Football Dataco Ltd & Ors v Sportradar GmbH & Anor

This case concerns the sui generis database right created by European Parliament and Council Directive 96/9/EC on the Legal Protection of Databases ("the Database Directive"). More particularly it concerns database right alleged to subsist in data relating to events at football matches and their timing.

There are two actions before the court, distinguished by, amongst other things, their defendants. In the first, which I shall call the Sportradar action, the defendants are Sportradar GmbH and Sportradar AG, a German and Swiss company respectively. I will refer to them together as “Sportradar” except where I need to distinguish between them. Sportradar are engaged in providing data relating to sports events to customers, for example betting companies. They do so under their brand name Betradar. The service which they provide which is of relevance in this case is called Live Scores. Although they have customers all over the world, including in the United Kingdom, they conduct their operations abroad. Mr Henry Carr QC and Mr Hugo Cuddigan presented the case for Sportradar.

In the second action, which I shall call the Stan James action, the defendants are the betting companies Stan James (Abingdon) Limited and Stan James PLC, a British and a Gibraltarian company respectively. The claim against Stan James (Abingdon) Limited has been stayed. The third defendant to that action is Enetpulse Limited, but Enetpulse is not concerned with the issues which have been ordered to be determined at this trial, and the claim against it has been stayed as well. When I refer to Stan James, I mean Stan James PLC, the second defendant in the Stan James action. The Stan James action also concerns allegations made in relation to fixture lists, but this action is not concerned with those issues either. Mr Philip Roberts presented the case for Stan James.

The Football Association Premier League Limited (“FAPL”) and the Football League Limited (“FL”) administer professional football in England, whilst the Scottish Premier League Limited (“SPL”) and the Scottish Football League Limited (“SFL”) do the same in Scotland. Football Dataco Limited (“FDL”), the first claimant in both actions, is owned in equal shares by FAPL and FL. FDL is concerned with the marketing of data from the matches organised by FAPL, FL, SPL and SFL. PA Sport UK (“PA”) is engaged as a sub-contractor by FDL to assist in the creation and marketing of the data in issue. The data in issue is collected in a database maintained by PA. This is a large database which contains much more data than that which relates to football. The data which is entered into the PA database concerned with the matches organised by FAPL, FL, SPL and SFL is entered into the database through a special graphical user interface, or front end, called Football Live. I will refer to FAPL, FL, SPL, SFL, FDC and PA collectively as the claimants. Mr James Mellor QC and Ms Lindsay Lane presented the case for the claimants.

As I have mentioned, Sportradar conduct their monitoring of sports events, including English and Scottish football matches, from locations outside the UK. The database which they create and maintain is uploaded onto servers in Austria, from whence it is accessible everywhere, including in the UK. The question of whether Sportradar can thereby be rendered liable for primary infringement of UK database right (if it subsists) by those acts has been stayed pending a reference to the CJEU by the Court of Appeal.

This trial is concerned with the claimants’ allegation that Sportradar and Stan James are jointly liable with customers of Stan James (“punters”) for acts of database right infringement committed by punters. Sportradar are also alleged to be jointly liable with customers of another betting company, Bet365, on a similar basis. Each of these companies has a website which includes a link identified as Live Score. When the punter clicks on the link a pop-up window appears on the screen of his or her computer, from which it is possible to access the Live Scores database.

The claimants allege that Sportradar have copied data relating to some of the matches from the PA database (the Dataco matches) and included them in Live Scores. The claimants rely on time stamped data relating to goals, goalscorers, own goals, penalties, cautions/expulsions and substitutions. However Sportradar contend, and the evidence shows that, since the date of the Defence, and except in relation to televised matches, Sportradar have only supplied data relating to goals (and their timings). It will be necessary, therefore, for me to make findings in relation to the pre- and post-defence activities of Sportradar.

Database right was created by the Database Directive. It is common ground that it has been duly implemented into our domestic law, and so no one suggested there was any point in looking at anything other than the terms of the Directive itself.

The policy and objectives of the Database Directive are contained in numerous recitals, the more pertinent of which are reproduced in the Appendix to this judgment. For present purposes I set out article 7(1), which provides:

Object of protection

1.

Member States shall provide for a right for the maker of a database which shows that there has been qualitatively or quantitatively a substantial investment in either the obtaining, verification or presentation of the contents to prevent extraction and/or re-utilization of the whole or a substantial part, evaluated qualitatively and/or quantitatively, of the contents of that database.

A database is defined in Article 1(2) of the Database Directive as:

“a collection of independent works, data or other materials arranged in a systematic or methodical way and individually accessible by electronic or other means.”

Article 7(2)(a) defines extraction as:

“…the permanent or temporary transfer of all or a substantial part of the contents of a database to another medium by any means or in any form;”

In addition to being infringed by the extraction or re-utilisation of the whole or substantial parts, Article 7(5) provides that other acts are not permitted, as follows:

"The repeated and systematic extraction and/or re-utilization of insubstantial parts of the contents of the database implying acts which conflict with a normal exploitation of that database or which unreasonably prejudice the legitimate interests of the maker of the database shall not be permitted.” (my emphasis)

This is the so called “little and often” provision. The right created by Article 7 applies irrespective of the eligibility of the database in question or its contents for copyright protection (see Article 7(4)), but, as recital 45 makes clear, the right given does not in any way amount to an extension of copyright protection in “mere facts or data”.

The Court of Justice has had to rule, on more than one occasion, on the eligibility for database right protection of data generated by organisers of sporting fixtures. The first group of cases, Case C-46/02 Fixtures Marketing Ltd v Oy Veikkaus AB (Finland); Case C-338/02 Fixtures Marketing Ltd v Svenska Spel AB (Sweden); Case C-444/02 Fixtures Marketing Ltd v Organismos Prognostikon Agonon Podosfairou (Greece), was concerned with fixture lists for football matches. The Court of Justice explained that the “substantial investment” necessary for subsistence of the right could not be demonstrated by investment in the creation as opposed to the obtaining, verification or presentation of the contents of the database. The Court accordingly rejected the claims to database right on the ground that substantially all the relevant investment in fixture lists was in the creation of the data. In the Finnish case, Case C-46/02, it said this:

“32.

Article 7(1) of the directive reserves the protection of the sui generis right to databases which meet a specific criterion, namely to those which show that there has been qualitatively and/or quantitatively a substantial investment in the obtaining, verification or presentation of their contents.

33.

Under the 9th, 10th and 12th recitals of the preamble to the directive, its purpose is to promote and protect investment in data 'storage' and 'processing' systems which contribute to the development of an information market against a background of exponential growth in the amount of information generated and processed annually in all sectors of activity. It follows that the expression 'investment in … the obtaining, verification or presentation of the contents' of a database must be understood, generally, to refer to investment in the creation of that database as such.

34.

Against that background, the expression 'investment in … the obtaining … of the contents' of a database must, as Veikkaus and the German and Netherlands Governments point out, be understood to refer to the resources used to seek out existing independent materials and collect them in the database, and not to the resources used for the creation as such of independent materials. The purpose of the protection by the sui generis right provided for by the directive is to promote the establishment of storage and processing systems for existing information and not the creation of materials capable of being collected subsequently in a database.

35.

That interpretation is backed up by the 39th recital of the preamble to the directive, according to which the aim of the sui generis right is to safeguard the results of the financial and professional investment made in 'obtaining and collection of the contents' of a database. As the Advocate General points out in points 61 to 66 of her Opinion, despite slight variations in wording, all the language versions of the 39th recital support an interpretation which excludes the creation of the materials contained in a database from the definition of obtaining…

37.

The expression 'investment in … the … verification … of the contents' of a database must be understood to refer to the resources used, with a view to ensuring the reliability of the information contained in that database, to monitor the accuracy of the materials collected when the database was created and during its operation. The expression 'investment in … the … presentation of the contents' of the database concerns, for its part, the resources used for the purpose of giving the database its function of processing information, that is to say those used for the systematic or methodical arrangement of the materials contained in that database and the organisation of their individual accessibility…

44.

Finding and collecting the data which make up a football fixture list do not require any particular effort on the part of the professional leagues. Those activities are indivisibly linked to the creation of those data, in which the leagues participate directly as those responsible for the organisation of football league fixtures. Obtaining the contents of a football fixture list thus does not require any investment independent of that required for the creation of the data contained in that list.

45.

The professional football leagues do not need to put any particular effort into monitoring the accuracy of the data on league matches when the list is made up because those leagues are directly involved in the creation of those data. The verification of the accuracy of the contents of fixture lists during the season simply involves, according to the observations made by Fixtures, adapting certain data in those lists to take account of any postponement of a match or fixture date decided on by or in collaboration with the leagues. As Veikkaus submits, such verification cannot be regarded as requiring substantial investment.

46.

…

47.

It follows that neither the obtaining, nor the verification nor yet the presentation of the contents of a football fixture list attests to substantial investment which could justify protection by the sui generis right provided for by Article 7 of the directive.

49.

In the light of the foregoing, the answer to the first question referred should be that the expression 'investment in … the obtaining … of the contents' of a database as defined in Article 7(1) of the directive must be understood to refer to the resources used to seek out existing independent materials and collect them in the database. It does not cover the resources used for the creation of materials which make up the contents of a database. In the context of drawing up a fixture list for the purpose of organising football league fixtures, therefore, it does not cover the resources used to establish the dates, times and the team pairings for the various matches in the league."

The same principle, distinguishing investment in creating from that in obtaining, was held to apply in a further case, this time concerned with horse racing, Case C-203/02 British Horseracing Board Ltd v William Hill Organisation [2004] E.C.R. I-10415; [2005] E.C.D.R. 1; [2005] R.P.C. 13 (the BHB case), a reference to the CJEU from the Court of Appeal in England. The BHB case concerned a database containing, amongst other things, the list of runners and riders prepared by BHB for horse races. The BHB database contained the data relating to many thousands of entries in races in the course of a year. BHB extensively checked its data using a number of sources, including tape-recorded telephone calls to verify each entry. In its judgment the ECJ held that the 'runners and riders' database was not protected by the sui generis right:

"38.

However, investment in the selection, for the purpose of organising horse racing, of the horses admitted to run in the race concerned relates to the creation of the data which make up the lists for those races which appear in the BHB database. It does not constitute investment in obtaining the contents of the database. It cannot, therefore, be taken into account in assessing whether the investment in the creation of the database was substantial.

39.

Admittedly, the process of entering a horse on a list for a race requires a number of prior checks as to the identity of the person making the entry, the characteristics of the horse and the classification of the horse, its owner and the jockey.

40.

However, such prior checks are made at the stage of creating the list for the race in question. They thus constitute investment in the creation of data and not in the verification of the contents of the database.

41.

It follows that the resources used to draw up a list of horses in a race and to carry out checks in that connection do not represent investment in the obtaining and verification of the contents of the database in which that list appears."

When the matter returned to the Court of Appeal, the claimants sought to argue that the ECJ had misunderstood the facts: British Horseracing Board Ltd v William Hill Organisation Ltd [2005] EWCA Civ 863; [2005] R.P.C. 35. At paragraphs 28 to 31 and 35 the Judgment, Jacob LJ explained why the ECJ had rejected the claim to sui generis right:

"28.

I now turn to what I think is the flaw in [BHB's counsel's] reasoning. He starts from the beginning of the process, working down to the final, officially published, list of riders and runners. By a series of steps he says Art 7(1) databases are created by a process of gathering in and checking.

29.

But the Court has, I think, implicitly rejected that approach. It focussed on the final database - that which is eventually published. What marks that out from anything that has gone before is the BHB's stamp of authority on it. Only the BHB can provide such an official list. Only from that list can you know the accepted declared entries. Only the BHB can provide such a list. No one else could go through a similar process to produce the official list.

30.

So if one asks whether the BHB published database is one consisting of "existing independent materials" the answer is no. The database contains unique information - the official list of riders and runners. The nature of the information changes with the stamp of official approval. It becomes something different from a mere database of existing material.

31.

It is only on this basis that one can understand the crucial paragraphs in the ECJ's reasoning [37 -41].

...

34.

It is true that in [38] the word "selection" is used. Out of context this might be taken to mean something like "creative choice" but in context it clearly does not have that meaning. Other language versions of the judgment (particularly the French determination) do not have the nuance of creative choice.

35.

It follows that despite all [BHB's counsel's] ingenuity, the answer from the Court is clear. So far as BHB's database consists of the officially identified names of riders and runners, it is not within the sui generis right of Art.7(1) of the Directive. And I think the same reasoning applies in those cases (big races) where the BHB publishes a list of provisional runners prior to final declarations. Again what is published is different in character from a mere list of gathered in information. It is a list of horses that BHB have accepted as qualifying to race – as properly and actually entered"

The CJEU has, however, made it clear that the fact that the creation of the materials contained in the database was carried out by the same entity as created the database does not preclude that entity from claiming the protection of the sui generis right. So, in the BHB case, the Court said:

“35 In that light, the fact that the creation of a database is linked to the exercise of a principal activity in which the person creating the database is also the creator of the materials contained in the database does not, as such, preclude that person from claiming the protection of the sui generis right, provided that he establishes that the obtaining of those materials, their verification or their presentation, in the sense described in [31] to [34] of this judgment, required substantial investment in quantitative or qualitative terms, which was independent of the resources used to create those materials. ”

In the case of lists of runners and riders, as in the BHB case, and in the case of fixture lists, as in the Fixtures Marketing cases, it is clear that the relevant lists do not fully exist as data until they are completed by the entity with control over the sporting event. Recently, in British Sky Broadcasting Group Plc v Digital Satellite Warranty Cover Ltd (In Liquidation) [2011] EWHC 2662 (Ch), Arnold J had to consider how far this principle went. That case concerned, amongst other things, a database containing Sky’s subscribers’ details, such as name, address, telephone number, details of their Sky equipment and installation date. The defendants contended that the investment relied upon by Sky fell on the wrong side of the line because Sky created the customer data rather than obtaining or verifying it. In particular the defendants contended that the data was created at the point where it appeared in Sky’s database. Arnold J rejected that argument:

21 …. In my judgment it is not supported by the British Horseracing Board or Fixtures Marketing cases. In those cases, the key information which the defendants had taken, namely the lists of horses in races and the lists of football fixtures, had been created by or on behalf of the claimants. It is true that the lists were created using pre-existing information (such as the names of the horses, the owners and the jockeys in British Horseracing Board), but nevertheless what resulted was new information: see the judgment of Jacob LJ when the British Horseracing Board case returned to the Court of Appeal [2005] EWCA Civ 863,[2005]RPC 35 at [29]-[30] and [35]. The CJEU held that the investment in creating that new information did not count for the purposes of subsistence or infringement of database right. By contrast, in the present case for the most part Sky do not create new information when they enter a customer’s details into the Chordiant database, they simply record pre-existing information in a systematic way. It might be argued that the installation date was created by Sky, but even if it that is right it does not assist the Personal Defendants with regard to information such as the customer’s name, address and telephone number. Counsel for the Personal Defendants’ argument would substantially deprive the contents of many databases of protection, contrary to the whole purpose of the Database Directive.”

Although not expressed by the CJEU in exactly these terms, it seems to me that a reason which supports the approach adopted by the CJEU in relation to the distinction between created data and obtained data is the following. Data which is created by an individual or organisation is, in most cases at least, not available to others until it is created. There is, accordingly, no alternative source for such data. If one allows a database right to attach to data which is created by the maker of the database, the creator obtains a true monopoly in that data. Such a result would be inconsistent with the objectives of the Directive. The Directive should not be construed in a way which gives a party a monopoly in facts, such as the runners and riders or the fixture lists. On the other hand where a database consists of data obtained from sources available to the public, such as existing published data, the balance of policy considerations is different. There is (or should be) nothing to prevent the public from investing in obtaining those data themselves. The owner of a database right in data which is obtained in this way does not achieve a stranglehold on the facts. The objectives of the Directive are therefore furthered by encouraging investment in the obtaining, verification and presentation of data, without creating monopolies in facts.

The meaning to be afforded to the term “extraction” is a wide one: see the BHB case at [51] and Case C-545/07 Apis-Hristovich v Lakorda at [40]. The “decisive criterion is to be found in the existence of an act of ‘transfer’ of all or part of the contents of the database concerned to another medium, whether of the same nature as the medium of that database or of a different nature”: Apis at [41].

What amounts to a substantial part was explained in the BHB case at [69] to [71]:

“69.

In that connection, it must be borne in mind that protection by the sui generis right covers databases whose creation required a substantial investment. Against that background, Article 7(1) of the directive prohibits extraction and/or re-utilisation not only of the whole of a database protected by the sui generis right but also of a substantial part, evaluated qualitatively or quantitatively, of its contents. According to the 42nd recital of the preamble to the directive, that provision is intended to prevent a situation in which a user ‘through his acts, causes significant detriment, evaluated qualitatively or quantitatively, to the investment’. It appears from that recital that the assessment, in qualitative terms, of whether the part at issue is substantial, must, like the assessment in quantitative terms, refer to the investment in the creation of the database and the prejudice caused to that investment by the act of extracting or re-utilising that part.

[70] The expression substantial part, evaluated quantitatively, of the contents of a database within the meaning of Article 7(1) of the directive refers to the volume of data extracted from the database and/or re-utilised, and must be assessed in relation to the volume of the contents of the whole of that database. If a user extracts and/or re-utilises a quantitatively significant part of the contents of a database whose creation required the deployment of substantial resources, the investment in the extracted or re-utilised part is, proportionately, equally substantial.

[71] The expression substantial part, evaluated qualitatively, of the contents of a database refers to the scale of the investment in the obtaining, verification or presentation of the contents of the subject of the act of extraction and/or re-utilisation, regardless of whether that subject represents a quantitatively substantial part of the general contents of the protected database. A quantitatively negligible part of the contents of a database may in fact represent, in terms of obtaining, verification or presentation, significant human, technical or financial investment.”

The Court went on to point out that, as the existence of the sui generis right does not give rise to the creation of a new right in the works, data or materials themselves, the intrinsic value of the materials affected by the act of extraction and/or re-utilisation does not constitute a relevant criterion for the assessment of whether the part at issue is substantial: see [72] and [78].

Mr Mellor QC accepted, rightly in my judgment, that in the present case he could not succeed on quantitative assessment if he did not succeed on a qualitative assessment. It will therefore be necessary to focus on a qualitative evaluation of the investment in any contents of the database which are shown to have been extracted.

Finally, the “little and often” provision of Article 7(5) was explained in BHB as follows:

“86.

It follows that the purpose of Article 7(5) of the directive is to prevent circumvention of the prohibition in Article 7(1) of the directive. Its objective is to prevent repeated and systematic extractions and/or re-utilisations of insubstantial parts of the contents of a database, the cumulative effect of which would be to seriously prejudice the investment made by the maker of the database just as the extractions and/or re-utilisations referred to in Article 7(1) of the directive would.

87.

The provision therefore prohibits acts of extraction made by users of the database which, because of their repeated and systematic character, would lead to the reconstitution of the database as a whole or, at the very least, of a substantial part of it, without the authorisation of the maker of the database, whether those acts were carried out with a view to the creation of another database or in the exercise of an activity other than the creation of a database.”

Thus, in the case of extraction of an insubstantial part, it is necessary to look more widely than the individual act of extraction itself and determine whether the acts, because of their repeated and systematic character, would lead to the reconstitution of a substantial part.

The work undertaken by the claimants in gathering in the information for inputting into the database in connection with the matches was not seriously in dispute.

As almost everyone knows, most of the matches in the English and Scottish football leagues are played on Saturdays in the football season. On a typical Saturday there may be 65 live games to be covered.

The data which is entered into the database is entered by Sports Information Processors (SIPs). When recruited, SIPs are given an introductory training and provided with a Guide to show them how data is to be entered as the match progresses. SIPs are required to be good at listening and quick at data inputting. A knowledge of football is a help but not essential.

The SIPs receive data from football analysts (FBAs) who attend the football ground where a game is being played. FBAs may be ex-professional footballers. Normally they will have press accreditation and sit in the press box at the ground. The FBA has a sheet which sets out the information he is required to give in respect of each match event. He or she is in communication by telephone with the SIP for the whole match. Each FBA is provided with two mobile phones (in case there is a network or other problem with one of them). When a goal is scored, for example, the FBA has to give the SIP the number of the player who has scored and the team of that player as well as further details such as body part, the pitch position from which it was struck, the distance in yards, the net position and the goal type. The SIP will read back the details to the FBA for him to confirm that they are correct.

The data relating to goals and goalscorers is not - by a long way - the only data relayed by the FBA to the SIP. Amongst the data identified in Annex 1A to the pleadings in the Sportradar action are the shots on goal. Every shot is recorded along with whether it was on or off target, the pitch position from which it was struck, the body part used (left foot, right foot, header, other), the shot type (normal, drilled, curled, volley, overhead kick, miss hit, lob), the distance in yards and the outcome (goal, blocked, hit wall, save, hit post, hit bar, miss left, miss right, over bar, throw in, clearance, disallowed goal, open play). Other events which are subject to data collection are throw ins (attacking or defending), fouls (player committed by, player fouled, outcome of foul), free kicks (pitch position, taken by, kick type, struck with, outcome of shot), penalties (reason, taken by, struck with, shot type, net position, keeper correct, outcome), offsides (offside by, outcome), crosses (crossed by, outcome), dead balls (taken by, kick length), corners (left or right, taken by, strike type, strike foot, played to, outcome), goal assists (by, pitch position, type), bookings (issued to, reason), substitutions (player on and player off), handballs (by, outcome), dismissals (dismissal to, reason), goals (as above), saves (by, type of save), blocks (by) and clearances (by).

I was played a recording of the conversation between an FBA and a SIP during a match. Not surprisingly, given the density of the data which the FBA is required to report, the effect is in the nature of a running commentary on the match, giving rise to a plentiful harvest of statistics for inputting by the SIP.

In fact the categories of data which I have summarised above are not the only data which the FBA is required to report. As exposed during the cross-examination of Mr Steeples, who is the Football, Rugby and Cricket Manager of PA, a number of other observations are referred to in the Guide and FBA’s sheets as being part of the data collected in Football Live. For example the FBA keeps the SIP up to date with who has had most possession: those reports allow the computational front end of the database to calculate possession statistics. Moreover, every ten minutes, the FBA must report the “dominator”, which refers to a key player, and a hot spot (an area on the field where most of the action has taken place) and a dominance rating must be entered. He is also required to “rate” the crowd and the referee. These are matters of judgment for the FBA, as is the selection of a “man of the match” and “worst man” which the FBA is also required to select. The FBA also is required to record the severity of fouls (normal or bad), ten-yard offences committed, whether missed shots were close, average or wild, reasons for substitutions (injury, tactical or serious injury), whether goals were easy, normal or incredible.

The claimants estimate that this operation costs them of the order of £600,000 per season. Only investment in “obtaining and verification” is relied on. No reliance is placed on investment in presentation of the information. Although this figure was chipped away at in cross examination, I am satisfied that the overall investment in the FBA/SIP operation is substantial.

FDC and PA market their data service by saying that they are the “official suppliers for the FA Premier League and the Coca Cola Football league”. FDC states on its website that the data it supplies is “official data”. Mr Steeples accepted that the data had the FAPL and FL stamp of authority on it.

As I have explained, Football Live is not a freestanding database, but is a front end or graphical user interface through which the data discussed above can be input into a much larger PA database. It includes some 17.46 gigabytes of data in 141 million rows. The data includes other sports such as rugby union and rugby league. The data relating to football matches is not just live data discussed above, but past statistics. It also includes data relating to other leagues and international games, which are not the subject of this claim.

Sportradar’s data gathering process since the date of the defence can be described in summary as follows.

In-game data (goals, penalties, goal scorers, yellow card, red cards, substitutions) is entered or approved manually by operators. Operators monitor all matches broadcast live on television channels accessible to them. Every FA Premier League Match is broadcast live on one of these channels. If an operator misses an event on that channel, he or she will use an alternative source.

Where no live broadcast is available, operators seek a reliable, fast live internet stream. These are rarely available.

If no broadcast or reliable, fast, live internet stream is available the most common source is Sky Sports News’ live TV broadcast. On Saturday this is called “Soccer Saturday”. The claimants point out that Soccer Saturday uses data under licence from PA, although it also includes live reports from grounds.

For Football Live Matches in the Championship, English League One, English League Two and Scottish Premier League and First Division where no live broadcast or internet stream is available, only goals and times are included in Live Score.

Operators also monitor manually a list of 16 online textual sources. The claimants maintain that at least seven of these sources use data under licence from PA. However amongst these are three websites which offer in game betting and for which the PA data would be too slow. Examples of the information available were produced in evidence. This information is independent of the claimants. Whilst not comprehensive, the information is not limited to Premier League matches, but extends to other leagues as well. It follows that PA are not by any means the only source of the relevant data, and, importantly, are not the first to disseminate it.

Sportradar also use automatic monitoring of sports broadcasts and sports information published online. Lists of the websites monitored in this way are in Schedules 2-9 to the Confidential Annex to the Defence. Information from these automatically monitored sources is only incorporated into Live Score once it has been assessed manually by an operator. The claimants point out that at least ten of these sources use data licensed from PA.

Prior to the date of the defence the processes differed in that, data for red cards, yellow cards and substitutions would be included for Championship matches and below. Moreover the automatic monitoring of this data allowed automatic incorporation into Live Score. The claimants say that these sources are licensed by PA.

Sportradar do not dispute that there is a potential chain of supply from the Football Live data to their operators and thus into their Live Score database. In his opening skeleton, Mr Carr accepted that the question was not whether there was any extraction, but how much was taking place. He pointed to the fact that it is not possible for Sportradar to know whether a score which they take, say, from Soccer Saturday was a score provided by PA, or one reported by a Sky Sports reporter at the ground, independently of the claimants. He also pointed to the availability of faster online data through the betting websites I have mentioned.

There was evidence about Sportradar’s use of scouts. Until November 2011 they did not use scouts in the UK. Although they now use scouts, they do not use the scout-collected data for the Live Score service.

It is sufficient to consider the Stan James website, as no significant difference was identified with the Bet365 website. By clicking on “Live Scores” on the sports home page, a pop-up box appears. The pop-up box is hosted at the address http://www.livescore.betradar.com/?alias=stanjames. When the user opens up the pop-up box, the web browser on the user’s computer communicates with betradar.com via the internet. Within the pop-up box is a list of featured games with an adjacent magnifying glass icon. Clicking on this magnifying glass reveals more information about that game.

Whilst not apparent to the user, it is not only the information about the match to which the magnifying glass is applied which is downloaded to the user’s computer. In fact, clicking on the Live Scores button will download all the information which is accessible by clicking on any of the buttons in the pop up box. The information that is downloaded exists only in machine readable form. It is only when the user clicks on a magnifying glass that he or she will be able to see information about a match in human readable form.

The Live Scores service is provided to Stan James by Sportradar by virtue of an agreement between Sportradar and Stan James. Customers of Stan James do not pay for the service. The service is used by Stan James to attract and bind users to its website.

Much of the evidence was devoted to the fact that certain errors and corrections reported by the claimants’ FBAs to the SIPs were found to have been present in Live Score. There were two parts to this: 15 genuine errors, and a seeding exercise performed for the purposes of the case. Given the concession which I have referred to above, the significance of this evidence is greatly reduced. Moreover the evidence needs to be approached from the point of view that it is quite possible, as Mr Mellor accepts, for the same error to be made independently by two observers. For example in the case of a goalmouth scramble it may appear that player A has scored a goal, only for it to be later established that player B got the last touch. The fact that two agencies might report the player A as the scorer is not surprising. However the detailed evidence which I received showed that, in some cases, the reason for the error was specific to the FBA or SIP concerned, such as one FBA who confessed to having dropped his papers at a critical moment, and a SIP who, in her own words, experienced a “blonde moment”. Those errors do establish a chain of copying in the case of those goals.

In the end the parties were not far apart on what the evidence of these errors showed. As to the 15 genuine errors, these were collated over a period of more than three years from November 2008 to February 2012. Mr Carr accepts that 5 errors are likely to have been derived from PA data. He compares this with 440 errors made by FDC over the relevant period. Mr Mellor submits that the number of probative errors is 7.

The seeded matches were taken from the third tier of the Scottish League and the fourth tier of the English League. Whether or not deliberately chosen to be so, these were matches of relatively small public interest, and were matches for which there was unlikely to be any alternative source for live data. Of the three matches deliberately seeded with an error, all were reproduced by Sportradar.

Sportradar submit, supported by Stan James, that the data identified in the annexes to the pleading are not existing data, but are created while the match is going on. They submit that the investment relied upon is investment in the creation of those materials, which are “mere facts” within recital [45]. They submit that, in order to satisfy the requirement for database subsistence, the materials in the database must be independent materials and pre-exist in material form. Until recorded by FDL’s representatives, they submit, the materials do not exist.

It is correct that, in general, this case concerns data which does not pre-exist in material form. The data relied on is captured by observation at the matches by the FBAs and recorded when it is input into the database by the SIP. Moreover, unlike fixture lists or lists of runners and riders, it is data over which the organisers of the match have no control. The organisers merely provide the environment in which the observations may be made. Does such data fall to be treated in the same way as the horse racing and fixture lists examples, or is the investment in its collection to be regarded as investment in obtaining the data?

This issue has not been the subject of any decision to which my attention was drawn. It has, however, attracted academic interest. An article to which I was referred by Davison and Hugenholtz, Football Fixtures, Horseraces and Spin-Offs: The ECJ domesticates the Database Right: [2005] EIPR 113, suggests that collected and recorded data are “created”.

“Meteorological data ... are records … of natural phenomena, not the phenomena themselves, and it would be difficult for scientists to argue that they have simply collected the data rather than creating them.”

The view that such data are created rather than collected is however questioned by Dr Estelle Derclaye in Databases Sui Generis Right: Should we Adopt the Spin-Off Theory? [2004] EIPR 402.

The spin-off theory, which both these articles mention, refers to the notion that organisations which produce data as part of their mainstream activities are undeserving of database right. The theory relies on the fact that the purpose of the Database Directive is to encourage investment of the relevant kind. Organisations which produce data in the course of their main activities do not require any encouragement to invest. The theory is discussed in the Opinion of the Advocate General in the BHB case, but does not form part of the reasoning of the Court. No such widely stated theory is, therefore, part of EU law.

A number of examples were discussed in argument. Mr Carr gave the example of a bus stopping three times in the Gray’s Inn Road. If those facts are recorded, he submits, the relevant materials are created. Unless the facts are recorded, there are no such materials.

I think this approach places too much weight on the term “materials” in the Directive and in the judgments of the CJEU. The term “materials” is used to signify the various things (to adopt a neutral term) which may go to make up the contents of a database. Recital [17] indicates that these things may include “facts”. There is no a priori reason why facts have to be recorded in order to exist.

I think there is a fundamental difference between runners and riders in a horse race and fixture lists, on the one hand, and goals and other sporting results recorded at a live match on the other. It is perfectly rational to say that the former lists are created by the organisers of the events. But the organisers of a football match do not create the goals: that is the province of the footballers. The organisers do no more than provide the environment in which the goals can be scored. They have no control over whether goals are scored or not.

In my judgment, factual data which is collected and recorded at a live event such as a football match about events outside the control of the person doing the collection and recording is not created by that person, but is obtained by him. I say so for a variety of reasons.

Firstly, collecting and recording data falls, in my view, squarely within the ordinary meaning of the term “obtaining”, but somewhat uncomfortably within the meaning of “creating”. Secondly, in determining whether there is creation involved in the investment relied on, it seems to me that the focus of the enquiry is whether the investment involves creation of new information: see in particular the BHB case on its return to the Court of Appeal, and Arnold J in British Sky Broadcasting. Recording existing facts is not creating new information. As Mr Mellor put it, the fact of the goal is created when the ball hits the back of the net. Thirdly, none of the policy objections which applied to the data in Fixtures or BHB cases apply. Subject to arguments about the restrictions placed on observers on the facts of the present case, in principle others are free to collect and record the observed data for themselves. There is nothing in the way in which the data comes into existence which prevents others from obtaining it without reference to the database. Fourthly, as recital [17] makes clear, data includes facts: a fact is as much a fact before as after it is recorded. Finally, I think it would be introducing a gloss on the wording of the legislation to read it as requiring that the materials collected in the database must pre-exist in material form, as opposed to merely pre-exist. The notion of a requirement for material form is one well known to those who draft legislation in this general field, for example in copyright.

Clearly some of the data which is input into the database through Football Live is not a pre-existing fact, but depends entirely on the judgment of the FBA. Take for example “worst man”: that data is created by the FBA. Others may have their own view as to the identity of this unfortunate person, but until the FBA states his opinion, that piece of data does not exist. I wondered, at one point, whether this was just an extreme example of something which applied to a greater or lesser degree to all observations made by the FBA. He or she could of course be wrong about the identity of a goal scorer, or indeed the fact that a valid goal had been scored. The data communicated to the SIP was only his or her view of the correct goal scorer, and the validity of the goal. Taking this approach, all the data, to a greater or lesser degree, simply represented the opinion of the FBA, which by definition is created by the FBA.

I think the flaw in this argument is that, whilst the initial information provided to the SIP may only represent the view of the FBA, the information is subsequently confirmed as correct. For example the validity of the goal will be confirmed by the match official ordering a new kick off, the identity of the goal scorer by an announcement or by being shown on a scoreboard. These are the things which will ultimately find their way into the database, and are obtained rather than created by the FBA. For example the FBA is not going to say: “do not record the goal because in my opinion it was offside, even though the game has re-started from the centre spot”.

I think this argument is also somewhat at odds with Sportradar’s position in relation to the same information which was at one stage relied on by the claimants as giving rise to copyright protection. That position was that the matters relied on – goals, goal scorers etc. - were pure matters of fact, involving no intellectual creativity on the part of the FBA. The present proposition, that these are all to some degree matters of opinion, is not easily reconciled with that position.

Mr Carr also submitted that the data were created because they were marketed as the official data of the leagues. By analogy with the BHB case, he submits that giving the data this official stamp of approval is analogous to BHB’s stamp of approval of the lists of runners and riders. I do not think this analogy works. The stamp of approval in the present case does not alter the character of the information in the same way as in BHB. There is no added information content of a relevant kind. As Mr Mellor pointed out, no doubt the Sky data in British Sky Broadcasting v Digital Satellite Warranty Cover (supra) was official, but it would be odd if that disentitled their database to protection on that ground.

I have mentioned the figures relied upon by the claimants for the annual investment in Football Live. I conclude that the investment, even if trimmed to the extent which Mr Carr suggested it should be, is substantial. It is wholly separate from the investment in the organisation of the leagues. However, I should record that no real attempt was made by the claimants to break these figures down so as to indicate how much of the investment was necessitated by the need to have an expert view on the wide variety of more or less subjective matters I have referred to. Even more importantly, there was no attempt to establish how much investment would be necessary simply for the recording of goals and timings. This is a matter to which I will have to return when considering the issue of substantial part.

I conclude that database right subsists in that portion of the PA database which contains data input through the Football Live front end.

Sportradar submitted that the relevant comparison was between the whole of the PA database and the part downloaded by the user when the pop-up window is activated. The evidence showed that it made technical nonsense to speak of a Football Live database. In technical terms there is only one database: there is no separate module for those matches which are included in this claim. The answer to the question of how substantial a part had been taken might receive different answers depending on whether one takes the whole PA database or only that part concerned with the relevant matches. This raises a problem which is acute where a case is sought to be made based on a quantitative evaluation of substantial part. As I have already indicated, however, Mr Mellor places no independent reliance on such an evaluation. Accordingly, it is not necessary for me to decide the point.

Sportradar and Stan James submitted that the only data which has to be considered was the data in relation to a single match highlighted in the pop-up box by means of the magnifying glass, and not the data in relation to all matches (not merely football) which was automatically downloaded when the punter pressed Live Scores. They stressed the fact that until the data is displayed in this way, it is not comprehensible to anyone.

In my judgment the data which falls to be considered is the data which has been transferred to the punter’s computer. I do not think that any other conclusion is open to me given the definition of extraction in the Directive. No doubt it would be possible for Sportradar to arrange matters differently, so that data in relation to one match is downloaded only when that match is selected. But that is not how the system works, and the evidence showed that the method in fact adopted would have the advantage that the speed of response would not be dependent on the speed of the link between the computer and the source of the data.

The principal question is whether the data extracted by the user in this way is qualitatively a substantial part. The data involved, prior to the defence, is goals, goal scorers, times of goals, cards and substitutions. Subsequent to the defence the data involved is limited to goals and times of goals for matches which are not televised.

Although, as I have said, there is no dispute that there existed a channel from PA’s database through which some data flowed through Sportradar to the computer of a punter, the evidence is far from precise as to how much of the data in relation to the matches on a given day will have flowed down that channel.

I find that in relation to televised matches the amount of data that has flowed from PA to the punter is minimal. It is true that the claimants have established one case of an error in the case of a televised SPL game. Nevertheless I accept the evidence of Sportradar’s witnesses that, in the case of televised matches, the facts are collected by watching the relevant broadcast, independently of PA.

In relation to other matches, I have to consider both the larger collection of data (goals, own goals, penalties, scorers, cards, expulsions and substitutions) which Sportradar collected before the defence, and the slimmed down collection (goals and timings) thereafter.

I think that the larger collection of data for non-televised matches was, on balance, likely to have used sufficient data derived from the PA database to amount to a qualitatively substantial part. I reach that conclusion on the basis that the available sources other than PA for that data were limited, and that it is realistic to regard the investment involved in obtaining a body of data of that nature as substantial, despite the fact that very much more data was in fact contained in the database. The range of data is, in my judgment, sufficiently wide to require an investment of the same order as that required for the full data collection operation. It is fair to say that Sportradar did not press very hard the suggestion that this would not amount to a substantial part. That is perhaps reflected by the fact that they had ceased to use this data after these proceedings were commenced.

The position is different, however, when one considers only the goals and timings. Mr Mellor submitted that the same set up would be necessary to collect even that amount of data, and that therefore the investment would be exactly the same. I reject that submission. It is unsupported by any evidence adduced by the claimants. It is entirely plausible that, with modern communications, the claimants could arrange for each goal scored and its timing to be recorded at a central data centre at virtually no additional cost. There would be no need for FBAs with football experience or SIPs. There would be no need for the one-to-one running commentary which the FBAs provide. Accordingly, even if every goal included in the data extracted by a punter was derived from the claimants’ database (which is not by any means established), I would hold that the data so extracted would not be sufficient to amount to a substantial part.

Without formally abandoning it, Mr Mellor recognised that this part of the claimants’ case faced formidable difficulties. He was right to do so. To begin with, there was no evidence that individual punters were extracting the data repeatedly and systematically. Such evidence as there was shows, as one might expect, that such usage was entirely ad hoc. Equally unsurprisingly there is no evidence that an individual user is accumulating data entries so as to reconstitute a substantial part of the original database. It follows that this aspect of the case must fail.

The primary acts of infringement relied upon are by punters using the Stan James and Bet365 websites. Stan James are said to be jointly liable with their customers. Sportradar are said to be jointly liable with those customers and with those of Bet365.

The law on accessory liability has recently been comprehensively summarised by Arnold J in L’Oreal SA v eBay International AG [2009] RPC 21. Neither side challenged any aspect of that summary. For present purposes both parties have focussed on the way the test was put by Peter Gibson LJ in Sabaf SPA v Meneghetti SPA [2002] EWCA Civ 976; [2003] RPC 14 at [59]:

“The underlying concept for joint tortfeasance is that the joint tortfeasor has been so involved in the commission of the tort as to make himself liable for the tort. Unless he has made the infringing act his own, he has not himself committed the tort.”

In that case the defendant had supplied goods ex works in Italy to MFI and made the transport arrangements. The Court of Appeal held that this was not sufficient to render the defendant liable for MFI’s infringing acts, as it was merely acting as a supplier of goods to a purchaser who was free to do what it wanted with the goods.

I consider first the joint liability of Sportradar with punters. The allegation of joint tortfeasance is based on the fact that Sportradar “supply” the relevant data by making it available on their servers in Austria, and accordingly have control over what the punter extracts onto his computer. On the facts which I have held and in relation to the data involved prior to the defence but not subsequently, the punter’s act of extraction will inevitably amount to infringement of the database right. Nevertheless I am not persuaded that Sportradar are thereby joint tortfeasors with the punter. The mere making available, abroad, of the means whereby a right may or will be infringed is not sufficient to make the supplier a jointfeasor. Something more is required. I do not think that it is realistic to say that by making data available in Austria, Sportradar make the act of extraction of data in England their own. I bear in mind that they have agreements with Stan James and Bet 365 for information to be made available, but what those companies do with the permission so granted is a matter for them. The case is no stronger than that of a foreign supplier who supplies infringing goods ex works in Italy: see Meneghetti above.

Next I need to consider Stan James. They provide the link on their website, and therefore to some degree directly encourage their customers to use the Live Scores service, albeit from their base in Gibraltar. They do so because it encourages punters to use their website: it binds them to it. Again, on the facts which I have held, prior to the defence, the act of extraction by the punter will inevitably result in infringement. On this basis Mr Mellor submits that there is procurement of the infringement.

Mr Roberts submitted that these facts did not establish anything more than facilitation of the infringement. He likened the situation of his clients to that of an intermediary such as eBay whose services are used for the commission of torts but who does not commit them. He drew attention to the distinction drawn in the eCommerce Directive, Directive 2000/31/EC of the European Parliament and of the Council, between a genuine tortfeasor and an intermediary whose services are to be used by a third party to infringe a right.

I think that these facts are sufficient to render Stan James liable as a joint tortfeasor. The link which they provide on their site is directed at their customers and encourages their customers to use the Live Scores service. They are not a mere intermediary, like eBay, whose services are used by a third party to infringe a right. Rather, in pursuance of a common design with their customers, they are offering a facility which is not essential to their service to their customers, but which is being used indirectly to promote and enhance that service. In so doing I consider it is right to say that they adopt the acts of extraction which their customers will perform and make them their own.

Stan James and Sportradar both submitted that to the extent that database right subsists in the database relied on, it is rendered unenforceable by the EU law doctrine of abuse of rights (abus de droit). According to this doctrine, EU law must be interpreted and applied so as to avoid conferring legitimacy on abusive behaviour. In Case C-110/99 Emsland-Stärke [2000] ECR I 11569 at paragraphs 52-53, the Court of Justice identified two essential conditions for application of this doctrine:

“52.

A finding of an abuse requires, first, a combination of objective circumstances in which, despite formal observance of the conditions laid down by the Community rules, the purpose of those rules has not been achieved.

53.

It requires, second, a subjective element consisting in the intention to obtain an advantage from the Community rules by creating artificially the conditions laid down for obtaining it. ...”

Mr Roberts submitted that FDC are seeking to exert a stranglehold over dissemination of information concerning events taking place on the pitch. He submitted that from their unparalleled position of control over the football clubs and stadia, they have attempted to put in place their own ‘single source’ data recording system to commodify something which was formerly public property; impose a blanket restriction on the right of every other person attending each match from recording or transmitting data relating to that match or any aspect to it; and assert their purported rights against everyone reporting on in-match events.

Despite the attractive way in which the argument was presented, I am entirely satisfied that this doctrine has no application in the present case. Whatever the defendants may think of the attempts by the claimants to place restrictions on the use of data collected in the way I have described, I do not accept that it is correct to say that the claimants have the subjective intention to create artificially the conditions laid down for obtaining the advantage of the Database Directive. I do not consider that there was anything artificial in any of the underlying facts which I have found which support the subsistence and infringement of the right. Furthermore there is no evidence that it was the subjective intention of any of the claimants to create an artificial situation. No such suggestion was put to any witness. It follows that this EU law defence fails.

Sportradar and Stan James also argued that the enforcement of the database right in the present case would amount to an unwarranted interference with the freedom of expression protected by Article 10(1) of the European Convention on Human Rights. This case is based in part on the proposition that no alternative source for the data in question is available. That is contrary to my factual findings in the present case. Whilst the claimants no doubt take steps to impose restrictions on the in-game reporting of match data, the evidence suggests that these steps are less than effective. Moreover, whilst there are games where the claimants are the only source of data, that is not because they are successfully enforcing restrictions on reporting: it is because there is relatively little interest in the game. In these circumstances, the claimants are not, by their assertion of database right “preventing reportage or dissemination of any meaningful information relating to incidents on the pitch” as Stan James assert in their opening written submissions. I am not persuaded that the exercise of the rights which I have held to exist in the present case is an unwarranted interference with the right to freedom of expression.

Sportradar suggested that if I came to the conclusion that database right subsisted, then I should refer a question to the CJEU. I do not propose to do so. I accept that there is scope for argument on the issue of subsistence. However, on the remaining issues, I have held that Sportradar are not liable. I am reluctant to burden the parties with a further reference to the CJEU on the basis of the historical liability of Stan James.

It follows that: