I "Maltigame selection" CrossReference to Related Applietdons The present application claims ptioity from4 Prvisional Patent Application No 2003905171 filed on 22 September 200, the contents of which is incorporated 5 herein by reference. Introductiun The present invention relates generally to the field of gaming and in particular the invention provides an improved method of game selection in a mltigame 10 environment, Background&9tte ietionh A player in a casino is presented with a choice from among potentially hundreds of gaming machines. Players develop a preference certain ganes from among all the 15 ones availble, and try various machines to find the ones they like. Given this large choice it can be difficult fbr a player to find games they like to play Multi-game machines offer more than one game on a single electronic gaming machine (EGNM) This can dramatically increase the choices available to players, but makes it more difftcult for the player to identify games that they like anongst the many 20 games provided, As technology progresses and multgane machines become more capable the number of games will increase even further as manufacturers entire game libraries could potentially become available on a single machine. Internet gaming web sites have the same problem, with many ganes avaiable via the one gaming terminal (eg a home computer). 25 naming machines in a Casino are often networked for the purposes of security, accounting, credit transfer, bonusing, and for player mrketin. These systems typically track players using a magactic card that the player inserts into the gaming machine, the magnetic card being associated with a player account on a central server US patent 4,636,951 (Harlick) shows a system for networking gaming machines to 30 provide a credit transfer mechanism between machines and a central cashier station, US patent 5,655,961 (Acres) shows a bonusing system, in which player tracking is provided via a magnetic card carried by a player and which the player inserts into a magnetic card reader in a gaming machine to identify the player to the machine and/or the entire system. 35 Throughout this specification the word comprised t or variations such as comprises" or "coimpring', will be understood to imply the inclusion of a stated 2 element integer or step, or group of elements, integers or steps, but not the exclusion of any other element, integer or step, or group of elements, integers or steps. Any discussion. of documents, acts, materials, devices, articles or the like which s has been included in the present specification is solely for the purpose of providing a context for the present invention, t is not to be taken as an admission that any or all of these matters form part of the prior art base or were common general knowledge in the field relevant to the present invention as it existed before the priority date of each claim of this application., 10 Summary of the Invention In accordance with a first aspect of the present invention, there is provided a method of gaming on a gaming system comprising a gaming server, a game selector, a is plurality of gaming machines, and a communications system connecting each of the plurality of gaming machines to the gaming server, the gaming machines each having a display, a player interface, and a game controller arranged to control images displayed on the respective display, the method comprising the steps of: tracking, on the gaming system, play of games played by a plurality of players 20 on the gaming machines; displaying, on the display, a selection of a plurality of games available on a respective gaming machine for the player to play; selecting, using the game selector, one of the plurality games, the selection being determined dependent at least in part on the players past history of playing games zs and a history of ganes played by other players; and initiating, using the game controller, play of the selected game at the respective gaming machine, the game having a game result which, if it is a winning result, will cause the game controller to award a prize to the player. so In an embodiment, the player's past history of games comprises ganes played on the current machine in a current playing session only. In an embodiment, the player's past history of games comprises games played on any machine in the system in a curent playing session and any previous session. W aij 4 G&Y ) WS01J=/70:3 3 In an embodiment, the history of games played by other players complies a history of play on that gaming machine and not other gaming machines. In an embodiment, the history of games played by other players comprises a history of play on any gaming machine in the system. a The selection of games for display by the game selector and their order of representation may be made by a recommendation engine. In an embodiment, the recommendation engine is located in the gaming machine. In an embodiment, the recommendation engine is located on a server connected to to the communications system. In an embodiment, the reconunendation engine collects player data on the playing habits of all distinguishable individuals who can be linked to a plurality of game plays. The player data may be processed to detennine the most popular games with is particular individuals. The player data may be processed to determine the groups of games that are popular with particular individuals. In an embodiment, the recommendation engine processes the player data to determine at least one popular game for the player, maintains for each game in the 20 system a list of other games that would be of similar enjoyment value to the player and uses the list to form the selection of games. In accordance with a second aspect of the present invention, there is provided a method of gaming on a gaming machine connected to a gaming system wherein the gaming system comprises a gaming server, a game selector, a plurality of said gaming 440U_ S& O,6ssc,N) fl kI 2A5J2ZVM'f 4 machines, and a communications system connecting each of the plurality of gaming machines to the gaming server, the gaming machine having a display, a payer interfae and a game controller arranged to control images displayed on the display, the method comprising the steps of: 5 displaying, on the display, a selection of a plurality of games available on a respective gaming machine for the player to play; selecting, using the game selector, one of the plurality games, the selection being determined dependent at least in part on the player's past history of playing games and a history of games played by other players; and 1o initiating, using the game controller, play of the selected game at the respective gaming machine, the game having a game result which, if it is a winning result, will cause the game controller to award a prize to the player. In accordance with a third aspect of the present invention, there is provided a method of gaming on a gaming machine comprising a display, player interface, and is game controller arranged to control images displayed on the display, the method comprising the steps of: tracking, on the gaming machine, play of gaies played on the gaming machine; selecting, using the game selector, one of a plurality games, the selection being determined dependent at least in par on the player's past history of playing games and a 20 history of games played by other players; and initiating, using the game controler, play of the selected game at the gaming machine, the game having a game result which, if it is a winning result, will cause the game controller to award a prize to the player. In an embodiment, for each game on the machine, the processed data results in a 25 list of other games that would be of similar enjoyment value to the player and the list is used in forming the selection of games. In an embodiment, the plurality of games available on the gaming machine comprise gaies stored on the system and accessible by the plurality of gaming machines.
5 In an embodiment, the player's past history of playing games and the history of games played by other players comprises a rank allocated to a game by the player and by the other players respectively. The selection of games may be determined dependent on age information s relating to each of the plumality of games. In an embodiment, the selection of gaies is displayed according to a rank of each game in the selection, the rank of each game being determined by the selection engine dependent on at least one of the player's past history of playing games and a history of games played by other players. io BriefDescription of the Drawins An embodiment of the invention is now described by way of example with reference to the accompanying diagrammatic drawings in which: Figure 1 shows a perspective view of a first style of gaining machine, suitable is for use in systems implementing embodiments of the present invention; Figure 2 shows a perspective view of a second style of gaming machine, suitable for use in systems implementing embodiments of the present invention; Figure 3 shows a block diagram of a control circuit of the gaming machines of figures 1 and 2; 20 Figure 4 shows a block diagram of a systen implementing an embodiment of the present invention; Figure 5 shows a first screen layout used in embodiments of the present invention; Figure 6 shows a variation of the screen of figure 5 used in embodiments of the 2S present invention; Figure 7 is a block diagram of a system which embodies the present invention showing the recommendation engine in detail; Figure 8 is a flow chart of a recommendation process; and Figure 9 is a flow chart of a process for building a similar games table, 3G 6 Detailed Description of the referred Embodimen Referring to Figure 1 a typical gaming machine is illustrated of a type to which the present invention can be appied. The machine illustrated in figure 1 is of a type that allows credit input by insertion of coins or bills but the invention cen aim, be 5 applied to machines that only allow credit input by transfer of credit fom a central cashier or from another gaming machine. In Figure 1, reference numeral 10 generally designates a gaming machine, including a game or games to be played by a player of the machine. The machine 10 includes a console 12 having a display means in the form of a video display unit 14 on wtich a game 16 is played. The video display unit 14 10 may be implemented as a cathode ray screen device, a liquid crystal display, a plasma screen, or the like, The game 16 as illustrated in Figure 1 is a spinning reel game which simulates the rotation of a number of spinning mels 19, however many other styles of game are also possible. A mid-trim 20 of the machine 10 optionally houses a keypad 22 for enabling a 15 player to play the game 16. The mid-trim 20 also houses a credit input mechanism 24 including a coin input chute 96 and a bill collector 97. As illustrated in Figure 2, sone gaming machines use a touch screen for player input, in which case the keypad 22 would not be required on the mid-trim in those machines Instead the keys of the keypad 22 of the Figure 1 machine would be represented as a graphic image 29 on the 20 screen 16 and touch sensors 38 (refer to Figure 3) located adjacent the screen surface would detect touching of the screen to record player selections. In all other respects the machines of Figures 1 and 2 are essentially functionally identiGal. The machine 10 of Figures 1 and 2 includes a top box 26 on which artwork 28 is carried. The artwork 28 includes paytables, details of bonus awards, etc. The artwork 25 28 in the top box 26 may optionally be displayed on a video display such as a CRT of LCD panel. A coin tray 30 is momted beneath the console 12 for cash payouts from the machine 10 In machines employing the present invention the machine may also be 30 connected via a computer network to other gaming machines and a system controller in which case player profiles may be saved on the system controller and credits can be applied to and cleared from the machine via the network. The credits can either be established at a cashier station and transferred to the machine or alematively a player might already have credits in another machine in the network and which they may wish 35 to transfer to a new machine that they wish to play.
7 To facilitate the establishment of player profiles, the use of player profiles to enhance player enjoyment and to enable the secure transfer of cash to a machine, each machine is provided with a card reader 98 and the player is issued with a player tracing card 27 either when entering the premises or when establishing credit in the 5 system, This tracking card 21 is inserted into the card reader 98 of a machine by the player after the player has established a credit on the system and has had the credit transferred to the desired machine. Alternatively the card may be a membership card permanently in the possession of the player and the player may establish a credit in an account associated with their membership record. By inserting the card 27 into the card 10 reader 98 of the machine he or she intends to play, the player identifies him or herself to the machine and establishes that a credit hed in the system belongs to them, In the illustrated embodiment, the card reader 98 is not connected directly to the machine's controller 36 but to the system interface $1, which is connected to the network via interconnection 52 and to the machine controller 36 as seen in Figure 3. 15 Referring to Pipre 3 of the drawings, a control means or control circuit 32 is illustrated, A program which implements the game and player interface is run on a processor 34 of the control circuit 32. The processor 34 forms part of a controller 36 that drives the screen of the video display unit 14 and that receives input signals from player inputs such as the optional keypad 22 (see Figue 1) or the optional sensors 38 20 associated with the pseudo-keypad 29 (see Figure 2). The sensors 38, if used, include touch sensors mounted in the screen of the video display unit 14 and associated with the represemation of pseude-buttons of the keypad 29, displayed on the display 16, thereby replicating the buttons of the keypad 22. The controller 36 also receives input pulses frm the mechanism 24 to determine whether or not a player has provided 25 sufficient credit to commence playing. The credit input mechanism 24 may comprise one or more of several credit input devices such as a coin input chute 96a bill collector 91 , and a card reader 98 or any suitable other type of validation device. In some embodiments of the present invention it is important hat there be a player tracking input device, such as the card reader 24.3, that can be used to associate a particular 30 player with a pardcular player profile and optionally a credit held in the system (either as data held in a machine or in the system controller or possibly in a futher controller reserved for financial information). Note that player tracking does not require knowing the actual identity of the player but is only used to associate the player with a particular player proie and/or credit. This is achieved in the preferred embodiment by using a 35 player tracking card 27 which is a simple magnetic stripe ead encoded with a unique code, that may be issued to the player either when they enter the establishment or when .8 establishment or when they establish a credit in the system and is read by the card reader 24.3. However other methods of player identification can be employed in tracking systems such as pin numbers, scamable tags of various known types sutch as magnetic stripe cards, smart cards, etc, iris recognition, fAger prints or other bio-sensor 5 systems. Finally, the controller 36 opti.oally drives a payout mechanism which, for example, may be ticket printer 41, or a coin hopper 40 for feeding coins to the coin tray 30 to make a pay out to a player when the player wishes to redeem his or her credit, Again however, in embodiments of the present invention, a payout mechanism is riot 10 essential as the player may remove the credit held in the machine by transfeig it to another machine or to a cashier. Referring to figure 4, a system in which the present invention is implemented is ilhsfrated. The system comprises a plurality of gaming machines 10 each connected to a network 63 by its respective system interface 51 and network connection 52. The 13 network connections 52 ave preferably connected to the remainder of the network via a hub 53 although other networking architectures such as daisy chaining may also be employed, Controlling the network is a system controller 54 and a cashier's terminal 55 is optionally connected, either to the system controller 54 directly, as illustrated in Figme 4, or alternatively via the network hub 53. The Cashier may be replaced or 20 supplemented by an electronic cashier or cash incash out terminal 59 comprsing a controller 56 to which is connected a player touch screen 58 and a card reader 57. The electronic cashier uses EFT transactions to debit or credit a player's account at a financial institution to tabhlish or refund a player's credit in the gaming system, While embodiments of the invention will be described by way of example in the 25 context of the conventional gaming machine and gaming system described above, it will be recognised that the invention is equally applicable to other gaming environnents such as intemet gaming where games are played on a personal computer (or, possibly a hand hdd device such as a Personal Digital Assistant (PDA) or Moble Phone) connected to an intemet gaming website, or a stand alone hand held device or 30 any other suitable gaming environment Referring to Figure 4, in the preferred implementation a network- connects EGMs to a server in the traditional manner the server tracks players by their use of a magnetic piayor tracking or identity card W collects statistics about the games played. The lists of games available to the player may be very large, such that it is 35 difficult for the player to page through large number of game selection screens to find games that they like.
9 The game play history for all players is analysed by a recommendation engine 60, which given a particular players game playing history produces a list of other recommended games that the player might enjoy playing based on the play history of other players, The list of recommendations is used to detennine the presentation of the 5 next available games to the player When the player wants to choose a new game to play the BM ranks the unplayed games according to the recommendation and preferentially displays the most highy recommended games to the player. For example, if the gaming machine has 100 games available to play, and games are presented 10 at a time on 10 selection screens, 10 then the first selection screen contains the 10 most highly recommended games. The next selecion screen presents the 10 next most highly recommended games, etc, antil the last screen has the least recommended games Of course the player may not want to accept a list of recommendations and may rather select a game on other criteda such as game type or prize distribution. Referring 15 to Figure 5, in order to accommodate the various differdng player requirements, the total list of games is preferably split into a number of categories according to common player preferences and presented to the player in a selection window 16 as illustrated, Such categories may include games with high wins, games with frequent wins, older (traditional) gams, card games, slot games, and currently popular games. The same 20 game may appear in more than one category, in the screen 16, illutated in Figure 5, games are grouped according to categories selected by touch screen "TuttonsL 201 to 210 and the game icons 211 to 216 displayed in the centre of the screen 16 represent the games belonging to the currently selected category or a subset of the games belonging to the selected category. One of the categores, represented by the 25 "ResconmmerA "button" 202, provides the set of reconunended game, generated by the recommendation system as described herein, Within each category the games may be further ordered, either with sub categories. If the selected games exceed the space available for display of their respective icons, multiple pages will be provided and the player may move through the 30 pages using page buttons 217 and 218. Refering to Figue 6, alternatively the list of game icons 311 to 316 displayed fbr selection by the player are generated according to a set of filters Each game is assigned one or more tags indicating certain attributes (corresponding to categories) of the game, and the tag is used by the filter to select the game (or not) for inclusion on 35 the displayed games list, For example a game may be 'high win' and 'popular% If the player selects either (or both) 'high win' and 'popular' filters then, by touching the 10 corresponding filter "hutton" 301 to 310, this game will appear in the selection. If another filter, such as frequent win, is selected the game will not appear in the selection list as it does not have a frequent win tag, Referably the filters are toggled on or off by selecting the respective filter Thuttor" 301 to 310, Some filters may disable other 5 filters when seleted, if the filters are mutually exclusive, The "Mgh wila" and 'frequent wins" filters are, for example, likely to be mutualy exclusive as it is not possible to design a game with frequent bigh wins (within the usual meaning of these terms within the gaming industry) without returning to the player more money tan they bet Note that the player may simultaneously select multiple filters as illustrated 10 by the grey buttons in figure 6, in which case only games that are associated with all of the selected categories will be displayed. If the selected games exceed the space available for display of their respective icons, multiple pages will be provided and the player may move through the pages using page buttons 317 and 318. In an further preferred embodiment, the number of categories displayed to the 15 player is a subset of those available, and the categories displayed can he selected by the player from the total set. Alternately the categories are dynamically selected by the gaming machine and/or system according to players playing history. Preferably also, some of the categories are selected using the recommendation technique described herein 20 The categories of games end/or games in each category may be changed dynamically, either manually or according to pre-programmed criteria such as time of day or number gaming machines in play. The command to change the games may be initiated from within the gaming machine or remotely from network control system. When the '"recommended" category is selected, the gmees selection player 25 interface may only be partially influenced by the recommendations. For example, the player's favounte games may be presented as well as the top choices from the recommendation list. The casino may also wish to override some of the recomnondations to promote other games. As games have a limited lifespan the age of the game being played is recorded 30 and used in making recommendations, Some players have preferences or games of different ages (eg some players prefer new games, some prefer older games). Numnemus variations on this implementation are clearly possible; * The recommendation engine may be provided by software rminig either on one of the existing servers used for accounting, bonusing, etc, or en a 35 separate reconmendation server II * The recommendation engine may also be part of the gaming machine. This can be beneficial if the gaming machine is not networked, or if there is a need to offload processing from the system for cost or performance reasons, If the reconmendation engine is iroplemented on the gaming machine, then 5 the gaming machine will contain a database of previously collected player statistics (either Ibr that machine only or collected from other machines in the establishment) to enable recommendations to be generated, * Instead of player tracking with a magnetic stripe identity card, a Siartcard may be used. Further, player hidsto could be stored on this card, providing 10 easy transportation of the data to other casinos, Overview of Gane Recommendation Functiou The gaming machines in installations employing embodiments of the present invention provide functionality that allows players to search for and select gaines to play from a hbrary of possibly hundreds of games stored on the system or the 15 individual gaming machine. While the player plays gaming machines on the system a tracking system records the games played and the frequency of playing each game. The tracking system includes a data stmrture and associated code which keeps track of games that have been selected and played by a player for play and this information is then used to provide recommendations for future game selections of the current player 20 and W a basis -for recommendations to others. The system also implements a variety of different auxiliary recommendation function for recom ending games to players. One such function allows players to interactively rate individual games to create a personal game ratings profile, and applies collaborative filtering techniques to each of 25 these profiles to generate personal recommendations. The primary function of the recommendation engine is to recommend games to players curently playing on the system, but it could also be used in the context of a website to make game recommendations to a player before they attend a gaming establishment or in the context of a nconunendation terming attached to the gaming 30 network of an establishment to recormmend games to a player ready in attendance at the establishment but not cuanmtly playing. Such a terminal might also direct a player to a suitable vacant machine for playing a selected game, Briefly, given a unary hting of games that are "mown" to be of interest to a player (e.g, a list of games previously played, by the player), the Recommendation Engine generates a list of additional games 35 ("recommendations") that are predicted to be of interest to the player, (As used herein, the term "interest refers generally to a player's liking of or affinity for an game; the 12 tem 'known" is used to distinguish games for which the player has implicitly or explicitly indicated some level of interest from games predicted by the Recommendation Function to be of intercat) A table is used to map games to lists of games haing "similar affiity for the 5 player similarr games lists"), without the player having to rate the games (although, as mentioned above, the use of mtings data is also envisaged). If a player is known to have shown interest ini say, five games (such as five games the player recently played), the function may retrieve the similar games lists for these five games from the table, and appropriately combiae these lists (as will he described below) to generate the 10 recommendations, The mppings of games to similar ganes ("gamnoaie mappings) may be generated pedodically, such as once per day (or more frequently), from data which reflects the collective interests of the community of players. More specifically, the garne-to-game mappings may be generated by an off-ine process which identifies 15 correlations between known interests of players in particular games For example, in the embodiment described in detail below, the mappings are generating by analysing player playing histories to identify correlations between use of particular ganes (eg, games A and B are similar because a relatively large portion of the players that played game A also played game B), The game-to-game mappings could also reflect other 20 types of similarities, including content-based similarities extracted by analysing game descriptions An important aspect of this embodiment of the Recommendation Function is that the relatively computation-intensive task of correlating game interests can optionally be performed off-line, and the results of this task (game-to-game mappings) 25 are stored in a mapping structure for subsequent look-up, This enables the personal recommendations to be generated rapidly and efficiently in real-tume in response to a request by the player, while still producing a broad varying comparison and not requiring intensive processing. Optionally, the similar games lits read from the table are appropriately 30 weighted (prior to being combined) based on indicia of the players affinity for or cw-ren interest in the corresponding games of known interest. For example, in one embodiment described below, if the game of known interest was previously rated by the player, the rating is used to weight the corresponding similar games list Similarly, the similar ganes list for a game that was played 1000 times in the last hour may be 35 weighted more heavily than the similt games list for a game that was played only 100 tmes in the past houn Embodiments of the invention may also make use of infArmation regarding games that the player has played in the current playing session as inputs to the Recommendation Function For example, if the player currently has two games that they have played today, these two games can be treated as the games of known interest 5 for purposes of generating recommendations, in which case the recommend ations may be generated and displayed automatically when the player views the game selection screen. Using games played in the current playing sessions as inputs, tends to produce recommendations that are highly correlated to the current short-term interests of the 10 player; even if these short term interests are not reflected by the player's long ten gaming history, For example, if a regular player is currently playing with an occasional partner who has different game interests than the regular player, this method will have a tendency to identify other games that are wel suited to the pair of players while they are playing together. is FIG, 7 illustrtes the basic components of the gaming system site 100, including the components used to implement the Recommendation Server 60, The arrows in FIG. 7 show the general flow of information that is used by the Recommendation Function. As illustrated by FIG. 7 Recommendation Server 60 processes requests received from gaming machines 34 over the network 63 and send recommendation lists 20 in reply. The Recommendation Server 60 aclessecs a Games Database 62 to generate lists of popular games and recommendation hits. The Gaming System 100 also includes a player profiles" database 64 which stores account-specific information about players' playing games on the system 100. Because a group of individuals can share an account by playing as a group, a given 25 "player" from the perspeedve of the Recommendation Engine 60 may include murliple actual players. As illustrated by FIG. 7, the data Stored for each player may include one or more of the following types of information (among other things) that can be used to generate recommendations in accordance with the invention: (a) the player's playing history, including dates of play, (b) the player's gamne ratings profile (if any), (c) the 30 player's personal recent playing listings, and (d) a listing of games that were recently (e.g. within the last six months) selected and then exited again without being played, or after having only been played a few times. The various processes 61, 66 of the recommendation server 60 may run, for example, on one or more Unix or NT based workstations or physical servers (not 35 shown) of the Gaminig System 100. The similar games table 65 is preferably stored as a B-tree data structure to permit efficient look-up, and may be replicated across 14 multiple machines (together with the associated code of the Recommendation Process 61) to accommodate heavy loads. The general form and content of the similar games table 65 will now be described with reference to FG. 7. As this table can take on many alternative forms, 5 the details of the table are intended to illustrate, and not limit, the scope of the invention, As indicated above, the similar games table 65 maps a list of games refereeee garmes") 67 to lists of similar games 6 based at least upon the collecive interests of the community of players, The similar games table 65 may be generated periodically (eg,, 10 once per day), or may be continuously updated by the table generation process 66. If the table generating process 66 is run once por day, it will preferably be run offlinm. The table generation process 66 generates the table 65 from data that reflects the collective interests of the community of players. In the embodiment described in detail herein, the similar ganes table is generated exclusively from the game playing histories 15 of the community of players (as depicted in PIC. 7), In otber embodiments, the table 65 may additionally or atematively be generated from other indicia of player-game interests, including indicla based on players viewing (browsing) activities, game playing histories, and game rating profiles, For example, the table 65 could be built exclusively from the long term and/or recent game playing histories of players. The 20 similar games table 65 could also reflect non.collaborative type game similarities including content-based similarities derived by comparing game or descriptions or stmotural game elements. rach entry in the similar games table 65 is pmferably in th form of a mapping of a list of all available games ("reference games ist") 67 to a coasponding list 68 of 25 similar gaines ("similar games lists"). Each similar games list 68 consists of the N (e., 5) games which, based on correlations between playing of games, are deemed to be the most closely related to the repective game 67. Each game in the similar games Ist 68 is stored together with a commonality index ("CP) value which indicates the relatedness of that game to the 30 reference game 67, based on play frequency of the respective games A relatively high commonality index for a pair of games Game J and Game K indicates that a relatively large percentage of players who played Game J also played Gamne K (and vice versa). A relatively low comnmonality index for Game J and Game K indicates that a relatively small percentage of the players who played Game J also played Game K (and vice 35 versa), As described below, the similar games lists are generated, for each game available on the system, by seleting the N other games that have the highest is commonality index values. Using this method, Game 3 may be itnluded in Game KWs similar games liat even though Game K is not present in Game is similar games lit. In the embodiment depicted by FIG, 7, the games are represented within the similar games table 65 using game IDs, such as game catalogue numbers or other 5 identifiers. Alternatively, the games could be represented within the table by title, where each fitle is the name by which the game is known to played. The general sequence of steps that are perfonned by the recommendation process 52 to generate a set of personal game recoimmendations will now be described with refrnce to FIG. S. 10 The FIG. 8 process 70 is invoked in real-time in response to an online action of the player, For example, the recommendations are generated and displayed in reaktime (based on the players play history and/or game ratings profile) in response to selection by the player of a new game instructions selection button 201-210 (See FIGS 5 & 6), The recommendations are generated (based on the player's current and/or recent game 15 playing history) in reaml-time when the player initiates a recommendation buttn 202, and are displayed on te gaming machine display screen 16 and are displayed on the same screen as icons 211-216. Any of a variety of other methods can be used to initiate the recommendations geneation process and to display the recommendations to the player, For example, the 20 reommendations can automatically be generated periodicaly and sent to the player as a message on the screen between games. Further, the personal commendations could be genemted in advance of any request or action by the player, and cached until requested. As illustrated by FIG, 8, the first stop (step 71) of the recommendations 25 generation process involves identfying a set of games that are of known interest to the player. The "knowledge" of the players interest can be based on explicit indications of interest (e.g, the player rated the game highly) or implicit indications of interest (e.g., the player selected the game and played it at least 5 times). In the embodiment depicted in FIG 7, the garnes of known interest are selected 30 from one or more of the following groups: (a) games in the players long term playing history; (b) gares in the playeT's short term playing history i.e, the games played in the current session and, (c) games rated by the player (optionally with a score that exceeds a certain threshold, In other embodiments, the ganes of known interest may additionally or alternatively be selected based on the viewing acdvities of the player. 35 For example, the recommendations process 52 oculd select games that were viewed by 16 the player for an extended perod of time and/or viewed more than once, Further, the player could be prompted to select games of interest from a list of popular games, For each game of known interest, the recommendation function 61 retrieves the corresponding similar gaies list 68 from the similar games table 65 (step 72), if such a 5 list exists. If no entries exist in the table 65 for any of the games of known interest, the process 52 may be terminated; alternatively, the process cold attempt to identify additional games of interest, such as by accessing other sources of interest information. In step 73, the similar games lists are filtered to remove unwanted games. The games removed during the filtering process may include, for example, games that have 10 already been played only a small number of times and then not played again by the player or games that have been rated lowly by the player. The filtering step could alternatively be performed at a diffrent stage of the process, such as after the subsequent combining step After the similar ganes lists 68 are filtered in step 73, the lists are appropriately 15 combined (step 74), such as by merging the lists. The similar games lists 68 are then preferably weighted 75 by muliplying the conunonalky index vales of the list by a weighthig value, The commonality index values as weighted by any applcable weighting value are refened to herein as "scores". In other embodiments, the recoUmendations may be generated without weighting the similar games lists 68. The 20 resulting lst is then sorted (step 76) in order of highes-to-loweat score. The result of step 76 is a lst ("recommendations list") of games to be recommended to the player, In step 75, the similar games lists 68 are opdonally weighted based on information about the player's affinity for the corresponding ganes of known interest. For example, a similar games list 68 may be weighted heavily if the player gave the 25 corresponding popular game a rating of "5" on a scale or 15, or if the player the game for a large number of plays. Weighting a similar ganes list 68 heavily has the effect of increasing the likelhood that the games :in that list we be included in the recommendations that are ultimately presented to the player, In one implementation described below, the player is premed to have a greater affmity for recently played 30 games over earlier played games,. In. step 77, one or more additional games arE optionally added to the recommendations list In one embodiment, the games added in step 77 are selected from the set of games (if any) in the players viewing history, As an important benefit of this stp, the recommendations include one or more games that the player previously 35 considered playing but did not actually play. The games added in step 77 may S17 additionally or alternatively be selected using another recommendations method, such as a content-based method. Finally, in step 78, a list of the t op M (e6g 6) games of the recommendations list are retumed to the gaming machine wher the request was generated. The gaming 5 machine ieorporates this list into one or nore screena of recommendations that are displayed to the player, with each recommrended game being presented as displayed selection icon on the display. The table-geeration process 66 is preferably executed periodically (e.g., once a day) to generate a similar games table 65 that reflects the most recent play history data. 10 The reconmendation process 61 cases the most recently generated version of the table 65 to generate recommendations, FIG 9 iihnstrates the sequence of steps that are perfonred by the table generation process 66 to build the similar games table 65, The general trn of temporary data structures that are generated during the process are shown at the right of 15 the drawing, As will be appreciated by those skilled in the art, any of a variety of alternative methods could be ued to generate the table 65. As depicted by FIG, 9, the process initially retrieves the gaming histores for all players (step 81). Each gaming history is in the general forn of the player ID of a player together with a list of the game IDs of the games played by that player and the 20 dates on which the player has played the ganes and the number of times the game was played on each occasion. In steps 82 and 84, the process generates two temporary tables 83 and 85. The first table 83 maps individual players to the games they played and includes dates and times played. The second table 83 naps games to the players that played such games 25 and again includes the total times played (within a relevant period of interest), In step 86, the process finds for each game in the system ("reference game") those other games played by a common player, and lists the players and games against the "reference game" in a further temporary table 87, This may be accomplished, for example, by selecting from the game-to-players table 85 those games that were played 30 by more than a threshold number (e.g, 10) of the players. For each <reference game, other game> pair/player combination, the number of plays of the game played least by the player is also listed. In step 8g, the pmcess generates the conmonality indexes (CI) For each referencee game, other game> pair in the table 8T As indicated above, the C1 values 35 are measures of the similarity between two games, with larger C1 vaiUes indicating greater degrees of similarity, The commonality indexes are preferably generated such 18 that, for a given reference game, the respective commonality indexes of the corespondiag other gaines take into consideration both (a) the total number of plays of each game in the respective period, and (b) the number of times each player played the least popular of each pair of games. A preferred method for generating the 5 commonality index values is set forth in the equadon below(Eqn ) ZX for all players of games A & B C1 (Game A, Game B}-- 0) Sqrt (Y x.Z) 10 where, for each player who has played Game A and Game B in a given period, 'X" i the nmuber of plays of the least played game (A or B), IY" is the total tumber of plays of Game A and " is the total number of plays of Came B in the period By way of example, foe three games, lameC Game _J and Game-K where 15 Game_G has been played 40,000 times by 500 players, CameI 90000 by 500 players, and GameK 100,000 by 5,000 players. In addition, GameG and GameJ have 50 players in common, and Game_G and Game K have 60 players in common. I X for games C and J is 2500 and Z X for gaens C and K is 960. Applying the equation above to the values shown in FIG. 4 produces the following results: 20 CI(GameG, GameJ2500sqrt(400.timnes.90000)0,042 CI(Ciame G, Game K)=960/sgrt(90000-timnes100,00O)>).010 Thus, even though Games G and K have more customers in common than Gmnes G and 1, Games G and J are treated as being more similar than Games G and K. This result desirably reflects the fact that for Game I players who played Game_C the 25 nber of plays of both games is much greater than for Game K players who played Gamne_G . Because this equation is symmetrica (ie, CI(Game A, Gamie -B)-Ci(Game_1, Game A)), it is not necessary to separately calculate the C1 value for every location in the table R9. 30 Following step 88 of FIG. 9, the similar games table 65 can be complied (step 91) such that each reference game in the reference game list 67 has a respective set of other games in the otherr ganes" list 68 which includes all of the other games fon the table 89 and their associated C1 values Using the vales in the example above, Camel would be positioned closer to the top of the Game G's list than Game K, since 0.042 > 35 0.010.
19 Optionally (not shown), the other games list is filtered by deleting all list entries that have fewer than a predetermined number of (say 10) players in common A non-gaming machine terminal may also be used to allow the player to get recommendations for new games to play This could he implemented at a casino or as 5 an Intemet website and may allow a player to input various likes (and dislikes) for games, then use the recommendation engine to recommend games. Promotions and Advertising Casinos often run promotions, posting advertising literature to the plays to entice them to retum to the casino for further play. A marketing promotion may use the 10 players' previously collected play information and use the recommendation engine to customise promotions to that player, For example the artwork of advertising material may be selected from games that have not yet been played, but are likely to appeal to the player. Alternately free credits may be provided for machines that the player has not yet played, but is likely to enjoy playing. 15 Attract Mode When a gaming machine is not being played after some time it enters an "attraCt mode, where it demonstrates the game it mns. A multi-game machine is able to demonrate any of it's games, so it is beneficial to demonstrate a game more likely to appeal to a nearby potential player. 20 There are eses when a potential player may be identified. 1. A potential player has inserted an identity card, but not yet selected a game or started playing. 2. A player is playing the gaming machine next to the unplayed gaming machine. The player can easily watch the unplayed machine and may either find a new game to 25 play, or play both at the same time (especially in the case of a couple of people playing together). When at least some of the players in the casino have been identified (eg though play machines), their average preference is determined and used. Although not as accurate as targeting an individual player, this has the advantage of continuously and 30 automatically adjusting the presentation of machines in the casino to optimum benefit. It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the invention as shown in the specific embodiments without departing from the spirit or scope of the invention as broadly described. The present embodiments aret therefore, to be considered in all respects as 35 illustrative and not resrictive,