Wolfire Games LLC et al v. Valve Corporation

Filing 391

ORDER (Redacted) re Plaintiffs' 181 SEALED MOTION to Certify Class and Appoint Co-Lead Class Counsel, Plaintiffs' 256 Redaction to Motion to Certify Class and Appoint Co-Lead Class Counsel, Valve 's 232 SEALED MOTION to Exclude Testimony of Steven Schwartz, Ph.D., and Valve's 321 Redaction to Motion to Exclude Testimony of Steven Schwartz, Ph.D. The Court GRANTS Plaintiffs' motion for class certification and the appointment of co-lead class counsel, Dkt. No. 256 , DENIES Valve's motion to exclude, Dkt. No. 321 , and ORDERS as follows: (1) The class detailed herein is certified. (2) Wolfire and Dark Catt are appointed class representatives; and (3) The firms listed herein are appointed Co-Lead Class Counsel. Signed by Judge Jamal N Whitehead. (KRA)

1
2
3
4
5

UNITED STATES DISTRICT COURT
WESTERN DISTRICT OF WASHINGTON
AT SEATTLE

6
7

In re VALVE ANTITRUST LITIGATION

8

CASE NO. 2:21-cv-00563-JNW
ORDER

9
10
11
12
13

1. INTRODUCTION
This matter comes before the Court on Plaintiff Wolfire Games, LLC, Dark

14

Catt Studios Holdings, Inc., and Dark Catt Studio Interactive LLC’s (together,

15

“Plaintiffs”) motion for class certification and the appointment of class counsel, Dkt.

16

No. 2561, and Defendant Valve Corporation’s motion to exclude expert testimony,

17

Dkt. No. 3212. The Court has considered the papers submitted in support of and

18

opposition to the motions, and being otherwise fully informed, finds oral argument

19
20

1 This document contains Court-approved redactions. See Dkt. Nos. 198, 236, 237

21

(motion to seal and resulting order). A sealed, unredacted version is located at
Docket Number 181.

22

2 This document also contains Court-approved redactions. See Dkt. Nos. 294, 330,

23

331 (motion to seal and resulting order). A sealed, unredacted version is located at
Docket Number 232.
ORDER - 1

1

unnecessary. For the reasons stated below, the Court GRANTS Plaintiffs’ motion

2

for class certification and DENIES Valve’s motion to exclude.
2. BACKGROUND

3
4
5
6
7
8
9
10
11
12
13
14
15
16
17

2.1

Valve is the uncontroverted market leader in digital PC game
distribution.3
Video games are big business. Early in their history, PC games were

distributed on physical media and purchased at brick-and-mortar retail locations,
but now they’re distributed digitally. Valve, through its Steam Store and Steam
platform (collectively, “Steam”), is the largest distributor in the world, offering over
50,000 titles for sale and play. See, e.g., Dkt. No. 127 ¶ 52. Valve started out
developing its own games, id. ¶¶ 46–54, but it is no longer primarily a developer. Id.
It leaves this task to others, ranging from small independent developers to largescale developers (“AAA” studios) like Electronic Arts and Epic Games. Id. ¶¶ 24,
136, 298, 355, 358. Valve then sells and distributes those games to end users
through its Steam platform and provides other game-related services. Id. ¶¶ 2, 3.
From the end-users’ perspective, the process works like this: A gamer creates
a Steam account and buys a video game through the Steam Store.4 Id. ¶¶ 11, 16, 40,
49. The gamer then downloads the game, uses Steam to launch it on their computer,

18
19
20
21
22
23

3 These facts are from Plaintiffs’ Consolidated Second Amended Class Action

Complaint (“SAC”) and are largely uncontested. Compare Dkt. Nos. 127 at 4–42
(SAC), with Dkt. Nos. 37 at 9–11 (Valve’s motion to dismiss prior complaint
containing similar allegations), 74 at 8–9 (second motion to dismiss), 319 at 1–5
(Valve’s brief opposing class certification).
4 In limited instances, the gamer may instead acquire a game’s Steam Key—a

unique alphanumeric code which Valve generates at no charge and provides to
gaming companies—either for a fee or for free through a promotion. The Court
discusses Steam Keys in more detail below.
ORDER - 2

1

and begins play. Id. As they play, Steam provides various features, including multi-

2

player functionality, in-game downloadable content (“DLC”), access to a gaming

3

community, and the streamlined delivery of technical updates. See, e.g., id. ¶¶ 48,

4

55, 236. Based on the availability and performance of these features, Steam has a

5

large and loyal following. See id. ¶¶ 103, 136, 138. Given Steam’s popularity, many

6

gamers have developed their game libraries, over years, on Steam’s platform. See id.

7

With few exceptions, Valve does not charge gamers for Steam—instead,

8

developers and publishers foot the bill through a commission for each game sold via

9

the Steam Store.5 See id. ¶ 9. Valve initially set its commission rate at 30 percent

10

and held it steady for years. Id. It now offers discounts for the best-selling games,

11

reducing the commission to as little as 20 percent. Id. ¶ 89 n.29. Game companies,

12

in turn, set the price for their games—whether that be $5.99 or $59.99—Valve

13

remits this amount, less its commission and fees. See id. ¶ 4. Regardless of the price

14

and resulting commission paid to Valve, it is the game company that decides how

15

much of the commission to pass on to the gamer. See, e.g., ¶ 328.

16

2.2

17
18
19
20

Plaintiffs allege Valve imposes a Platform Most Favored Nations
policy, which gives rise to this antitrust lawsuit.
Plaintiffs allege that Valve requires content and price parity for any game it

distributes; this means Valve will not sell or support a game if a game company
sells it for less or offers a better version of it elsewhere. See, e.g., id. ¶¶ 159, 164,
179, 195. Plaintiffs call this a Platform Most Favored Nations (“PMFN”) Policy. Id.

21
22

5 “Developer” and “publisher” carry different meanings within the PC gaming

23

industry but Plaintiffs use the terms interchangeably in the SAC. See Dkt. No. 127
at 4 n.1.
ORDER - 3

1

¶¶ 154–84. Plaintiffs argue Valve implements the PMFN Policy to maintain its

2

market control. Id. ¶¶ 156–94. And Plaintiffs contend its effect is clear and largely

3

consistent: the industry knows that if you undercut Steam you pay for it. See, e.g.,

4

id. ¶¶ 83–90, 165–84, 198–205.

5

If a game company violates the PMFN Policy, Valve may start with a simple

6

conversation. See, e.g., id. ¶¶ 208–09. From there it turns to more punitive

7

measures. Valve is known to remove violating games from its in-platform marketing

8

efforts and, sometimes, delist them from the Steam Store altogether. See, e.g., id.

9

¶¶ 146, 194, 195, 210, 202, 219, 233–36. Even AAA studios are not immune. See id.

10

¶ 136. Given the breadth of Steam’s gamer base, ostensibly all game companies rely

11

on Steam and cannot afford to lose access to its customers. Id.

12

Plaintiffs assert that Valve’s PMFN Policy has the following anti-competitive

13

impacts, among others: (1) Steam earns a supracompetitive commission, (2) game

14

companies cannot compete between distribution platforms, and (3) rival platforms

15

cannot succeed. Id. ¶¶ 258–374. This harms consumers and game companies alike

16

by increasing consumers’ prices and reducing publishers’ profits (among other

17

anticompetitive ills). Id. Plaintiffs assert this violates Sections 1 and 2 of the

18

Sherman Act, 15 U.S.C. § 1 et seq., and Washington’s Consumer Protection Act

19

(“CPA”), RCW 19.86 et seq. Id. ¶¶ 389–436. Plaintiffs seek class-wide remedies and

20

propose a class encompassing the following:

21
22
23

All persons or entities who, directly or through an agent, paid a
commission to Valve in connection with the sale or use of a game on the
Steam platform on or after January 28, 2017, and continuing through
the present until the effects of its scheme are eliminated (the “Class
Period”), and where either (1) the person or entity was based in the
ORDER - 4

United States and its territories or (2) the game was purchased or
acquired by a United States-based consumer during the Class Period.
Excluded from the Class are (a) Defendant, its parents, subsidiaries,
affiliate entities, and employees, and (b) the Court and its personnel.

1
2
3
4

Id. ¶ 375.
Valve strongly denies all but a few of Plaintiffs’ allegations. See, e.g., Dkt. No.

5
6

128 ¶¶ 154–84.
3.

7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

3.1

DISCUSSION

Valve’s motion to exclude Plaintiffs’ expert’s testimony.
Because Plaintiffs’ motion for class certification relies heavily upon the work

of their expert Steven Schwartz, Ph.D., the Court addresses Valve’s Daubert motion
to exclude before turning to class certification.
According to Dr. Schwartz, Valve is the largest PC video game distributor in
the world. Its market share exceeds that of all other platforms combined. See Dkt.
Nos. 182-1, 343 ¶ 130.6 While Valve largely reached this position through
competitive practices, Dr. Schwartz concludes that it maintains its dominance
through its PMFN Policy. Dkt. No. 343 ¶¶ 130–72. And without the PMFN—that is,
in a “but-for world”—competitive pressures would drive Valve’s market share down,
forcing it to reduce its commission structure to maintain any market share. Id.
¶¶ 274, 279, 375, 377. Valve contends Dr. Schwartz’s opinion is based on faulty
methods and improper assumptions and thus does not satisfy Rule 702’s admission
standards. See generally Dkt. No. 321. On this basis, it moves to exclude. Id.

6 Docket Number 182-1 is a sealed unredacted version of Docket Number 343. See

Dkt. Nos. 294, 330, 330-1 (motion to seal and resulting order).
ORDER - 5

1

“Rule 702 of the Federal Rules of Evidence tasks a district judge with

2

‘ensuring that an expert’s testimony both rests on a reliable foundation and is

3

relevant to the task at hand.’” Hyer v. City & Cnty. of Honolulu, 118 F.4th 1044,

4

1055 (9th Cir. 2024) (quoting Elosu v. Middlefork Ranch Inc., 26 F.4th 1017, 1024

5

(9th Cir. 2022)). The district court has “broad discretion” in rendering such

6

evidentiary rulings. Id. (quoting City of Pomona v. SQM N. Am. Corp., 866 F.3d

7

1060, 1065 (9th Cir. 2017)).

8

Rule 702 guides the Court’s discretion. Opinion testimony from an expert is

9

allowed if: (1) their scientific, technical, or other specialized knowledge will help the

10

trier of fact to understand the evidence or determine a fact in issue; (2) the

11

testimony is based on sufficient facts or data; (3) the testimony is the product of

12

reliable principles and methods; and (4) they have reliably applied the principles

13

and methods to the facts of the case. FED. R. EVID. 702. The rule is to be construed

14

liberally. United States v. Hankey, 203 F.3d 1160, 1168 (9th Cir. 2000). And the

15

resulting “inquiry is a flexible one,” with a focus on principles and methods, rather

16

than conclusions reached. Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579, 595

17

(1993).

18

Valve challenges Dr. Schwartz’s methodology, not his qualifications, arguing:

19

(a) Dr. Schwartz applied an incorrect method as matter of law by analyzing Valve

20

as a one-sided rather than two-sided platform; (b) Dr. Schwartz assumed a PMFN

21

price parity requirement without sufficient evidence, i.e., the analytical gap

22

between his opinion and the evidence is too great; (c) he impermissibly speculated

23

when determining Valve’s but-for market share; (d) his analysis is incomplete in
ORDER - 6

1

that it ignores the import of Steam Keys; and (e) his passthrough determination is

2

methodologically unsound. Dkt. No. 321 at 2–3.

3

The Court takes each argument in turn.

4

3.1.1

5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

Dr. Schwartz accounts for Steam as a two-sided platform.

According to Dr. Schwartz, a platform “facilitates interaction between two or
more groups of users, like consumers and producers.” Dkt. No. 343 ¶ 30. It can be
one-sided or two-sided. Id. A platform is one-sided if it “connects groups of users
directly to people like them,” and it is two-sided if it “connects two or more different
groups in a . . . virtual space.” Id. Two-sided platforms often “exhibit network
effects,” where “the product or platform’s value changes as the number of users of
that product or platform changes.” Id.
According to Valve, Ohio v. Am. Exp. Co., 585 U.S. 529 (2018) (“AMEX”), the
network effects of antitrust conduct involving a two-sided platform must be
analyzed on a two-sided basis. Dkt. No. 321 at 3–4. Valve argues that Dr. Schwartz,
by utilizing his Platform Competition Model (“PCM”) to deduce solely the reduced
commission game developers and publishers would pay in a but-for world (without a
similar analysis of consumer effects), see Dkt. No. 343 ¶¶ 272–77, applied an
incorrect methodology as a matter of law. Dkt. No. 321 at 3–6. This critique is not
persuasive.
In AMEX, the district court permanently enjoined American Express
(“Amex”) from engaging in anticompetitive conduct that allegedly drove inflated
merchant fees (i.e., the service fee charged to a merchant for a credit card sale). See
United States v. Am. Exp. Co., 88 F. Supp. 3d 143, 199, 238–39 (E.D.N.Y. 2015). The

ORDER - 7

1

Second Circuit reversed and the Supreme Court affirmed. See Ohio v. Amex, 585

2

U.S. at 552, aff’g, 838 F.3d 179 (2d Cir. 2016). In so doing, the Supreme Court held

3

that a credit card is, ostensibly, a two-sided platform with merchants on one side

4

and consumers on the other. Id. at 535. And that both sides of a card’s network

5

effects—increased value to cardholders when more merchants accept it and

6

increased value to merchants when more cardholders use it—must be accounted for.

7

Id.

8

As Dr. Schwartz explains, his PCM utilized sales commissions as a proxy for

9

antitrust injury in quantifying class-wide damages because this reflects Valve’s

10

historic pricing practice7— an agency model imposing a one-sided charge. See Dkt.

11

Nos. 343 ¶¶ 338–40, 348-1 ¶¶ 146–52. The PCM is not intended to model a complete

12

but-for world. See Dkt. No. 348-1 ¶¶ 146, 208–10. Indeed, it is one of three models

13

Dr. Schwartz utilized to assess antitrust impact(s). See Dkt. No. 343 ¶¶ 284–95,

14

302–13. He also employed a yardstick approach, comparing the PC game delivery

15

platform to those for online retail and vacation home rentals, and an empirical

16

analysis of existing market data. Id. Both approaches address consumer impacts.

17

Id. Also worth noting—Dr. Schwartz’s method is based on the same Landes and

18

Posner research that the Supreme Court relied on in Amex. Compare id. ¶¶ 334–37,

19

with Amex, 585 U.S. at 535–36.

20
21
22
23

7 On reply, Valve offers Steam’s “community market” as an example of a transaction

where consumers pay fees. Dkt. No. 363 at 6. But that market occurs solely between
consumers, i.e., on one side of the platform. See Dkt. No. 383-1 ¶ 27. Dr. Schwartz
did not err in ignoring it.

ORDER - 8

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

3.1.2

The analytical gap between antitrust injury and common
proof is not large.

Dr. Schwartz’s opinion is, necessarily, predicated on the existence of Valve’s
PMFN Policy. See Dkt. No. 343 ¶¶ 154, 197. Valve suggests that his report cites
insufficient common facts or data to support its existence, at least with respect to
the alleged price parity component. Dkt. No. 321 at 6–10. Thus, there is too great an
“analytical gap” between Dr. Schwartz’s ultimate opinion and common proof. Id. at
7 (citing Gen. Elec. Co. v. Joiner, 522 U.S. 136, 146 (1997)). Or put differently, Dr.
Schwartz relies on unreasonable inferences and assumptions not supported by
common proof. Id. at 7–9. The critique in inapt.
Valve allegedly cloaks its PMFN Policy to guard against antitrust actions.
See Dkt. No. 182-22 at 2. Therefore, Dr. Schwartz set about deducing its existence
and Valve’s enforcement efforts. See Dkt. No. 343 ¶¶ 154–96. He provides evidence
supporting his reasoning, citing to (1) contractual terms that video game companies
agree to before receiving Steam Keys, which suggest a price parity requirement;8 (2)
examples of one-on-one communications between Valve and video game companies
explicitly describing the requirement;9 and (3) a variety of other messaging,

17
18
19
20
21
22
23

8 Those terms require, among other things, that game companies sell their games

“‘[i]n other stores in a similar way to how I am selling my game on Steam,’” and that
discounts they offer outside the Steam Store be offered to “‘Steam customers within
a reasonable amount of time.’” Dkt. No. 343 ¶ 157 (quoting Steam Key Guidelines).
9 This includes communications from Valve that “‘the price on Steam [must be]

competitive with where it’s being sold elsewhere’” and that Valve “‘wouldn’t be OK
with selling games on Steam if they are available at better prices on other stores,
even if they didn’t use Steam keys.’” Dkt. No. 343 ¶ 158, 160 (quoting emails
produced at VALVE_ANT_0598921, 0605087).

ORDER - 9

1

including internal communications at Valve and other game companies, suggesting

2

the same. Id. ¶¶ 157–67. This is common evidence of the varied ways that Valve

3

allegedly communicates its PMFN Policy. Valve may disagree with the conclusions

4

Dr. Schwartz’s derived from this information and it may even offer countervailing

5

evidence, but Valve cannot render Dr. Schwartz’s opinion unsound under Rule 702

6

simply by advancing a contrary position.

7

Expert testimony need only be “based on sufficient facts or data.” FED. R.

8

EVID. 702(b) (emphasis added); see Bosley v. DePuy Synthes Sales Inc., No. C21-

9

1683-MLP, 2023 WL 6038010, at *4 (W.D. Wash. Sept. 15, 2023) (finding that

10

challenged expert did more than “baldly state” a conclusion) (citing United States. v.

11

W.R. Grace, 455 F. Supp. 2d 1148, 1152 (D. Mont. 2006)). “[A]n expert [may] rely on

12

hypothetical facts that are supported by the evidence.” FED. R. EVID. 702 advisory

13

committee’s notes to 2000 amendments. Rule 702(b) “is not intended to authorize a

14

trial court to exclude an expert’s testimony on the ground that [it] believes one

15

version of the facts and not the other.” Bosley, 2023 WL 6038010, at *4 (internal

16

quotation omitted).

17

Valve, in making its analytical gap argument, takes an overly exacting view

18

of Rule 702’s requirements. Dr. Schwartz provides common evidence of the varied

19

ways in which Valve establishes its PMFN expectation. Whether that evidence and

20

Dr. Schwartz’s conclusions deserve credence is an inquiry for a different day.

21
22

3.1.3

Dr. Schwartz’s but-for market share approach satisfies
Rule 702.

In support of his opinion about class-wide injury, Dr. Schwartz concludes

23
ORDER - 10

1

that, in a but-for world, Valve would hold a

2

game distribution market, rather than the roughly

3

Dkt. Nos. 182-1, 343 ¶¶ 130, 375. In reaching these figures, Dr. Schwartz relies on

4

Valve’s share of Steam Store game sales between 2008–2012. Id. Valve contends

5

this is untethered to an analytical base and thus produces a mere speculative

6

conclusion. Dkt. No. 321 at 10–12. Valve questions why Dr. Schwartz would assume

7

that the market as it existed in the Steam Store between 2008–2012, when Valve

8

sold a significant amount of its own games, is relevant today when Valve mostly

9

sells others’ games. Id.

10

10 percent share of the PC video

percent share it holds now.

The purpose of Dr. Schwartz’s analysis is not to estimate market share at a

11

temporal point but at a circumstantial one: A mature competitive PC game

12

distribution market “where the alleged [antitrust] conduct never took place” and

13

presumably where Valve lacks the market power it holds today, ostensibly forcing

14

third-party game companies to sell through the Steam Store. Dkt. No. 348-1 ¶ 227.

15

In this competitive market, Valve would be relegated to generating revenue mostly

16

from the sale of its own games.

17

According to Dr. Schwartz, because this market does not currently exist, he is

18

left with no choice but to use the closest historic analogue. See Dkt. Nos. 343 ¶ 302,

19

348-1 ¶ 227. Before 2008, not all large developers and publishers distributed their

20

games on Steam. Dkt. No. 343 ¶¶ 302, 375. And after 2012, some distributed their

21
22
23

10 Valve is not a publicly traded company and therefore has no external reporting

obligation for proprietary information. So this redaction, like others, is necessary to
maintain the confidentiality of Valve’s business information. See, e.g., Dkt. Nos. 95,
159, 169, 236, 237, 330, 331 (protective orders and orders granting motions to seal).
ORDER - 11

1

titles through their own platform(s). Id. Therefore, Dr. Schwartz focuses on Steam

2

Store sales between these two points because this is when the largest game

3

developers and publishers—Valve included—sold their games primarily through the

4

Steam Store. See id. ¶¶ 332–410.

5

Ultimately, Dr. Schwartz’s resulting opinion may be incorrect. But that is not

6

the standard for admissibility. And for Rule 23 purposes, he need only provide a

7

reliable method to establish the existence of antitrust impact and injury, and then

8

show that class-wide damages are susceptible to common proof. See Lytle v.

9

Nutramax Lab’ys, Inc., 114 F.4th 1011, 1026–27 (9th Cir. 2024). His “calculations

10

need not be exact.” Comcast Corp. v. Behrend, 569 U.S. 27, 35 (2013). They need

11

only provide “some basis from which to estimate reasonably, and without undue

12

speculation, the damages flowing from the antitrust violations.” Moore v. Jas. H.

13

Matthews & Co., 682 F.2d 830, 836 (9th Cir. 1982). Dr. Schwartz’s market share

14

estimate, as utilized in his but-for analysis, however limited it might be in

15

predicting real world outcomes, meets this standard.

16

3.1.4

Steam Key activity is irrelevant to Dr. Schwartz’s opinions.

17

Dr. Schwartz also concludes that Valve’s commission on Steam Store sales in

18

a competitive market would be

19

averaged.11 Dkt. Nos. 182-1, 343 ¶¶ 352, 377. Dr. Schwartz did not factor in Steam

20
21
22
23

percent, rather than the

it recently

11 Dr. Schwartz determined this by averaging the overall commission rate on Steam

Store sales from the beginning of the proposed class period, January 28, 2017,
through December 31, 2022. See, e.g., Dkt. No. 343 ¶¶ 36, 82, 308, 334, 352, 400.
Presumably, data beyond 2022 was unavailable to Dr. Schwartz. According to
Valve’s expert, though, its average rate is slightly lower, averaging
percent.
Dkt. No. 233-1 ¶ 155.
ORDER - 12

1

Keys in determining these amounts. See Dkt. No. 343 ¶¶ 332–410. Valve contends

2

this is error. Dkt. No. 321 at 12–13.

3

Steam Keys are unique alphanumeric codes used to access Steam-enabled

4

games, which Valve provides to video game companies to distribute as they see fit.

5

Dkt. No. 343 ¶¶ 45–47. Steam Keys allow gamers to access Steam-enabled games

6

without purchasing them through the Steam Store. Id. Sometimes game companies

7

give Steam Keys away; sometimes they charge. Id. ¶ 45. This varies by video game

8

and game company, as does the number of Keys which Valve provides at its

9

discretion. Id. ¶ 46. According to Dr. Schwartz, Valve makes Steam Keys available

10

to protect its market position and bring new customers into Steam. Id. ¶ 47.

11

To assess antitrust impact and injury, Valve suggests that Steam’s

12

commission rate must include the revenue it generates on all Steam-enabled games,

13

including those distributed outside the Steam Store through Steam Keys. See Dkt.

14

Nos. 321 at 12–13, 363 at 3–5. If true, this would mean Valve’s effective commission

15

rate is much lower than Dr. Schwartz’s calculation. Id. But this ignores the fact that

16

Valve utilizes a revenue share model, with Steam Store sales as the driver. It is

17

through this mechanism that Valve allegedly exacts its antitrust toll from the

18

putative class. See Dkt. No. 343 ¶¶ 332–410.

19

In addition, Valve’s argument skates over antitrust netting requirements.

20

Dr. Schwartz surmises that in a competitive market, Valve would continue to

21

provide Steam Keys. Dkt. No. 348-1 ¶¶ 180–85. If anything, to maintain its

22

competitive position, Valve would need to provide more Keys than it does now. See

23

Dkt. No. 383-1 ¶ 16. This means that there is no legal basis to net the savings
ORDER - 13

1

provided to the putative class through Steam Keys (sales commissions not paid)

2

against Valve’s allegedly inflated Steam Store commissions. See Los Angeles Meml.

3

Coliseum Commn. v. Nat’l Football League, 791 F.2d 1356, 1370 (9th Cir. 1986) (in

4

determining antitrust damages, plaintiffs need not net savings they would incur

5

regardless of the antitrust conduct against the harm associated with that conduct);

6

see also In re Glumetza Antitrust Litig., 336 F.R.D. 468, 480 (N.D. Cal. 2020) (the

7

antitrust impact is determined at the time of overcharge—later savings are

8

irrelevant) (citing Hanover Shoe, Inc. v. United Shoe Mach. Corp., 392 U.S. 481, 489

9

(1968)). Again, according to Dr. Schwartz, Valve would provide Keys irrespective of

10

whether it charged inflated commissions. See Dkt. No. 348-1 ¶¶ 180–85.

11

3.1.5

Dr. Schwartz’s passthrough approach satisfies Rule 702.

12

For purposes of determining class-wide antitrust impact and injury,

13

Dr. Schwartz estimated what portion of the savings from a reduced (i.e.,

14

competitive) commission a video game company might pass through to game

15

consumers. See Dkt. No. 343 ¶¶ 379–97. If they were to pass through all of it, there

16

would be no antitrust impact or injury.

17

Because a competitive market does not presently exist, Dr. Schwartz

18

compared the prices certain companies charged on a selection of Steam Store games

19

before and after Valve adopted its present volume discount policy. Id. From this

20

information, Dr. Schwartz could see how much savings those companies passed on

21

to consumers. What he found was, on average, they passed through

22

of their savings. Dkt. Nos. 182-1, 343 ¶ 397. Dr. Schwartz then used this figure to

23

gauge the impact and injury Valve’s alleged antitrust conduct caused publishers.
ORDER - 14

percent

1

Dkt. No. 343 ¶¶ 379–97.

2

Valve challenges the statistical method Dr. Schwartz used to select games for

3

his analysis, focusing on the sample size and his practice of removing outliers. Dkt.

4

No. 321 at 13–15. But as Dr. Schwartz explains, companies are highly likely to

5

retain at least a portion of their savings. See Dkt. No. 348-1 ¶ 253–54. He applied

6

an average passthrough, negating the need for individualized inquiry about

7

antitrust impact and injury. Id.

8

This type of sampling largely fits what other courts have permitted in a

9

similar context. See, e.g., In re Broiler Chicken Grower Antitrust Litig. (No. II), No.

10

6:20-MD-02977-RJS-CMR, 2024 WL 2117359, at *18 (E.D. Okla. May 8, 2024); In re

11

Takata Airbag Prod. Liab. Litig., No. 15-MD-02599-MORENO, 2022 WL 4594123,

12

at *5 (S.D. Fla. July 19, 2022). So this Court considers it to be generally accepted.

13

Statistical averaging is a “generally reliable econometric technique to control for the

14

effects of the differences among class members and isolate the impact of the alleged

15

antitrust violations on the prices paid by class members.” Olean Wholesale Grocery

16

Coop., Inc. v. Bumble Bee Foods LLC, 31 F.4th 651, 677 (9th Cir. 2022).
In summary, Dr. Schwartz’s methods are reliable and within the norms for

17
18

admission, as provided by Rule 702. At bottom, Valve’s critiques go to the weight of

19

the evidence, rather than admissibility. See Apple iPod iTunes Antitrust Litig., No.

20

05-cv-0037-YGR, 2014 WL 4809288, at *4 (N.D. Cal. Sept. 26, 2014).

21
22

3.2

Plaintiffs’ motion for Class Certification.
Plaintiffs move to certify a class of game companies (or persons) who paid a

23
ORDER - 15

1

commission to Valve to sell their games on Steam from January 2017 through

2

present. See generally Dkt. No. 256.

3

To certify a class action under Rule 23, Plaintiffs must show (1) the class is so

4

numerous that joinder is impracticable, (2) there are common questions of law or

5

fact, (3) the claims or defenses of the representatives are typical of those of the

6

class, and (4) the representatives will fairly and adequately protect the interests of

7

absent class members. FED. R. CIV. P. 23(a); see Mazza v. Am. Honda Motor Co., 666

8

F.3d 581, 588 (9th Cir. 2012). If the Rule 23(a) requirements are met, the proposed

9

class must also satisfy at least one of the three requirements found in Rule 23(b).

10

Plaintiffs seeks to certify a (b)(3) class. To do so, they must establish that “questions

11

of law or fact common to the class members predominate over any questions

12

affecting only individual members, and that a class action is superior to other

13

available methods for fairly and efficiently adjudicating the controversy.” FED. R.

14

CIV. P. 23(b)(3).

15

The decision to certify a class is within the Court’s discretion. Vinole v.

16

Countrywide Home Loans, Inc., 571 F.3d 935, 944 (9th Cir. 2009). In determining

17

whether Plaintiffs have carried their burden by a preponderance of the evidence,

18

the Court must conduct a “rigorous analysis.” Gen. Tel. Co. of Sw. v. Falcon, 457

19

U.S. 147, 161 (1982). This inquiry may “entail some overlap with the merits” but

20

only if it overlaps with the Rule 23 requirements. Ellis v. Costco Wholesale Corp.,

21

657 F.3d 970, 981 (9th Cir. 2011).

22
23

In opposing certification, Valve objects to Plaintiffs’ market definition; it also
attacks whether Plaintiffs have shown sufficient Rule 23(b)(3) commonality and
ORDER - 16

1

predominance. See Dkt. No. 319 at 6–36. None of Valve’s critiques, however, are

2

enough to defeat certification.

3

3.2.1

Plaintiffs present a cogent market.

4

Plaintiffs allege that Valve, through Steam, holds and maintains monopoly

5

power in the PC video game digital distribution market. See, e.g., Dkt. No. 127 ¶ 52.

6

Valve suggests Plaintiffs’ market definition is “incoherent” because it excludes other

7

game forms (such as console and mobile games) and other distribution methods

8

(such as brick-and-mortar stores, cloud gaming, and multi-game subscription

9

models). Dkt. No. 319 at 6–8. But Plaintiffs present significant evidence supporting

10

their market definition. See, e.g., Dkt. Nos. 257 ¶¶ 21–121, 343 ¶¶ 16–28 (Professor

11

Joost Rietveld, Ph.D. and Dr. Schwartz’s expert reports). To be sure, Valve presents

12

countervailing evidence, Dkt. No. 344 ¶¶ 22–57, 84–133, but it is simply not

13

convincing.

14

As Plaintiffs’ experts make clear, downloaded PC games are materially

15

distinct from other forms. This is based on the purchase and game play experience

16

(such as the hardware inputs and outputs, the computing capacity of the PC itself,

17

the opportunities for game customization, the lack of latency that downloaded

18

games offer, and the game play experience). See, e.g., Dkt. No. 257 ¶¶ 49–61, 66–82,

19

104, 108, 109, 116–21. This matches both the regulatory view and even Valve’s, as it

20

has outlined in other venues. Id. ¶¶ 63, 65.

21

3.2.2

22
23

Numerosity, typicality, adequacy, and superiority are
satisfied.

Valve does not contest that Plaintiffs satisfy numerosity, typicality, adequacy

ORDER - 17

1

of representation, or superiority. See generally Dkt. No. 319 at 21–36. Nor would it

2

appear that Valve reasonably could.

3

As to numerosity, “[w]here the exact size of the class is unknown, but general

4

knowledge and common sense indicate that it is large, the numerosity requirement

5

is satisfied.” In re Abbott Laboratories Norvir Anti-Tr. Litig., No. C 04-1511 CW,

6

2007 WL 1689899, at *6 (N.D. Cal. June 11, 2007) (internal quotation omitted). “[I]n

7

general, a class include[ing] at least 40 members” is sufficient. Rannis v. Recchia,

8

380 F. App’x 646, 651 (9th Cir. 2010); see also Troy v. Kehe Food Distrib., Inc., 276

9

F.R.D. 642, 652 (W.D. Wash. 2011) (certifying a class of 43 to 54 workers). Plaintiffs

10

submit unrebutted evidence that the proposed class consists of thousands of game

11

developers or publishers. Dkt. Nos. 182-1, 343 ¶ 399. This is enough to satisfy

12

numerosity.

13

As to typicality, at issue is “whether other members have the same or similar

14

injury, whether the action is based on conduct not unique to the named plaintiffs,

15

and whether other class members have been injured by the same course of conduct.”

16

Ellis, 657 F.3d at 984 (cleaned up). In the antitrust context, typicality is established

17

when “all class members allege the same antitrust violation.” In re Qualcomm

18

Antitrust Litig., 328 F.R.D. 280, 295 (N.D. Cal. 2018) (citing In re High-Tech

19

Employee Antitrust Litig., 985 F. Supp. 2d 1167, 1181 (N.D. Cal. 2013)); see Joseph

20

McLaughlin, Typicality—Application of typicality requirement to particular cases—

21

Antitrust, McLaughlin on Class Actions: Law and Practice § 4:21 (20th ed. 2023).

22

Here, the alleged antitrust conduct—imposition of an alleged PMFN—is the same

23

for each member of the putative class. See Dkt. Nos. 127 ¶¶ 156–388, 343 ¶¶ 150–
ORDER - 18

1
2

240. Typicality is satisfied.
As to adequacy of representation, the Court examines whether the named

3

plaintiffs and their counsel have any conflicts of interest with other class members

4

and whether they will prosecute the action vigorously on behalf of the class. Ellis,

5

657 F.3d at 985. This requires “an absence of antagonism between representatives

6

and absentees and a sharing of interest between representatives and absentees.” Id.

7

(cleaned up). The Court knows of no such antagonism, nor has Valve suggested any.

8

See generally Dkt. No. 319 at 21–36. And as Plaintiffs explain, Wolfire and Dark

9

Catt both share a common injury as putative class members (allegedly inflated

10

commissions). See Dkt. No. 256 at 23–24. Wolfire and Dark Catt will adequately

11

represent their game company peers.

12

As for superiority, the Court must find that a “class action is superior to other

13

available methods for fairly and efficiently adjudicating the controversy.” FED. R.

14

CIV. P. 23(b)(3). When undertaking this inquiry, the Court considers (1) the interest

15

of individuals within the class in controlling their own litigation, (2) the extent and

16

nature of any pending litigation commenced by or against the class involving the

17

same issues, (3) the convenience and desirability of concentrating the litigation in a

18

particular forum, and (4) the manageability of the class action. See FED. R. CIV. P.

19

23(b)(3)(A)–(D); Zinzer v. Accufix Rsch. Inst., Inc., 253 F.3d 1180, 1190 (9th Cir.

20

2001). This analysis must “focus on the efficiency and economy elements of the class

21

action so that cases allowed under subdivision (b)(3) are those that can be

22

adjudicated most profitably on a representative basis.” Id.

23

Given the complexity of the legal and evidentiary issues raised in this
ORDER - 19

1

consolidated suit, resolving them in a uniform manner through a class action is

2

superior to resolving them through individual suit(s). Moreover, this is the most

3

appropriate forum to do so. Valve admits that its headquarters is within this

4

district. See Dkt. No. 128 ¶ 34. And while individual gamers have filed additional

5

suits against Valve in this district,12 this does not defeat the notion that a class

6

action remains a superior vehicle for resolving the claims presented here, and if

7

anything, supports this notion. The Court reserves judgment, for now, on whether

8

further consolidation is warranted.

9
10
11
12
13
14
15
16
17
18
19

3.2.3

Valve’s attacks on commonality and predominance are
unpersuasive.

Commonality requires that the “class members’ claims depend upon a
common contention such that determination of its truth or falsity will resolve an
issue that is central to the validity of each claim in one stroke.” Mazza, 666 F.3d at
588 (internal citation omitted). The key inquiry is not whether a plaintiff has raised
common questions but whether “class treatment will generate common answers apt
to drive the resolution of the litigation.” Abdullah v. United States Sec. Assocs., Inc.,
731 F.3d 952, 957 (9th Cir. 2013) (emphasis in original) (internal citation omitted).
Every question of law or fact need not be common to the class. Rather, all Rule
23(a)(2) requires is “a single significant question of law or fact.” Id. (internal
citation omitted). The existence of “shared legal issues with divergent factual

20
21
22
23

12 See Elliott v. Valve Corp., Case No. 2:24-cv-1218-JNW, Dkt. No. 1 (W.D. Wash

August 9, 2024) (putative antitrust class action brought by individuals who
successfully challenged the enforceability of Valve’s gamer arbitration provision);
Hepler et al. v. Valve Corp., No. 2:24-cv-01735-JNW (W.D. Wash.), and Drake et al.
v. Valve Corp., No. 2:24-cv-01743-MLP (W.D. Wash.).
ORDER - 20

1

predicates is sufficient, as is a common core of salient facts coupled with disparate

2

legal remedies within the class.” Hanlon v. Chrysler Corp., 150 F.3d 1011, 1019 (9th

3

Cir. 1998) (amended).

4

Predominance “tests whether a proposed class is sufficiently cohesive to

5

warrant adjudication by representation.” Amchem Prods., Inc. v. Windsor, 521 U.S.

6

591, 615–16 (1997) (cleaned up). This inquiry presumes the existence of common

7

factual or legal issues required under Rule 23(a)’s “commonality” element and

8

instead “focuses on the relationship between the common and individual issues.”

9

Hanlon, 150 F.3d at 1023. “When common questions present a significant aspect of

10

the case and they can be resolved for all members of the class in a single

11

adjudication, there is a clear justification for handling the dispute on a

12

representative rather than on an individual basis.” Id.

13

In attacking commonality and predominance, Valve takes a buckshot

14

approach. See generally Dkt. No. 319. Its specific arguments include the following:

15

(a) Plaintiffs cannot establish the existence of a PMFN Policy through common

16

evidence; (b) alleged antitrust impact, injury, and damages are too particularized to

17

allow for class treatment; and (c) the statute of limitations presents sufficient

18

individualized issues to preclude a predominance finding. Id. at 8–36. As described

19

below, none of Valve’s arguments are compelling.

20

a. Plaintiffs present common evidence of a PMFN Policy.

21

Valve admits that its standard distribution agreement requires content

22

parity; but it contends this only applies to DLC, not applications (i.e., games). Dkt.

23

No. 242 at 14–15. According to the agreement, game companies that sell games
ORDER - 21

1

within and outside Steam, and who also distribute DLC for that game, must

2

maintain “material parity” between their Steam and non-Steam customers. Dkt.

3

No. 182-1813 at 6. Frankly, this provision is not a model of clarity, as its meaning is

4

open to interpretation. But Plaintiffs present Rule 30(b)(6) testimony suggesting

5

that Valve understands it to require content parity in all instances, not just for

6

DLC. See Dkt. No. 182-3314 at 15. This is common proof of a content parity

7

requirement.

8

As to price parity, as mentioned above, see supra Part 3.1.2, Plaintiffs present

9

evidence that Valve communicates this portion of the PMFN Policy in a roundabout

10

manner and through various means. See Dkt. No. 343 ¶¶ 157–67. They include

11

certain contract terms associated with providing Steam Keys, one-off discussions

12

with game companies, and various messages in industry forums. Id. All this

13

contributes to a common industry understanding of Valve’s price-parity expectation.

14

See Dkt. Nos. 343 ¶ 320, 348-1 ¶¶ 18–21. This is sufficient, under the rule of

15

reason,15 to establish common evidence of a price parity expectation that drives

16
17
18

13 Docket 182-18 is a sealed unredacted version of Docket Number 204-18. See Dkt.

Nos. 198, 236, 237 (motion to seal and resulting order).

14 Docket Number 182-33 is a sealed unredacted version of Docket Number 204-33.

19

See Dkt. Nos. 198, 236, 237 (motion to seal and resulting order).

20

15 The rule of reason is a burden-shifting framework used to gauge whether non-per

21
22
23

se anticompetitive conduct violates the Sherman Act. See Amex, 585 U.S. at 541–42.
While the rule of reason is mainly a merits determination, it applies with equal
force in assessing whether an antitrust case satisfies Rule 23. See, e.g., In re Beer
Distrib. Antitrust Litig., 188 F.R.D. 557, 563 (N.D. Cal. 1999); Kamakahi v. Am.
Socy. for Reprod. Med., 305 F.R.D. 164, 174 (N.D. Cal. 2015). It thus implicates
issues of commonality and predominance.
ORDER - 22

1

common market behavior. See Epic Games, Inc. v. Apple, Inc., 67 F.4th 946, 974

2

(9th Cir. 2023); see In re High-Tech Employee Antitrust Litig., 985 F. Supp. 2d at

3

1172 (discussing the notion of industrywide or “common impact[s]”).

4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22

b. Common issues drive the inquiry about antitrust impact, injury, and
damages.
As to the antitrust impact of Valve’s supposed PMFN Policy, Valve provides a
litany of challenges to Plaintiffs’ theories. They include examples of games selling
for less off Steam and testimony from game companies explaining why they prefer
keeping their game prices uniform across platforms (i.e., to meet consumer
expectations). See Dkt. No. 319 at 27–28 (citing Dkt. Nos. 322-10 at 18, 322-24 at 9,
344 ¶¶ 287–97). None of this defeats predominance. Rather, this evidence raises
common questions about the class-wide response to Valve’s alleged PMFN
expectation, an issue which predominates over any individualized questions. See
Amgen Inc. v. Connecticut Ret. Plans and Tr. Funds, 568 U.S. 455, 466 (2013) (“the
focus of Rule 23(b)(3) is on the predominance of common questions”) (emphasis in
original).
Concerning antitrust injury, Valve contends that Plaintiffs cannot “meet
their predominance burden” to “prove that antitrust impact is ‘capable of being
established through a common body of evidence, applicable to the whole class.’” Dkt.
No. 319 at 29 (citing Olean Wholesale Grocery Coop., Inc., 31 F.4th at 666). In
support, Valve makes the same arguments it did in moving to exclude
Dr. Schwartz’s testimony: (a) the import of Steam Keys cannot be ignored, (b)
Plaintiffs have no means of accurately determining pass through, and (c) any

23
ORDER - 23

1

antitrust overcharge must be offset by the benefits received. See id. at 29–34. Those

2

arguments are no more persuasive for purposes of Rule 23(b)(3) than they were for

3

purposes of Rule 702. See supra Part 3.1.4–3.1.5. Through Dr. Schwartz’s

4

testimony, Plaintiffs establish a common antitrust injury: inflated Steam Store

5

commissions that all class members must pay. See Dkt. Nos. 343 ¶¶ 241–397; 348-1

6

¶¶ 160–95; 383-1 ¶¶ 9–14, 29–34.

7

Finally, as to antitrust damages, Valve suggests Plaintiffs cannot measure

8

them across the entire class. See Dkt. No. 319 at 34–36. In support, it repeats the

9

methodological challenges raised against Dr. Schwartz’s testimony. Compare id.;

10

with Dkt. No. 321 at 3–6, 12–15. As before, those arguments fail here. At this stage,

11

Plaintiffs may utilize statistical means to estimate antitrust injury across the class.

12

The apportionment of damages to individual class members is not required. See In

13

re Glumetza Antitrust Litig., 336 F.R.D. at 479 (“‘[a]ntitrust plaintiffs may satisfy

14

the predominance requirement by using a model that estimates the damages

15

attributable to the antitrust injury, even if more individualized determinations are

16

needed later to allocate damages among class members.’”) (quoting In re Suboxone

17

(Buprenorphine Hydrochlorine and Naloxone) Antitrust Litig., 967 F.3d 264, 269 (3d

18

Cir. 2020)). Said another way, “[t]he need for individual damages calculations does

19

not, alone, defeat class certification.” Vaquero v. Ashley Furniture Indus., Inc., 824

20

F.3d 1150, 1155 (9th Cir. 2016).

21
22

c. The statute of limitations for antitrust claims does not defeat
predominance.
Finally, Valve suggests that, because the antitrust statute of limitations is

23
ORDER - 24

1

four years, and the proposed class period extends beyond that to January 28, 2017,

2

individualized inquiry is required to establish when each putative class member

3

was first harmed by Valve’s alleged conduct. See Dkt. No. 319 at 24–27 (citing 15

4

U.S.C. § 15b). Valve contends that individualized inquiries would be required to

5

determine when each class member first heard of the PMFN. Id. at 25. This is

6

because the continuing violation doctrine would not apply if premised on the

7

“‘passive receipt’” of revenue derived from Valve’s alleged anticompetitive conduct,

8

rather than an “‘overt act of enforcement.’” Id. at 26 (quoting Eichman v. Fotomat

9

Corp., 880 F.2d 149, 160 (9th Cir. 1989)). But Plaintiffs point to common proof of

10

Valve’s overt acts occurring within the proposed class period. See Dkt. No. 347 at

11

20–21 (citing Dkt. No. 343 ¶¶186 n.503, 187 n.505, 191). The anticompetitive effect

12

of these acts present common issues that predominate. Moreover, even if the statute

13

did present individualized issues, this alone would not be enough to defeat

14

predominance. See Williams v. Sinclair, 529 F.2d 1383, 1388 (9th Cir. 1975); Nitsch

15

v. Dreamworks Animation SKG Inc., 315 F.R.D. 270, 308 (N.D. Cal. 2016).
In summary, none of Valve’s commonality and predominance attacks defeat

16
17
18
19

class certification.
3.3

Plaintiffs’ Request for the Appointment of Class Counsel.
The Court previously appointed numerous attorneys as Interim Co-Lead

20

Class Counsel in this matter. See Dkt. No. 92. Plaintiffs now ask that those same

21

attorneys be appointed Co-Lead Class Counsel for the rest of the case (through

22

trial). See Dkt. No. 256 at 35. In considering this request, the Court must consider:

23
ORDER - 25

1

(1) “the work counsel has done in identifying or investigating potential claims in the

2

action”; (2) “counsel’s experience in handling class actions, other complex litigation,

3

and the types of claims asserted in the action”; (3) “counsel’s knowledge of the

4

applicable law”; and (4) “the resources that counsel will commit to representing the

5

class[.]” FED. R. CIV. P. 23(g)(1)(A). To date, interim class counsel have pursued the

6

putative class’s claims in an effective, collaborative, and well-resourced manner.

7

This matches their collective performance in other large class actions. See Dkt. Nos.

8

204-80, 204-81, 204-82, 204-83. The Court sees no reason that, if reappointed,

9

interim class counsel’s performance would not continue just like it has to date.
4. CONCLUSION

10
11

In sum, the Court GRANTS Plaintiffs’ motion for class certification and the

12

appointment of co-lead class counsel, Dkt. No. 256, DENIES Valve’s motion to

13

exclude, Dkt. No. 321, and ORDERS as follows:

14

1. The following class is certified:
All persons or entities who, directly or through an agent, paid a
commission to Valve in connection with the sale or use of a game
on the Steam platform on or after January 28, 2017, and
continuing through the present until the effects of its scheme are
eliminated (the “Class Period”), and where either (1) the person
or entity was based in the United States and its territories or (2)
the game was purchased or acquired by a United States-based
consumer during the Class Period. Excluded from the Class are
(a) Defendant, its parents, subsidiaries, affiliate entities, and
employees, and (b) the Court and its personnel.

15
16
17
18
19
20
21

2. Wolfire and Dark Catt are appointed class representatives; and

22

3. The firms below are appointed Co-Lead Class Counsel:

23
ORDER - 26

Quinn Emanuel Urquhart & Sullivan, LLP; Constantine Cannon
LLP; Lockridge Grindal Nauen P.L.L.P.; and Wilson Sonsini
Goodrich & Rosati, P.C.

1
2
3

The case management schedule, which the Court previously adopted, Dkt.

4

No. 98, remains in force. The Court will defer setting a schedule beyond summary

5

judgment until it disposes of such motions, currently due 180 days from today. See

6

id. at 2.

7

This ruling has no bearing on a putative consumer class, as described in this

8

and related cases. See Dkt. No. 370; Elliott, Case No. 2:24-cv-1218-JNW, Dkt. No. 1;

9

Drake, No. 2:24-cv-01743-JNW, Dkt. No. 1; Hepler, No. 2:24-cv-01735-JNW, Dkt.

10

No. 1.

11
12

Dated this 25th day of November, 2024.

13
Jamal N. Whitehead
United States District Judge

14
15
16
17
18
19
20
21
22
23
ORDER - 27