The Commentator Information System: A Usability Evaluation of a Real-Time Sport Information Service Marie-Anne Midy

Carlos Jensen

Yunrim Park

School of EECS Oregon State University 1147 Kelley Engr. Ctr Corvallis, OR, 97331, USA

School of EECS Oregon State University 3061 Kelley Engr. Ctr. Corvallis, OR, 97331, USA +1-541-737-2555

School of EECS Oregon State University 1147 Kelley Engr. Ctr. Corvallis, OR, 97331, USA

[email protected]

[email protected]

[email protected] ABSTRACT Many of the most vivid recollections we have of major sporting events and accomplishments as TV viewers and sports fans are things like scores, statistics, or the TV images of specific events. What made many of these live in our memories, or hold our attention in the first place was often the commentator or sportscaster, their words of wisdom, or lack thereof, or the excitement and emotion in their voices, echoing our own hearts. The commentator is a vital part to the TV sports experience, part of what sometimes makes sports more enjoyable through this medium than at the stadium. In order to provide accurate and timely data to TV viewers, radio, or online listeners, commentators sometimes rely on a system called the Commentator Information System (CIS). In this paper we present an in-depth usability study of a CIS, how commentators do their job, and how the process may be improved.

Categories and Subject Descriptors H.3.5. Online Information Services: Data sharing, Web-based services, XML; H.5.1. Multimedia Information Systems: Hypertext navigation and maps, video; H.5.m. Information Interfaces and Representation (e.g., HCI): miscellaneous; K.m.k. [Computing Milieux] Miscellaneous: Sports

General Terms Design, Reliability, Experimentation, Human Factors

Keywords Sports IT, Olympics, Entertainment, Television, Commentator Information System, Usability, HCI, Ethnography

1. INTRODUCTION The use of computer technology in sports is a pervasive phenomenon. Athletes and their coaches use computers to analyze results and performance, create simulations, and manage Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. ACE’07, June 13–15, 2007, Salzburg, Austria. Copyright 2007 ACM 1-58113-000-0/00/0004…$5.00.

schedules [16]. Event organizers and sports federations, especially those responsible for large international events such as the Olympic Games, World Cup Soccer, or World and European Athletics Championships, use computers extensively in the planning, management, and dissemination of information about their events. Additionally, sporting events are increasingly relying on networked sensors, or high-speed camera systems like HawkEye [5], to gather and relay results, rulings, and other data which may affect the event. This data has five main consumers; the athletes, judges, and organizers hosting the event, the audience (whether in the stadium or at home), and the media. Most recent work in the HCI field has focused on the needs of spectators, both on-site, and at home, watching television or online [12]. Relatively little work has been done on understanding and meeting the needs of broadcasters. The information needs of this group are more complex, as they need detailed, timely, and correct information in order to plan and produce compelling programming. Sports broadcasting is big business, regularly attracting prime-time crowds, generating large revenue streams, and requiring large investments in acquiring broadcast rights, purchasing or moving equipment (especially when the events take place abroad), coordination, and production. For the Turin 2006 Winter Olympic Games, broadcasters paid over $833 million in licensing rights alone, and it is estimated that this investment will reach $1.7 billion dollars for the Beijing 2008 Summer Olympic Games [11]. Obviously, with such a large investment, broadcasters are under pressure to capitalize by producing the best viewing experience possible. This is not always an easy job, especially due to the scale of many of these events. During the 14 days of the Athens 2004 Summer Olympic Games there were 301 events in 28 different sports, with 11,099 athletes taking part from 202 countries. Over 44,000 hours of TV coverage were dedicated to these games globally [7], which translated into 34.4 billion viewer hours. In order to produce this content, TV and radio broadcasters had more than 12,000 journalistic and support staff on-site during these games. NBC Universal alone paid a record $793 million for the broadcast rights, and produced 1,200 hours of coverage of the Athens 2004 Olympic Games [10]. Given the magnitude, and concentration of competitions at events like the Olympic Games, broadcasters and commentators are highly dependent on accurate, reliable, and timely computer support systems to help them do their job. To help broadcasters handle the complexity of these major events and help them plan and produce better programming, organizers will sometimes offer

1

a broadcaster access to a system called the Commentator Information System (CIS). While different events will have different CIS implementations, these follow a general template. From usability and software engineering perspectives, applications such as the CIS pose very interesting challenges. The user population, though highly knowledgeable and skilled, have to operate these systems while delivering compelling and engaging live commentary to millions of viewers. Timing is crucial, and the cognitive load and level of stress is very high. Mistakes, while not life-threatening, can effectively end a lucrative career. In addition to its usefulness to commentators and broadcasters, the potential for further enhancing the local or remote viewing experience through the presentation of CIS-like systems to viewers is very promising. One example of where this possibility is currently being explored is NASCAR and Formula 1 racing in the USA, with the NEXTEL FanView™ [9]. The NEXTEL FanView™ is a portable device with a touch-screen interface providing custom views of official live and past data about racing events. This system could easily evolve to be used in other sports, or similar systems could be implemented online by tapping into the CIS information. We decided to examine these CIS from a usability perspective, particularly the system used for the Olympic Games and Athletics Championships, to determine how well these systems met the needs of broadcasters. In order to do this we performed a study of TV sports broadcasts, the accompanying CIS, its layout and content, and conducted a field study of commentators doing their job. The rest of this paper is structured as follow: First, we give a brief summary of related work before describing the CIS in more detail, followed by a description of our broadcast study, usability study, and field study. We wrap up by presenting our findings and observations, and suggestions for future work.

2. RELATED WORK Sports and sports-related entertainment are very big business. Companies have invested heavily in sports, and maximizing the benefit from their broadcast rights. This includes a sizable investment in the exploration of technology meant to enhance the viewing experience, and to reach wider audiences. Great examples of such technologies include Hawk-Eye [5] and Cyclops [4], both within tennis, which are changing the way local and remote audiences, as well as athletes and officials experience the game. Another example, aimed primarily at improving the experience of the fans in the stands, is the NEXTEL FanView™ [9], which not only allows NASCAR fans in the stands to have direct access to official results, statistics and team information, it also gives fans access to some telemetric data from the cars on the track. A number of systems have also been developed to make sporting events available online, where streaming video is complicated because of bandwidth constraints, demand, or license rights. One of the systems with the longest and most successful track record in this area has been the Wimbledon PointTracker by IBM [17]. Taking advantage of advanced camera technology, remote viewers are presented with live information and visualization of the balls position on the court, effectively allowing the remote viewer to replay the set from any angle at any time, albeit without seeing the players.

A similar tool is the BBC Virtual Replay, [3], mainly used for English FA Cup and European soccer championships. With this Shockwave application, online sports fans can replay their favorite soccer games or goals, played out with 3D animated characters, controlled by data gathered from the game. In addition, they have the option to focus on a specific player and see the game from their perspective. Like in PointTracker, the camera can be set by the user to best fit their preferences. With a multitude of options, again, the remote spectator is in full control of what they want to see. On the broadcast side things are somewhat different. While major broadcast corporations are able to develop and use sophisticated visualization and information systems, most get by with the systems provided by the event organizers. For the most part, the tools available to commentators consists of a TV monitor and, when the organizer provides, a touch-screen system called the Commentator Information System (CIS). These systems are closely guarded by their owners (Atos Origin – [1], ST SportService – [13], and others depending on event). Only very limited amounts of documentation exists for this class of systems, and no studies of their design or evaluation have been published. Studies conducted about the overall organization of the Olympic Games IT infrastructure have been published for the past few events. Usually they focus on the resources used, the technologies involved and the results obtained through data and feedback from viewers, athletes, and media [15, 14, 6]. Though not the primary focus of these papers, some details and information related to the production, deployment, and use of the CIS is available. Starting from the 1994 Winter Olympic Games in Lillehammer, [15], to the 2004 Summer Olympic Games in Athens, [8], the technical resources employed in the organization of the Olympic Games have greatly evolved in quantity and quality. While the CIS remains essentially the same, interacting with the same components – the INFO database and servers – the network architecture has gone from coaxial cable to optic fibers and the communication from wired to cell. Furthermore, the number of people involved in the planning and organization of the Games is growing [6], as is the demand for information during the events. Because some reports are issued prior the end of the Games and without final statistics about the event [8], they give an idea of the planning of such events, and their expectations. The focus seems to be mainly on the success of the Olympic Games from a local audience perspective and is not so much on the information delivered by radio or TV journalists. The CIS seems to be the biggest concession on the part of the organizers to meeting the information needs of broadcasters. There are a few statistical reports about the Olympic Games that focused on number of CIS terminals, sports, venues, etc. that have been published. It is interesting to compare the resources available during Winter and Summer Olympic Games: 500,000 historical records in winter vs. 1,500,00 in summer, or, 535 CIS screens in winter vs. 800 in summer [2].

3. THE COMMENTATOR INFORMATION SYSTEM The CIS gives broadcasters and commentators access to both realtime and archive information about athletes, events, and sports. On the archive side this includes information on the sport or events history, each countries or team participating, including

2

athlete and coach bio’s, sport rules, and official and personal records. On the real-time side this includes access to information from sensors in the venues, qualifications and rankings, judges decisions as they happen, as well as start-lists (information on athletes pre-start, including starting number) (see Figure 1). In addition, most venues are wired to sensors providing meteorological data if this is relevant to the sport. As such, the CIS is only the interface to a much greater IT system (Figure 2 gives an overview of a major sporting events’ typical IT system).

action...

Figure 1. CIS for the 2006 Winter Olympic Games in Turin. Sport selection screen (back) and Biathlon results (front).

While some of the interface and implementation details differ depending on the organizers and type of event, the CIS is typically implemented as a custom web application running on a touch-screen system. The touch-screen interface is chosen to minimize the use of peripherals, desk clutter, and simplify the use of the system. The application itself is fed through XML messages dispatched from a central server (see Figures 2 and 3). A great deal of effort and work goes into developing the IT infrastructure of a major sporting event, and the CIS back-end, as has been described in [1, 8, 14, 15], and these sources should be consulted for more detail. From the perspective of the usability evaluation we are interested in performing, the underlying implementation details are of relatively little consequence.

4. METHODOLOGY To examine the effectiveness and usability of the CIS we approached several major European broadcasters, as well as the service providers of these systems at two major sporting events: the Turin 2006 Winter Olympic Games (10-26 February 2006), and the 2006 European Athletics Championships in Gothenburg, Sweden (6-13 August 2006). While different systems were used at these two events (developed and administered by Atos Origin, [1], and ST SportService, [13] respectively), with a different set of sports in mind, the basic design and usability concepts examined in this study were comparable.

Figure 2. Event IT network diagram including the CIS. ID: 210914 DateTime: 20060223120012742 RawMessage:

Figure 3. Sample CIS message for biathlon (women’s 4x6 relay).

We received copies of, and analyzed the design and implementation documents for the Turin CIS. Additionally we were given a direct, live data-stream for our analysis. To limit the amount of work, we chose to focus this analysis on three sports: ski jumping, biathlon, and ice hockey (sports with varying degrees of CIS information and complexity, from high to low respectively). For each of these sports we did an in depth analysis of two whole transmissions. These were randomly chosen

3

television broadcasts (from US Universal network), where we tried to use live coverage whenever we could in order to avoid post-production editing. In this analysis we looked at on-screen information as well as the content of commentaries. We created transcripts for each event (see Figure 5) in which we stored the data gathered from the video, from the commentary, and from the official XML feed to the CIS. Then, we compared the availability of on-screen and commentary information to the information available over CIS. The final step in our analysis was an ethnographic study during the Gothenburg 2006 European Athletic Championships, where we observed radio commentators performing their job in the stadium, and their interactions with the CIS.

Figure 4. TV capture of biathlon event.

5.1.1 CIS Architecture and Design Because the CIS is used by a select few, no official tutorial or manual is available. To get to know the system we had to go to the implementation and requirements documents used in the development period in order to find information on the Turin 2006 CIS before it became available in February of 2006 (see Figure 6). From the available information we produced full-scale mockups and mapped the links between them in order to get a sense of the information architecture. As part of this process we noted how similar the architecture and screen layout of different sports were.

Figure 6. CIS display layout.

Figure 5. Transcript for biathlon event (women’s 4x6km relay)

5. RESULTS 5.1 CIS Analysis As part of our analysis we focused on two things, the design/layout, and the content of the system. In our study of the design, we paid more attention on the layout and presentation of information, including the number of screens available to commentators for each sport/event, as well as the navigation between those. In terms of content, we focused on determining what information was presented on what screen, and when it became available to the commentators.

Figure 7. Navigation of Biathlon Relay events.

For a competition like the Winter Olympic Games where each of the 15 sports feature an important number of events (236 total), there can be over 100 distinct screens in the CIS per sport. Figure 7 gives a mapping of the different screens available for a single event in the CIS (a Biathlon Relay event). One of the most surprising findings is the flatness of the CIS page hierarchy. Obviously designed for speed of access, it is virtually possible to navigate directly between any two pages in an event. While this may initially seem like a good design decision, this may not be an ideal choice given the mental load of the target user. The amount

4

of screen clutter and available options may in this case outweigh any gain from direct access.

5.1.2 CIS Data-stream The amount of information generated during one of these events can also be truly massive. For example, the women’s 4 x 6km biathlon relay event during the Turin 2006 Games resulted in 1,267 XML messages being generated and sent to the CIS. This information is then distributed across a total of 11 screens in the CIS. (Figure 7).

messages to the official clock to compare the data available in the CIS with the journalists’ comments. An example of this transcription work can be seen in figure 5. This transcript shows the events on-screen at the moment of the image in figure 8. The left-hand column in figure 5 contains the references to the XML messages that match the time for each action / information / comment. What we obtained from this method is detailed in the following section.

During an event like the Turin 2006 Games, a total of 260,100 messages were generated over the 23 days the system was active, and an average of 14,865 XML messages per day during the actual games. Each single message has a general format including time, event name, origin… as shown on Figure 3. Over all sports, an average event generated 1,028 messages. Table 1 shows a breakdown of messages by sport. As expected, different sports result in different amounts of traffic, as the number of competitors, duration, rounds, and measurable results vary. Table 1. CIS data by sport and event, sorted by average number of messages per event Curling Speed Skating Figure Skating Ice Hockey Freestyle Short Track Snowboard Alpine Skiing Biathlon Cross Country Ski Jumping Skeleton Nordic Combined Bobsleigh Luge Weather Other Total

Messages 34,985 5,173 2,089 36,289 2,619 7,064 5,909 12,564 16,176 28,359 12,293 10,535 19,498 23,283 25,667 1,176 16,421 260,100

Events 98 12 4 58 4 8 6 10 10 12 3 2 3 3 3

Average 357 431 522 626 655 883 985 1,256 1,618 2,363 4,098 5,268 6,499 7,761 8,556

236

1,028

5.1.3 Event Transcription Knowing the information available in the system and the average number of messages generated, the next phase was to map these messages to the commentary and on-screen information for the events we focused on. As mentioned earlier these were two complete transmissions of a ski jump, biathlon and hockey events. In this analysis we looked for one of three events: 1)

An event takes place on-screen: an athlete takes the lead, a team scores, etc.

2)

Relevant information is shown on screen: official clock, current score, biography, etc.

3)

A commentator makes a statement

Our goal was to keep track of all occurrences of these by transcribing them into Excel sheets with the intention to examine how the commentaries were related to the information on screen, the actions or both. In addition, we mapped the CIS XML

Figure 8. Specific screenshot of biathlon (women’s 4x6km relay).

5.2 Information Use The next step was to try and determine what information was conveyed to the viewer and how, and where possible, determine the origin of said information. We used our transcripts of the six randomly selected transmissions (two biathlon, two ski jumping, and two ice hockey events from the US Universal network’s coverage of the 2006 Olympics), as described earlier, sports with different CIS complexity (see Table 1). For each of these transmissions we tried to match the content of commentaries with the available CIS information and/or on-screen graphics. A total of 198 minutes of video was analyzed in this part of the project. Examining the information the CIS provides, comparing it to what it is supposed to deliver, we start to see some of the shortcomings of this implementation of the system. In Table 2 we show a sample of the errors and information gaps found in the two biathlon events. Important information is missing from the CIS stream. Not only is information missing for more than 15% of events (shooting results or partial times), but shooting results are only available after an athlete fires at all targets rather than after each shot, as is available on-screen. A random sampling of shooting messages shows that on average, these messages are put on the system 3.5 seconds after the final shot is taken, with some messages being delayed 30 seconds. Table 2. CIS information gap for Biathlon Coverage Total number of actions

212

No information

33

15.57%

Only available on TV

64

30.19%

115

54.25%

Available from either on TV or CIS

5

Biathlon

Bobsleigh

CrossCountry Skiing

Curling

Figure Skating

Freestyle Skiing

Ice Hockey

Luge

Nordic Combined

Short Track

Skeleton

Ski Jumping

Snowboard

Speed Skating

Analysis (Min) Events Comments Weather Bio Rules/Strategyy Live Results Past Results

Alpine Skiing

Table 3. Commentary by Sport

Summary

180 6 316 3% 19% 16% 51% 12%

180 4 421 3% 14% 18% 49% 16%

100 3 263 4% 16% 12% 51% 17%

123 5 215 5% 13% 16% 46% 20%

180 3 251 0% 7% 33% 51% 8%

180 2 148 0% 20% 29% 42% 9%

132 4 208 4% 19% 21% 50% 7%

180 3 308 0% 17% 19% 53% 12%

59 3 122 2% 16% 23% 52% 7%

107 3 192 3% 16% 17% 43% 20%

129 7 221 0% 15% 25% 47% 13%

21 2 60 0% 15% 12% 62% 12%

39 2 71 0% 18% 24% 44% 14%

130 4 202 4% 12% 31% 47% 5%

180 9 334 0% 19% 15% 49% 17%

1920 60 3332 2% 16% 20% 49% 13%

Our investigation clearly indicates that there were some serious quality of service problems with the Turin 2006 Olympic CIS, which is meant to function as a real-time system. Information was usually more promptly and easily available to commentators and viewers on TV monitors than on the CIS. Looking at the timing of commentary, we saw clear indications that commentators were relying more on the TV images than the CIS for live results, which accounted for almost half of all commentary. Given the shortcomings of this system, we believe it is likely that commentators at the Turin 2006 Olympic Games would have relegated the CIS to a secondary tool in favor of the live TV feed.

provider [13], for a different set of sports, the basic interface and data provided was similar enough for our purposes (see Figure 9). Waiting an additional 4 years for the next Winter Olympics did not seem a reasonable course of action. For the study we received access to the press areas of the stadium used in the event. We also received the collaboration of several major European TV and radio stations, whose commentators allowed us to observe them while doing live commentary of events, as well as interviewing them after the fact.

1,920 minutes of additional audio commentary were analyzed, in less detail, to determine whether the content of commentary seen in our sample was generalizable to a wide selection of sports. Commentary was classified into one of 5 categories; Weather (commentary regarding meteorological or venue conditions), Bio (commentary related to past achievements or personal details of athletes or coaches), Rules/Strategy (Commentary seeking to explain the sport or the rational for actions), Live results, and Past results (commentary about results from other races/heats in the same or previous Games). Our findings (Table 3) show some variability in the content of commentary as function of the sport. Obviously, commentators of some sports and events have more facts and figures to report than others. Some sports are likewise more affected by weather conditions, rule changes, or have more athletes participating. However, there is a remarkable regularity to the data, and it is possible that if the data were somehow normalized to take into account factors like whether rules were changed and the number of competitors, an even closer match could be obtained. Note that, except Curling, which is mainly based on strategy, the two sports with the most Rules/Strategy discussion were those featuring new scoring systems or new events: Figure Skating and Snowboard. To keep viewers up-to-date the, commentators have to explain the rules anew.

5.3 Field study Given the questions raised by our analysis of the Olympic CIS, we decided to conduct an ethnographic study to determine if our conclusions about the systems’ use were correct. This study also allowed us to gather first-hand accounts of CIS use and any other tools or information sources used by commentators. We decided to study the work of commentators and the CIS used at the 2006 European Athletics Championships in Gothenburg. While this was a different system developed by a different service

Figure 9. Home and result screens from the 2006 European Athletics Championships CIS.

While this implementation of the CIS may not have had the same shortcomings as that of the 2006 Olympics, we found that as we predicted, its use by commentators at this event was limited. In most cases we observed reporters initially browsing the information available in the system, but quickly switch the CIS to the “Live Results” page and leaving it there for the duration of the session. We also observed commentators relying heavily on the results and images presented on TV monitors (the same the viewers see at home), more so than the CIS results. However, the CIS was used extensively to compare results to those of previous heats or partial results from earlier in the same event, the only place such information was available. For information on previous results as well as biographical information on the athletes and coaches, we observed reporters keeping extensive and detailed paper-based dossiers with their own notes, as well as the printed version of some of the information available in the CIS (see Figure 10). This printed material, facilitated by the event organizers, offered

6

commentators the ability to add their own notes and bookmarks, making it easier to access compared to the CIS. Usually, printed pages with updated results for each new heat and round were handed out to all the journalists every 10 to 15 minutes. As a result, after half a day of commentaries, the accumulation of paper and information on a commentators’ desk was significant. We noticed that journalists usually only took a quick look at the sheets of paper and either stored these sheets under other documents or threw them away without further consideration. It seemed as if these papers were only of interest as a backup to CIS or TV failure.

Figure 10. Commentator at the 2006 European Athletics Championships.

In addition to using the CIS less than expected, journalists did not notice the new features added to the system. For this event for instance, journalists were not familiar with the streaming videos, which were an innovation of the system. We learned that journalists did not know what they were for, or how to use these features. Since these systems are only available in the stadium or other official venues, commentators had no opportunity to explore and familiarize themselves with the system before arrival. However, the stress associated with entertaining a live TV or radio audience meant that commentators were extremely hesitant about taking risks and exploring new technologies at that time. This is probably one of the main reasons why the system is not used more actively as well. The wealth of information available, together with the large number of screens available adds to the potential for confusion. Commentators therefore seem to prefer to switch the system to the one screen they know they need, and use other sources for the remaining information. It is important to note that despite these shortcomings, commentators were uniformly positive in their evaluation of the CIS system, and considered it an important, if limited tool for doing their job. Our own experience using the system taught us that though the system contained a wealth of information, we kept returning to the “Live Results” and “Start Lists”. This handful of screens containing the most relevant information to enhancing the viewing experience.

6. DISCUSSION AND FUTURE WORK The CIS is a complex yet essential tool for commentators. While we found significant shortcomings in the current generation of CIS, there is little doubt about the utility and desire for such a system to be available. All commentators which we spoke to were

in agreement on this point. What does seem obvious is that rather than attempting to maximize the types and amount of information available, the system should be more carefully designed around the information needs of commentators. One significant drawback of the design was commentators’ reliance on the “Live Results” and “Start List” information, only available on CIS, and constantly referred back to in commentary. Because this information was only available on the CIS, commentators were strongly discouraged from exploring the rest of the system and the information it offered, for fear of loosing watch of these two vital screens. Allowing users to make their own short-list of screens, or giving them access to more than one display could be potential solutions to explore. In addition to the technical and usability problems which we identified as part of this study, there are several important policy decisions which prevent the wider use of CIS. The first is that until now, these systems have only been available in the stadiums, and not until the start of the Games themselves. This gives commentators very little time to become familiar with the systems. The second is that these systems, though similar from event to event, are different enough in details like the placement of links and the naming of functions to impact familiarity and learnability. This coupled with the lack of training and familiarization opportunities means that journalists are less prepared and less confident in using the system than they should. Some journalists specifically noted the lack of available documentation and learning materials as one of the primary factors affecting their use of the system. From our observations of the note-taking by commentators it is also obvious that there is a very real need for CIS to incorporate commentators’ private notes and observations in, for instance athlete biographies. Without the ability to add their own private observations, commentators are forced to rely on paper copies, diminishing the value of the equivalent CIS functions, and the likelihood that these will be used. Another issue mentioned previously is the availability of the system in the time prior to an event. Indeed, when media are on site a few days before the beginning of an event, they can only access the systems an hour or so before they are going to be live. In the future, the CIS should be hosted on machines provided by the broadcasters rather than the organizers, thereby allowing commentators to customize the information and presentation, as well as familiarize themselves with the system functionalities and navigation. The problems we discovered in terms of the reliability and availability of information in the Turin 2006 Olympic CIS are highly significant and troubling. Our observations indicate that commentators are highly risk-averse before and during transmissions, as can be expected. Any breakdown in the CIS, or its real-time delivery of information serves to seriously undermine confidence in the application for all future events. It is therefore essential to address the reliability issues we detected. As of late there has been significant discussion around the issue of providing CIS outside of the venues. This has not been possible in the past, primarily because of the difficulty and cost associated with the real-time delivery of information to remote clients. These attitudes are becoming more relaxed, at the same time as new opportunities are arriving (more and better Internet connections, the possibility to distribute data via satellite radio, or embedded in video signals, etc.). This presents interesting opportunities for

7

small broadcasters to participate in events that would normally be cost prohibitive by providing them in their home country and home studio the same information they would enjoy in the stadium. Likewise, it opens up opportunities for sports with smaller followings, making broadcasters more likely to provide coverage, even if they don’t attend the event in person.

[3] BBC soccer Virtual Replay official website http://news.bbc.co.uk/sport2/hi/football/eng_prem/virtual_re play/3643048.stm. Accessed Jan. 11, 2007.

The same advances in infrastructure are also opening the doors for new and promising opportunities for viewers. With potential access to the same source of information as commentators, more engaging and interactive viewing experiences could be designed. The possibility for viewers to overlay records of interest to them, such as the last lap time of their favorite athlete rather than waiting for the generic list of all athletes to appear is appealing. Consider viewers rooting for the underdog, who seldom gets screen-time, and who is generally ignored by commentators. Using the CIS information is equally appealing for viewers using devices with small screens, such as mobile phones, where small and often less than perfect video quality can be significantly enhanced with relevant data either as text overlays, or in the form of custom animations or visualizations similar to those seen in PointTracker.

[5] Hawk-Eye official website. http://www.hawkeyeinnovations.co.uk/Flasharea/Hawkeye.h tm. Accessed Jan. 11, 2007.

In the future we intend to take our work down two tracks; first doing a redesign of the architecture of the current CIS system to develop a system based around commentators actual use and needs. Secondly, we intend to explore the opportunities these data-streams open for interactive TV services, and viewer control of on-screen information and visualization. While promising, we believe there are a number of significant usability questions which must be answered before this can become a reality.

7. ACKNOWLEDGMENTS The authors wish to thank the European Broadcasting Union,’s Sports Operation Group (EBUOPS), Spanish National TV (RTVE), French National TV (RTF), Swedish National TV (SVT), and the organizers of the 2006 European Athletics Championship for their help and the access they granted us. We also wish to thank the support and technical staff at the College of Engineering of Oregon State University for their help in making this research possible. Finally, we thank the people who gave us their comments on this topic before, during and after the completion of this paper.

8. REFERENCES [1] Atos Origin official website. http://www.atosorigin.com. Accessed Jan. 11, 2007. [2] Bassin, K., Biyani, S., and Santhanam P. Metrics to evaluate vendor-developed software based on test case execution results. IBM Systems Journal. Vol. 41, Issue 1 (2002), p. 1330. http://www.research.ibm.com/journal/sj/411/bassin.html. Accessed Jan. 11, 2007.

[4] Cyclops. http://news.bbc.co.uk/sportacademy/hi/sa/tennis/features/ne wsid_3001000/3001768.stm. Accessed Jan. 11, 2007.

[6] Hunter, J. Telecommunications delivery in the Sydney 2000 Olympic Games. IEEE Communications Magazine, Volume 39, Issue 7 (Jul. 2001), p. 86-92, 0163-6804. [7] International Olympic Committee. Athens 2004 Olympic Games Global television Report. http://multimedia.olympic.org/pdf/en_report_1086.pdf. Accessed Jan. 11, 2007. [8] Koulouris, J. Planning telecommunications for the Athens 2004 Olympic Games. IEEE Communications Magazine, Volume 39, Issue 7 (Jul. 2001), p. 100-104, 0163-6804. [9] NASCAR NEXTEL FanView™ official website. http://www.nextel.com/en/promotions/racing/fanview.shtml. Accessed Jan. 11, 2007. [10] Olympic Games official website. http://www.olympic.org. Accessed Jan. 11, 2007. [11] Olympic Games official website. Global Broadcast revenue. http://www.olympic.org/uk/organisation/facts/revenue/broad cast_uk.asp. Accessed Jan. 11, 2007. [12] Page, M., and Moere, A.V. Towards Classifying Visualization in Team Sports. International Conference on Computer Graphics, Imaging, and Visualisation (2006). p. 24-29. [13] ST SportService official website. http://www.st-sportservice.com/index_eng.html. Accessed Jan. 11, 2007. [14] Takizawa, Y. Telecommunications at the Nagano 1998 Winter Olympic Games. IEEE Communications Magazine, Volume 39, Issue 7 (Jul. 2001), p. 76-84, 0163-6804. [15] Toftemo, O., and Ekholdt, R. Telecommunications in the 1994 Winter Olympic Games. IEEE Communications Magazine, Volume 39, Issue 7 (Jul. 2001), p. 72-75, 01636804. [16] Vicon official website. Analysis Process. http://www.vicon.com/applications/sports.html#analysis. Accessed Jan. 11, 2007. [17] Wimbledon-IBM PointTracker http://championships.wimbledon.org/en_GB/pointtracker/pth elp.html. Accessed Jan. 11, 2007.

8