or slot machines), and that within the Video Slot area the House section is the largest, meaning that house games contain the majority of the preference filtered revenue. Simply put, the tree map is a visual market basket analysis.

Figure 5: Game Preference Bank A
Figure 5: Game Preference Bank A
Using preference filters and tree maps, we can see that customers who are moderately interested in Paradise Fishing have much different product preferences when compared to customers who are moderately interested in Bank A. From this, we immediately see that Bank A players are a lot less inclined to play video slot participation products. They are most likely to play house WMS video slot games, some of the older IGT penny titles and Bally’s Blazing 7s.

Just like we did with the target group, we select Bank A’s most preferred games and are able to locate an area with decent (if scattered) activity and several non-preferred banks. This area should prove successful for Bank A. (See Figure 6.)

Where is the Money?
Figure 6: New Location for Bank A
Figure 6: New Location for Bank A
The numbers on Jackpot Wharf are hard to argue with: An ongoing 50 percent increase in net revenue resulted from a focused and well-informed strategy of gaming floor optimization.3 As illustrated in this article, Jackpot Wharf has even more opportunity for improvement, and the decisions about these games are simply not based on outcome metrics—they are based on optimization metrics. We showed how preference filters are a powerful way of conducting gaming floor optimization and that traditional outcome metrics would have missed these opportunities altogether. We look forward to further updates on Jackpot Wharf as we dig deeper into revenue improvement.

Footnotes
1 CEM January 2012, Cardno, Thomas, Evans: Jackpot Wharf, Part 1
CEM February 2012, Cardno, Thomas, Evans: Jackpot Wharf, Part 2
CEM December 2010, Cardno, Singh, Thomas, Evans: Penny Alley, Part 1
CEM January 2011, Cardno, Singh, Thomas, Evans: Penny Alley, Part 2
CEM February 2011, Cardno, Thomas, Evans: Penny Alley, Part 3
CEM March 2011, Cardno, Thomas, Evans: Penny Alley, Part 4
2 CEM January and February 2012, Cardno, Thomas, Evans: Jackpot Wharf, Parts 1 and 2
3 CEM December 2011, Cardno, Thomas, Where’s the Money, Part 6: Player Experience and Slot Optimization

Leave a Comment