OPINION AND ORDER

ANDREW J. PECK, United States Magistrate Judge:

Plaintiffs' "Motion for Recusal or Disqualification" (Dkt. No. 169) is based not on any claim that the Court has an actual bias, but rather on "an appearance of partiality." (Dkt. No. 192: Pls. Reply Br. at 1 n. 1: "Plaintiffs have never accused Judge Peck of actual bias or sought to impugn Judge Peck's integrity. Plaintiffs' only ground for recusal is that the facts taken together create an appearance of partiality.") Plaintiffs' recusal motion is DENIED.

FACTUAL BACKGROUND

The main ground of plaintiffs' motion is that my support for predictive coding showed bias favoring MSL and coerced plaintiffs into assenting to the concept of predictive coding. The chronology of events in this case puts the lie to plaintiffs' claim.

This case was referred to me on November 28, 2011. (Dkt. No. 48.) Well before that date, MSL had informed plaintiffs that it "proposes using keywords as well as the analytical tools available in the Axcelerate review platform, which includes predictive coding...." (Dkt. No. 178: Anders Aff. Ex. 2: Anders 10/21/11 Letter to Wipper at 3; see also Anders Aff. Ex. 3: Anders 11/3/11 Letter to Wipper at 4-9.) Plaintiffs requested that MSL fully disclose its proposed predictive coding methodology for plaintiffs' consideration. (See Dkt. No. 124: Nurhussein 3/19/12 Aff. Ex. E: Wipper 11/9/11 Letter to Anders at 5; see also id. Ex. D: Wipper 10/25/11 Letter to Anders at 2.)

December 2, 2011 Conference

I held my first conference with the parties on December 2, 2011. (Dkt. No. 51: 12/2/11 Conf. Tr.) During that conference, MSL's counsel stated that an open issue was "plaintiff's reluctance to utilize predictive coding to try to cull down" approximately three million electronic documents from the agreed-upon custodians. (12/2/11 Conf. Tr. at 7-8.) Because of my Search, Forward article that I would call to the parties' attention at the conference, I stated that: "You must have thought you died and went to Heaven when this was referred to me," to which MSL's counsel responded: "I'm just thankful that, you know, we have a person familiar with the predictive coding concept." (12/2/11 Conf. Tr. at 8-9.) Plaintiffs did not move to recuse me at that time for making this comment. To the contrary, plaintiffs' counsel clarified that MSL had "over simplified [plaintiffs'] stance on predictive coding," i.e., that they were not opposed but had "multiple concerns ... on the way in which [MSL] plan to employ predictive coding" and plaintiffs wanted "clarification." (12/2/11 Conf. Tr. at 21.)

The Court did not rule on any predictive coding issue but offered the parties the following advice:

(12/2/11 Conf. Tr. at 20-21, emphasis added.)1 The conference adjourned with the parties agreeing to further discuss the ESI protocol. (12/2/11 Conf. Tr. at 34-35.)2

My October 2011 Search, Forward Article

As noted, at the December 2, 2011 conference I alerted the parties to my article, Search, Forward: Will manual document review and keyword searches be replaced by computer assisted coding?, appearing in the October 2011 issue of Law Technology News. See Andrew Peck, Search, Forward, L. Tech. News, Oct. 2011, at 25-26, 29. The article reviewed the problems with manual review and keyword searches (if poorly done), and generally described how computer-assisted review, a/k/a predictive coding, worked. See Da Silva Moore v. Publicis Groupe, 11 Civ. 1279, ___ F.R.D. ___, ___-___, 2012 WL 607412 at *2-3 (S.D.N.Y. Feb. 24, 2012) (Peck, M.J.) (quoting Search, Forward, L. Tech. News, Oct. 2011, at 29), adopted, 2012 WL 1446534 (S.D.N.Y. Apr. 26, 2012). I ended the article with the comment that lawyers could view the Search, Forward article "as a sign of judicial approval" of predictive coding, but only for appropriate cases, stating:

Andrew Peck, Search, Forward, L. Tech. News, Oct. 2011, at 29. Having had this article and my general opinion about predictive coding brought to the parties' attention at the December 2 conference, plaintiffs did not move for my recusal.

January 4, 2012 Conference

Plaintiffs' counsel wrote asking the Court to postpone the next scheduled conference, stating that "[a]lthough plaintiffs are prepared to consider the use of predictive coding as a search method in general, Plaintiffs need more time to evaluate and provide feedback on [MSL's] draft proposal and its methodology." (Dkt. No. 178: Anders Aff. Ex. 5: Wipper 12/19/11 Letter to Court, at 2.) Plaintiffs said that after their consultant (DOAR Litigation Consulting) examined MSL's proposal, the "parties then may engage in dialogue in order to address the issues, if any, that require resolution by the Court." (Id.) The Court granted the extension (over MSL's objection), and reminded the parties to read my article: "As to predicative coding, you should read my article, `Search, Forward' in the Oct. 2011 issue of Law Technology News." (Dkt. No. 58: 12/20/11 Memo Endorsed Order.)

On January 3, 2012, in anticipation of the January 4, 2012 conference, plaintiffs submitted a letter to the Court containing their discovery proposals, including an ESI protocol utilizing predictive coding. (Anders Aff. Ex. 6: Wipper 1/3/12 Letter to Court.) Plaintiffs informed the Court that they "have attempted to work within Defendants' proposed methodology while honoring their restrictions" but that "Plaintiffs believe that Defendants' proposed use of predictive coding while arbitrarily imposing a cap [on cost] runs counter to the reasoning behind the use of predictive coding." (Id. at 2.)3 Plaintiffs' proposed protocol stated that "the following is a summary of the Parties' agreement on the use of Predictive Coding" for search of MSL's email archive. (Anders Aff. Ex. 6: 1/3/12 Wipper Letter to Court, attached "Plaintiffs Proposed Protocol Relating to the Production of Electronically Stored Information ('ESI')" at 12.) Plaintiffs' Proposed Protocol, in a section entitled "General Overview of Predictive Coding Process," described that MSL would use Recommind's Axcelerator software and develop a "seed set" to train the computer, it would be a transparent process, and it would include quality testing. (Pls. Proposed Protocol at 12-20.) In other words, after discussions with its consultant DOAR, plaintiffs were proposing using predictive coding, but with certain differences in the details compared to MSL's proposal.

At the January 4, 2012 conference, after addressing other discovery issues, the parties turned to the email search issue.4 The following colloquy occurred:

(1/4/12 Conf. Tr. at 51, emphasis added.) The parties spent time conferring with each other while the Court handled another case, and later in the afternoon reported that they had made "a lot of progress." (1/4/12 Conf. Tr. at 53-55.) In discussing what needed to be done and when the next conference would be, MSL counsel Anders noted he would be on vacation, and the Court suggested that another lawyer such as Ms. Chavey cover in his absence so as not to lose time. (1/4/12 Conf. Tr. at 60-61.) The Court also noted that while he might not be working on the case, Anders could involve his firm's Florida ediscovery counsel, Ralph Losey, who I stated I knew "very well." (1/4/12 Conf. Tr. at 61.)5 Counsel then suggested a date for the next conference, and the Court noted that "That's LegalTech week," but that February 2 would work. (1/4/12 Conf. Tr. at 66.) Counsel did not ask what I would be doing at LegalTech.6

Legal Tech

I was a speaker at seven panels in connection with LegalTech on January 30, 31 and February 1, 2012.7 LegalTech is run by ALM (the company that publishes The American Lawyer, Law Technology News and other publications). LegalTech New York 2012 had thirty-nine sponsors, including Recommind, and 186 exhibitors, not to mention hundreds of paying attendees.8 I did not speak at any panels sponsored by Recommind, nor did I receive any expense reimbursement or compensation from Recommind during LegalTech 2012 nor at any other conference. The panels in which I participated at LegalTech 2012 involved general ediscovery issues (such as cooperation, preservation, proportionality, etc.), effective keyword searching, and predictive coding.9 My participation in these educational panels (for which attendees received CLE credit), and that of the other panelists, was arranged before this case was referred to me.

Another panelist on two of the panels was Ralph Losey, national ediscovery counsel in the Florida office of Jackson Lewis, MSL's counsel. (Dkt. No. 178: Anders Aff. Ex. 10: Losey Aff. ¶ 8.) Losey's affidavit makes clear that we have never had any ex parte communication about this lawsuit. Losey stated:

(Anders Aff. Ex. 10: Losey Aff. ¶ 6.) I confirm that. (See also Dkt. No. 158: 4/2/12 Order.)10 Those panels (and my preparation for them) involved only the subject of computer-assisted review in general terms and in comparison to other search techniques (like manual review and keywords). In other words, the level of the panel discussion was similar to that of my Search, Forward article. There was absolutely no discussion of the details of the predictive coding protocol in this case, or in what a predictive coding protocol should look like in general. While plaintiffs' recusal motion makes it sound like I spoke at length about this case on these panels, that simply is not true. Typical of my minor reference during panels to this case (without mentioning the case name, at a time when the case was not yet all over the ediscovery blogs) is this passing reference:

(Dkt. No. 171: Wittels Aff. Ex. B: Excerpt of a Video Recording of the Judicial Perspectives on Technology-Assisted Review Panel at LegalTech N.Y.2012.)11

February 8, 2012 Conference and DOAR's Press Release

The Court's next conference in this case was on February 8, 2012. (Dkt. No. 97: 2/8/12 Conf. Tr.) In advance of the conference, plaintiffs and MSL wrote to the Court enclosing their proposed ESI protocols and highlighting their areas of difference. (See Dkt. No. 124: Nurhussein 3/19/12 Aff. Ex. H: Wipper 1/25/12 Letter to Court; Dkt. No. 171: Wittels Aff. Ex. Z: Anders 1/25/12 Letter to Court.) Plaintiffs' proposed protocol, like MSL's, used predictive coding. (Nurhussein 3/19/12 Aff. Ex. H: Wipper 1/25/12 Letter to Court at 3-5 & Proposed Protocol at 20-35.) Plaintiffs, however, proposed training the computer with two rounds of iterative review of 16,555 documents reviewed each round, while MSL's proposal was for up to seven rounds of review of at least 500 documents per round. (See Nurhussein 3/19/12 Aff. Ex. H: Wipper 1/25/12 Letter to Court, Proposed Protocol at 31-34; Wittels Aff. Ex. Z: Anders 1/25/12 Letter to Court at 6-7; 2/8/12 Conf. Tr. at 70-89.) The Court heard argument on these issues from the parties' counsel and their ediscovery vendors at the February 8, 2012 conference. (See 2/8/12 Conf. Tr.)12

The Court's rulings at the February 8 conference are summarized in my February 24, 2012 opinion, and will not be repeated here. Da Silva Moore v. Publicis Groupe, 11 Civ. 1279, ___ F.R.D. ___, ___-___, 2012 WL 607412 at *3-6 (S.D.N.Y. Feb. 24, 2012) (Peck, M.J.), adopted, 2012 WL 1446534 (S.D.N.Y. Apr. 26, 2012). Suffice it to say that the Court was comparing plaintiffs' predictive coding protocol and defendant MSL's predictive coding protocol. (2/8/12 Conf. Tr. at 57-89.) See Da Silva Moore v. Publicis Groupe, ___ F.R.D. at ___-___, 2012 WL 607412 at *5-6. The Court accepted MSL's predictive coding protocol, with seven iterative review rounds, but with the caveat that if the system was not stabilized at that point, the parties would present quality control verification information to the Court and further rounds would be ordered to stabilize the system. (2/8/12 Conf. Tr. at 76-77, 83-84, 88.) See Da Silva Moore v. Publicis Groupe, ___ F.R.D. at ___, 2012 WL 607412 at *6.

Moreover, plaintiffs' consultant, DOAR's CEO Paul Neale, agreed that, in general, computer-assisted review works and works better than most alternatives. (2/8/12 Conf. Tr. at 76.) Neale noted that "it is fair to say we are big proponents of it." (2/8/12 Conf. Tr. at 76.) Neale added, however, that "[t]his is new technology and it has to be proven out" and that plaintiffs "don't at this point agree that this is going to work." (2/8/12 Conf. Tr. at 75.)

A few days after the February 8 conference, plaintiffs `consultant DOAR issued a press release about my February 8 oral ruling and touted DOAR's role in this "`important step in the wider adoption of predictive coding technologies.'" (Dkt. No. 178: Anders Aff. Ex. 8: 2/13/12 DOAR Press Release.) It is worth quoting DOAR's press release in full:

(Anders Aff. Ex. 8: DOAR 2/13/12 Press Release.)13

The Court notes that even at the May 14, 2012 conference, DOAR's Neale repeated under oath his company's approval of predictive coding generally, stating:

(5/14/12 Conf. Tr. at 51-52, emphasis added.)14

After the February 8 conference (and DOAR's February 13 press release), the parties submitted their "final" ESI Protocol which the Court "so ordered." (Dkt. No. 92: 2/17/12 ESI Protocol & Order.) Plaintiffs included a paragraph in the Protocol objecting to it "in its entirety," noting that "Plaintiffs [had] submitted their own proposed ESI protocol [including predictive coding] to the Court, but it was largely rejected." (2/17/12 ESI Protocol & Order ¶ J. 1 at 22.)

Plaintiffs' Objections to the Court's February 8, 2012 Rulings

Plaintiffs filed their objections to my February 8 rulings on February 22, 2012. (Dkt. No. 93: Pls. Rule 72(a) Objections; see Dkt. No. 94: Nurhussein Aff.; Dkt. No. 95: Neale Aff.) Plaintiffs' consultant, DOAR CEO Neale, reiterated that: "As stated during the February 8, 2012 hearing and cited in Judge Peck's opinion, I am a proponent of the use of predictive coding, when it can be validated as reliable." (Dkt. No. 125: Neale 3/19/12 Aff. ¶ 9.) Plaintiffs also conceded that they were willing to consider the use of predictive coding, "if it was established as reliable." (Dkt. No. 123: Pls. Rule 72 Objections Reply Br. at 5, 10.) Plaintiffs distinguished between generalities and details: "As Plaintiffs repeatedly warned, although the use of predictive coding may be appropriate under certain circumstances, the devil is in the details." (Id. at 1, emphasis added.) Plaintiffs' objections referred to my Search, Forward article and my panels at LegalTech, noting that I "primarily discussed, as a general matter, the propriety of computer-assisted document review (no doubt meaning some variations of predictive coding), but the real question was whether MSL's specific protocol would adequately address its Rule 26 obligations." (Id. at 1-3, 7-8.)

The Court's February 24, 2012 Decision is Affirmed by Judge Carter

The Court issued its formal opinion on the ESI protocol on February 24, 2012. Da Silva Moore v. Publicis Groupe, 11 Civ. 1279, ___ F.R.D. ___, 2012 WL 607412 (S.D.N.Y. Feb. 24, 2012) (Peck, M.J.). Judge Carter overruled plaintiffs' objections and adopted my decision on April 26, 2012. Da Silva Moore v. Publicis Groupe, 11 Civ. 1279, 2012 WL 1446534 (S.D.N.Y. Apr. 26, 2012).

On March 9, 2012, the Court held a further discovery conference. (Dkt. No. 209: 3/9/12 Conf. Tr.) Plaintiffs did not raise recusal at all at the conference. (See generally 3/9/12 Conf. Tr.)

Plaintiffs' Recusal Request and Motion

It was not until March 28, 2012 that plaintiffs wrote a letter asking me to recuse myself from the case. Because MSL wrote that it would want a chance to respond, I entered an order asking plaintiffs whether they wanted me to rule based on letter-briefs or whether they wanted to file a formal motion, and added the following comment:

(Dkt. No. 158: 4/2/12 Order.)

On April 13, 2012, plaintiffs filed their recusal motion. (Dkt. No. 169: Motion; see also Dkt. No. 170: Pls. Recusal Br.; Dkt. No. 171: Wittels Aff.)

Subsequent Events in the Case

On April 20, 2012, plaintiffs asked Judge Carter to hold off on ruling on their objections to my February 8 and February 24 rulings about predictive coding until the recusal motion was decided; on April 25, 2012, Judge Carter rejected their request. (Dkt. No. 174: 4/25/12 Memo Endorsed Order.)

Also on April 25, I held a further discovery conference. (Dkt. No. 180:4/25/12 Conf. Tr.) At the start of the conference, plaintiffs asked me to stay discovery until Judge Carter ruled on their pending motions to further amend their complaint and for FLSA collective action certification. (4/25/12 Conf. Tr. at 2-3.) Only after I denied that application did plaintiffs ask that I take no further action in the case until after I (and presumably Judge Carter) ruled on plaintiffs' recusal motion. (4/25/12 Conf. Tr. at 13.) I denied that request. (4/25/12 Conf. Tr. at 14.) The order in which plaintiffs raised those two applications is indicative of their apparent strategy of seeing whether they prevail on matters and when they do not, only then raising recusal.

On May 14, 2012, I sua sponte reconsidered plaintiffs' request and stayed MSL's review and production of ESI pending Judge Carter's ruling on plaintiffs' motion to amend the complaint and for collective action certification; MSL originally objected but then did not oppose the stay, which I entered in order to avoid the expense of redoing discovery if collective action certification was granted. (Dkt. No. 202: 5/14/12 Conf. Tr. at 71-84.) See Da Silva Moore v. Publicis Groupe, 11 Civ. 1279, 2012 WL 1698980 at *1 (S.D.N.Y. May 14, 2012) (Peck, M.J.).

ANALYSIS

I. RECUSAL LEGAL STANDARD

Under 28 U.S.C. § 455(a), "[a]ny justice, judge, or magistrate judge of the United States shall disqualify himself in any proceeding in which his impartiality might reasonably be questioned." 28 U.S.C. § 455(a). "The purpose of § 455(a) `is to promote confidence in the judiciary by avoiding even the appearance of impropriety whenever possible.'" Green v. N.Y.C. Health & Hosps. Corp., 343 Fed. Appx. 712, 713 (2d Cir.2009) (quoting Liljeberg v. Health Servs. Acquisition Corp., 486 U.S. 847, 865, 108 S.Ct. 2194, 2205, 100 L.Ed.2d 855 (1988)).15

"In determining whether Section 455(a) requires recusal, the appropriate standard is objective reasonableness — whether `an objective, disinterested observer fully informed of the underlying facts, [would] entertain significant doubt that justice would be done absent recusal.'" United States v. Carlton, 534 F.3d 97, 100 (2d Cir.), cert. denied, 555 U.S. 1038, 129 S.Ct. 613, 172 L.Ed.2d 468 (2008); accord, e.g., Green v. N.Y.C. Health & Hosps. Corp., 343 Fed.Appx. at 713-14; In re Basciano, 542 F.3d 950, 956 (2d Cir.2008), cert. denied, 555 U.S. 1177, 129 S.Ct. 1401, 173 L.Ed.2d 596 (2009); United States v. Wecht, 484 F.3d 194, 213 (3d Cir.2007) ("`The test for recusal under § 455(a) is whether a reasonable person, with knowledge of all the facts, would conclude that the judge's impartiality might reasonably be questioned.'"); United States v. Lauersen, 348 F.3d 329, 334 (2d Cir.2003) ("Disqualification under section 455(a) requires a showing that would cause `an objective, disinterested observer fully informed of the underlying facts [to] entertain significant doubt that justice would be done absent recusal.'"), cert. denied, 541 U.S. 1044, 124 S.Ct. 2190, 158 L.Ed.2d 735 (2004).16 Thus,

United States v. Bayless, 201 F.3d 116, 126-27 (2d Cir.), cert. denied, 529 U.S. 1061, 120 S.Ct. 1571, 146 L.Ed.2d 474 (2000).17 "The Court of Appeals has cautioned that when answering this question [i.e., the objective test], `the grounds asserted in a recusal motion must be scrutinized with care, and judges should not recuse themselves solely because a party claims an appearance of partiality.'" Barnett v. United States, 2012 WL 1003594 at *1.

To establish a basis for recusal, "[m]ovants must overcome a presumption of impartiality, and the burden for doing so is substantial." Metro. Opera Ass'n, Inc. v. Local 100, Hotel Emps. Int'l Union, 332 F.Supp.2d at 670 (quotations omitted); accord, e.g., United States v. Denton, 434 F.3d 1104, 1111 (8th Cir.2006) ("A judge is presumed to be impartial, and `the party seeking disqualification bears the substantial burden of proving otherwise.'").18

"Discretion is confided in the district judge in the first instance to determine whether to disqualify himself. The reasons for this are plain. The judge presiding over a case is in the best position to appreciate the implications of those matters alleged in a recusal motion." In re Drexel Burnham Lambert Inc., 861 F.2d at 1312 (citation omitted); accord, e.g., In re Basciano, 542 F.3d at 956; In re Certain Underwriter, 294 F.3d 297, 302 (2d Cir.2002); United States v. Roldan-Zapata, 916 F.2d 795, 802 (2d Cir.1990), cert. denied, 499 U.S. 940, 111 S.Ct. 1397, 113 L.Ed.2d 453 (1991).19 "[A] judge has an affirmative duty ... not to disqualify himself unnecessarily, particularly `where the request for disqualification was not made at the threshold of the litigation and the judge has acquired a valuable background of experience.'" Nat'l Auto Brokers Corp. v. Gen. Motors Corp., 572 F.2d 953, 958 (2d Cir.1978), cert. denied, 439 U.S. 1072, 99 S.Ct. 844, 59 L.Ed.2d 38 (1979); accord, e.g., LoCascio v. United States, 473 F.3d 493, 498 (2d Cir.), cert. denied, 552 U.S. 1010, 128 S.Ct. 554, 169 L.Ed.2d 374 (2007); United States v. Yonkers Bd. of Educ., 946 F.2d 180, 183 (2d Cir.1991). "A judge is as much obliged not to recuse himself when it is not called for as he is obliged to when it is." In re Drexel Burnham Lambert Inc., 861 F.2d at 1312; accord, e.g., In re Literary Works in Elec. Databases Copyright Litig., 509 F.3d 136, 140 (2d Cir.2007); Metro. Opera Ass'n, Inc. v. Local 100, Hotel Emps. Int'l Union, 332 F.Supp.2d at 670; In re Certain Underwriter, 294 F.3d at 302.

"In deciding whether to recuse himself, the trial judge must carefully weigh the policy of promoting public confidence in the judiciary against the possibility that those questioning his impartiality might be seeking to avoid the adverse consequences of his presiding over their case." In re Drexel Burnham Lambert Inc., 861 F.2d at 1312; accord, e.g., In re Basciano, 542 F.3d at 956; In re Certain Underwriter, 294 F.3d at 302.20 This is because "[l]itigants are entitled to an unbiased judge; not to a judge of their choosing." In re Drexel Burnham Lambert Inc., 861 F.2d at 1312; accord, e.g., Mulligan v. Loschiavo, 173 Fed.Appx. 26, 28 (2d Cir.2006).21 Moreover,

Metro. Opera Ass'n, Inc. v. Local 100, Hotel Emps. Int'l Union, 332 F.Supp.2d at 670.22

II. DA SILVA MOORE'S RECUSAL MOTION IS UNTIMELY

A. Legal Standard

Although 28 U.S.C. § 455 does not explicitly address timeliness, such a requirement "has been read into this section," addressing two underlying concerns. Apple v. Jewish Hosp. & Med. Ctr., 829 F.2d 326, 333 (2d Cir.1987); accord, e.g., Taylor v. Vt. Dep't of Educ., 313 F.3d 768, 794 (2d Cir.2002) (Sotomayor, C.J.); United States v. Brinkworth, 68 F.3d 633, 639 (2d Cir.1995); Polizzi v. United States, 926 F.2d 1311, 1321 (2d Cir.1991).23 "First, judicial resources should not be wasted; and, second, a movant may not hold back and wait, hedging its bets against the eventual outcome." Apple v. Jewish Hosp. & Med. Ctr., 829 F.2d at 334; accord, e.g., LoCascio v. United States, 473 F.3d 493, 497 (2d Cir.2007), cert. denied, 552 U.S. 1010, 128 S.Ct. 554, 169 L.Ed.2d 374 (2007); United States v. Brinkworth, 68 F.3d at 639; Gil Enters., Inc. v. Delvy, 79 F.3d 241, 247 (2d Cir.1996); In re IBM Corp., 45 F.3d 641, 643 (2d Cir.1995) (A "prompt application avoids the risk that a party is holding back a recusal application as a fall-back position in the event of adverse rulings on pending matters."); United States v. Yonkers Bd. of Educ., 946 F.2d 180, 183 (2d Cir.1991).24

Recusal motions must be made "at the earliest possible moment after obtaining knowledge of facts demonstrating the basis for such a claim." Apple v. Jewish Hosp. & Med. Ctr., 829 F.2d at 333; accord, e.g., Weisshaus v. Fagan, 456 Fed. Appx. 32, 34 (2d Cir.2012); United States v. Amico, 486 F.3d 764, 773 (2d Cir.2007).25 Courts have held that:

Universal City Studios, Inc. v. Reimerdes, 104 F.Supp.2d at 349-50; accord, e.g., United States v. Daley, 564 F.2d 645, 651 (2d Cir.1977) (motion for recusal untimely because, inter alia, the facts upon which it was based "as a matter of public record, were at all times ascertainable by counsel"), cert. denied, 435 U.S. 933, 98 S.Ct. 1508, 55 L.Ed.2d 530 (1978).26

In deciding whether a recusal motion is timely, a court looks to a number of factors, including whether: "(1) the movant has participated in a substantial manner in trial or pre-trial proceedings; (2) granting the motion would represent a waste of judicial resources; (3) the motion was made after the entry of judgment; and (4) the movant can demonstrate good cause for delay." Apple v. Jewish Hosp. & Med. Ctr., 829 F.2d at 334 (citations omitted); accord, e.g., Weisshaus v. Fagan, 456 Fed.Appx. at 34; United States v. Amico, 486 F.3d at 773; Taylor v. Vt. Dep't of Educ., 313 F.3d at 794-95; United States v. Brinkworth, 68 F.3d at 639.27

B. Application

Plaintiffs' seek my recusal due to my advocacy of predicative coding, relationship with Losey and speaking engagements at LegalTech. (See generally Dkt. No. 170: Pls. Br.; Dkt. No. 192: Pls. Reply Br.) Plaintiffs and defendants, however, have been aware of my view on predicative coding since at least December 2, 2011, and my relationship with Losey and engagement with LegalTech since January 4, 2012. (See page 143 above.)

During the first conference on December 2, 2011, I made the parties aware of my knowledge of predictive coding and recommended that parties read my Search, Forward article. (See pages 140-41 above.) Specifically, I said:

(See page 141 above, emphasis added.) Indicating that I am familiar with the ediscovery industry and know several people involved, when discussing whether a special master would be useful, I said, "You know, I know enough people in the industry that I can recommend some, or you all can get your vendors to recommend somebody...." (See page 141 n. 1 above.) I also made the comment, "You must have thought you died and went to Heaven when this was referred to me" (see page 140 above), and MSL's counsel responded that counsel was "just thankful that ... we have a person familiar with the predictive coding concept" (see page 140 above).

On December 20, 2011, I granted plaintiffs' request for an adjournment of the upcoming discovery conference, and in doing so reminded the parties to read my article, stating, "As to predictive coding, you should read my article, `Search, Forward' in the Oct. 2011 issue of Law Technology News." (See page 142 above.)

During the January 4, 2012 conference, while discussing substitute counsel for MSL counsel Anders during his vacation, I made the parties aware that I knew Ralph Losey, a partner in MSL's counsel's ediscovery practice group. (See page 143 above.) The following colloquy occurred:

(Dkt. No. 71: 1/4/12 Conf. Tr. at 61, emphasis added; see page 143 above.) Also at the January 4 conference while setting the date for the next conference, I implicitly made it known that I would be unavailable due to my speaking engagements at LegalTech. (See page 143 above.) The following colloquy occurred:

(1/4/12 Conf. Tr. at 66; see page 143 above.) Plaintiffs' ediscovery consultant, DOAR, is very familiar with the Legal Tech conferences. (See page 143 n. 6 above.)

Plaintiffs argue that their motion should "not be denied as untimely [because]: (1) the case remains in its early stages; (2) granting the motion would not represent a waste of judicial resources; (3) the motion was not made after entry of judgment; and (4) Plaintiffs have acted promptly upon discovering the relevant facts." (Pls. Reply Br. at 6.)

With respect to the first factor, plaintiffs have been active participants in pretrial proceedings since before my first conference on December 2, 2011. While plaintiffs are correct that the case is in its early stages, discovery has been ongoing since at least October 2011 (see Dkt. No. 44: 10/12/11 Order), if not before. At conferences and through written communications, plaintiffs have enlisted both myself and the District Judges with respect to the ediscovery protocol, adjournments, issues arising during depositions, stipulations and other pretrial issues. (See, e.g., Dkt. No. 88: 2/8/12 Conf. Tr.; Dkt. No. 91: 2/2/12 Telephone Conf. Tr.; Dkt. No. 92: 2/17/12 ESI Protocol & Order; Dkt. No. 108: 3/8/12 Stipulation & Order; Dkt. No. 109: Am. Joint Scheduling Order; Dkt. No. 118: 3/19/12 Stipulation & Order; Dkt. No. 128: 3/28/12 Memo Endorsed Order; Dkt. No. 209: 3/9/12 Conf. Tr.; Da Silva Moore v. Publicis Groupe, 11 Civ. 1279, ___ F.R.D. ___, 2012 WL 607412 (S.D.N.Y. Feb. 24, 2012) (Peck, M.J.), adopted, 2012 WL 1446534 (S.D.N.Y. Apr. 26, 2012).)

With respect to the second factor, this case was referred to me on November 28, 2011 for general pretrial supervision. (See page 140 above.) Since that time, I have familiarized myself with this case and have expended considerable time and attention in responding to the parties' discovery issues and disputes. While plaintiffs assert that "granting the motion would not represent a waste of judicial resources" (Pls. Reply Br. at 6), another magistrate judge would have to spend ample time to familiarize himself or herself with this complex case. Moreover, plaintiffs have stated that if recusal is granted, they will ask Judge Carter to review and overturn my decisions and orders in the case. (Pls. Br. at 25.)

With respect to the third factor, while it is true that, as plaintiffs assert, "the motion was not made after entry of judgment" (Pls. Reply Br. at 6), it is also true that plaintiffs waited to seek my recusal until after I adopted MSL's predictive coding protocol (see pages 147-48 above). Plaintiffs themselves assert that "[p]redictive coding without proper safeguards will impact the merits of this case." (Pls. Reply Br. at 7.) Courts in the Second Circuit have required "a prompt application [to avoid] the risk that a party is holding back a recusal application as a fall-back position in the event of adverse rulings on pending matters." In re IBM Corp., 45 F.3d 641, 643 (2d Cir.1995). It appears that plaintiffs are improperly using the recusal motion as "fall-back position" to an unfavorable ruling. See, e.g., Weisshaus v. Fagan, 456 Fed.Appx. 32, 34 (2d Cir.2012) (Motion held untimely. "Although there was no dispositive ruling as to [defendant] at the time [plaintiff] brought her recusal motion, the district court aptly noted that the motion came on the heels of its direction that [plaintiff] submit to a deposition, thus strongly suggesting that the motion was a mere fall-back position in response to an adverse ruling."); Silver v. Kuehbeck, 217 Fed. Appx. 18, 23-24 (2d Cir.2007) (Motion held untimely. "[A]fter the April 18, 2005 settlement conference, from which [plaintiff] was `left with no choice but to assume that what transpired ... may have colored' the district court's perception, [plaintiff] continued to litigate the case for two more months, even appearing before the Court for oral arguments on the motion to dismiss on May 23 without requesting recusal. It thus appears that [plaintiff] held back his `recusal application as a fall-back position in the event of adverse rulings on pending matters.'"); LoCascio v. United States, 473 F.3d 493, 497 (2d Cir.2007) (recusal motion untimely where plaintiff "made no mention of the above remark until after the District Court had denied his motion to amend and after it had denied his § 2255 petition."), cert. denied, 552 U.S. 1010, 128 S.Ct. 554, 169 L.Ed.2d 374 (2007); Gil Enters., Inc. v. Delvy, 79 F.3d 241, 247 (2d Cir.1996) ("In the face of defeat, [plaintiff] now seeks to do just that which this Court warned against in In re IBM and use its post-hoc recusal motion as a `fall-back position.' Since [plaintiff]'s objection was raised well later than `the earliest possible moment,' and because [plaintiff] has failed to demonstrate any bias on the part of the district court beyond [the Judge]'s own expressed frustration, we reject this ground for relief"); Armenian Assembly of Am., Inc. v. Cafesjian, 783 F.Supp.2d 78, 88 (D.D.C.2011) ("[I]t appears that the filing of the [recusal] motion was motivated by the fact that Plaintiffs received a largely adverse decision from the Court rather than by the sudden discovery of an alleged bias."); Katzman v. Victoria's Secret Catalogue, 939 F.Supp. 274, 278 (S.D.N.Y.1996) (Plaintiff "and her counsel now seek to do just that which the Second Circuit and other courts have warned against; namely, use a post-judgment recusal motion to try to get a second bite at the apple."), aff'd, 113 F.3d 1229 (2d Cir.1997).

With respect to the fourth factor, plaintiffs assert that they "have acted promptly upon discovering the relevant facts" but provide no explanation for their delay. (See Pls. Reply Br. at 6.) Despite plaintiffs' knowledge as of December 2, 2011 of my views on predicative coding, and by January 4, 2012 as to my relationship with Losey and my speaking at LegalTech (see pages 140-41, 143 above), plaintiffs did not request my recusal until March 28, 2012 by letter (see pages 147-48 above) and did not file their formal recusal motion until April 13, 2012 (see page 148 above). The movant "is charged with knowledge of all facts `known or knowable, if true, with due diligence from the public record or otherwise.'" Universal City Studios, Inc. v. Reimerdes, 104 F.Supp.2d 334, 349 (S.D.N.Y.2000); accord, e.g., Armenian Assembly of Am., Inc. v. Cafesjian, 783 F.Supp.2d at 87 (timeliness rule "`has been applied when the facts upon which the [recusal] motion relies are public knowledge, even if the movant does not know them.'"). Plaintiffs here had the requisite knowledge no later than January 4, 2012, but the recusal request did not come until nearly three months later. (See 147-48 above.) Courts have found shorter delays to be untimely. See, e.g., Apple v. Jewish Hosp. & Med. Ctr., 829 F.2d 326, 334 (2d Cir.1987) (motion untimely where party waited two months after events giving rise to charge of bias or prejudice before making its recusal motion, despite fact that other of the "factors do not support a finding of untimeliness"); Six W. Retail Acquisition, Inc. v. Sony Theatre Mgmt. Corp., 97 Civ. 5499, 2003 WL 282187 at *1, *4 (S.D.N.Y. Feb. 7, 2003) (two-month delay untimely), aff'd, 124 Fed. Appx. 73 (2d Cir.), cert. denied, 546 U.S. 1016, 126 S.Ct. 660, 163 L.Ed.2d 526 (2005); Katzman v. Victoria's Secret Catalogue, 939 F.Supp. at 278 (two-month delay untimely); Lamborn v. Dittmer, 726 F.Supp. 510, 515 (S.D.N.Y.1989) ("Recusal motions are often denied on the basis of untimeliness where there has been only a short delay." (citing, inter alia, In re Martin-Trigona, 573 F.Supp. 1237, 1244-45 (D.Conn.1983) (motion untimely based on twelve-day delay), appeal dismissed, 770 F.2d 157 (2d Cir.1985), cert. denied, 475 U.S. 1058, 106 S.Ct. 1285, 89 L.Ed.2d 592 (1986))). I have made no efforts to hide my views, relationships or affiliations. If plaintiffs truly believed that any of these issues, individually or collectively, created a bias or the appearance of partiality, they should have promptly moved for my recusal.

Accordingly, plaintiffs' recusal motion is untimely. In any event, it also is meritless, as will be discussed in the following sections.

III. DA SILVA MOORE'S RECUSAL MOTION IS MERITLESS

Plaintiffs assert that my recusal is required due to: (1) my "public comments concerning the case" (Dkt. No. 170: Pls. Br. at 12-14; Dkt. No. 192: Pls. Reply Br. at 8-9), (2) my "participation on pro-predictive coding panels with defense counsel Ralph Losey while presiding over the parties' dispute on predictive coding" (Pls. Br. at 14-15; Pls. Reply Br. at 7-8), (3) my "numerous speaking engagements in favor of predictive coding, which were at least indirectly sponsored and funded by Recommind and other e-discovery vendors" (Pls. Br. at 15-17), and (4) my "failure to disclose [my] activities enhances the appearance of impropriety" (Pls. Br. at 19-21).

A. Speaking Engagements

With respect to my speaking engagements on the subject of computer-assisted review, I only spoke generally about computer-assisted review in comparison to other search techniques. (See pages 143-44 above.) The fact that my interest in and knowledge about predictive coding in general overlaps with issues in this case is not a basis for recusal. See, e.g., Hoatson v. N.Y. Archdiocese, 280 Fed.Appx. 88, 90 (2d Cir.2008) (In a case involving the Catholic Church, fact that the district judge was "member of the Guild of Catholic Lawyers of the Archdiocese of New York," had received an award from that organization and occasionally attended its meetings discussing Catholic education and Catholic commitment to social justice was not a basis for appearance of impropriety recusal.); Hu v. Am. Bar Ass'n, 334 Fed.Appx. 17, 19 (7th Cir.2009) (A "judge's membership in a bar association, or his receipt of reimbursement for participating in bar-association activities, does not create the type of relationship that would cause us to doubt his ability to preside impartially over a case in which the bar association is a party."); Lunde v. Helms, 29 F.3d 367, 370-71 (8th Cir.1994) (no basis for recusal where judge was an alumnus of the university defendant, had made financial contributions to the university and had participated in the university's educational programs), cert. denied, 513 U.S. 1155, 115 S.Ct. 1111, 130 L.Ed.2d 1076 (1995); Wu v. Thomas, 996 F.2d 271, 275 (11th Cir.1993) (recusal not required where judge made past contributions to the university defendant and held position as unsalaried adjunct professor), cert. denied, 511 U.S. 1033, 114 S.Ct. 1543, 128 L.Ed.2d 195 (1994); Sierra Club v. Simkins Indus., Inc., 847 F.2d 1109, 1117-18 (4th Cir.1988) (judge's Sierra Club membership before appointment to the bench did not require recusal from case where the Sierra Club was a party), cert. denied, 491 U.S. 904, 109 S.Ct. 3185, 105 L.Ed.2d 694 (1989); United States v. Alabama, 828 F.2d 1532, 1543-44 (11th Cir.1987) ("all judges come to the bench with a background of experiences, associations and viewpoints.... A judge is not required to recuse himself merely because he holds and has expressed certain views on a general subject." Judge's background as civil rights lawyer and state legislator did not require disqualification in desegregation case, but other factors did.), cert. denied, 487 U.S. 1210, 108 S.Ct. 2857, 101 L.Ed.2d 894 (1988); Shaw v. Martin, 733 F.2d 304, 316 (4th Cir.) ("One who has voted as a legislator in favor of a statute permitting the death penalty in a proper case cannot thereafter be presumed disqualified to hear capital cases as a judge or predisposed to give a death sentence in any particular case."), cert. denied, 469 U.S. 873, 105 S.Ct. 230, 83 L.Ed.2d 159 (1984).28

While speaking on ediscovery panels in January 2012, I mentioned that I had a case in front of me using computer-assisted review, but I did not mention the parties or counsel involved. (See pages 144-45 & n. 11 above.) The only arguably identifiable statement was the "died and went to Heaven" comment. (See pages 144-45 & n. 11 above.) This comment, however, was originally made in open court at the December 2, 2011 conference and was available in the public transcript. (See pages 140-41 above.) Consequently, this is not a recusable statement. See, e.g., United States v. Pitera, 5 F.3d 624, 626-27 (2d Cir.1993) (upholding district court's denial of recusal motion by defendant in narcotics prosecution, where judge had lectured to the DEA Task Force including advice on steps to take to increase convictions, but also spoke at a PLI program for criminal defense lawyers, and the "record discloses that the Judge commendably lectures to a variety of trial practice seminars"), cert. denied, 510 U.S. 1131, 114 S.Ct. 1103, 127 L.Ed.2d 415 (1994); United States v. Yonkers Bd. of Educ., 946 F.2d 180, 184-85 (2d Cir.1991) (finding "no impropriety" in the District Judge's public comments in the media about a pending case because the judge "only restated what he had been saying in open court for the past few years and did not discuss the details of remedy implementation"); Wilborn v. Wells Fargo Bank, N.A. (In re Wilborn), 401 B.R. 848, 863 (Bankr. S.D.Tex.2009) ("Here, the undersigned judge's disagreement with certain legal arguments previously rejected by this Court and others, expressed at the Dallas Seminar, are not sufficient to warrant recusal."); Metro. Opera Ass'n, Inc. v. Local 100, Hotel Emps. Int'l Union, 332 F.Supp.2d 667, 674-75 (S.D.N.Y.2004) ("[T]he summary of factual findings from the Opinion, used as a springboard in the [CLE] Presentation to discuss `best practices' in electronic discovery to avoid such findings, is insufficient to require recusal, even when a motion to reconsider was pending.").29

The Metropolitan Opera case is particularly instructive. There, Judge Preska issued a "lengthy opinion" granting judgment for plaintiff and awarding it attorneys' fees based on defendants' discovery abuses. See Metro. Opera Ass'n, Inc. v. Local 100, Hotel Emps. Int'l Union, 332 F.Supp.2d at 669. While a motion for reconsideration was pending, Judge Preska gave a presentation at a BNA CLE program "on the topic of electronic discovery and how technological advances might affect discovery obligations." Id. at 669-70. "The Presentation began with a summary of selected discovery failures set out in the Opinion, ... and proceeded to recommend ... steps to take to avoid these and other pitfalls in electronic discovery, including steps that echoed findings in the Opinion." Id. Judge Preska denied defendants' § 455(a) disqualification motion. Id. at 676. Movants "argue[d] that a few of the phrases in the course of the Presentation are indicative of [the judge's] `personal involvement' or `emotional commitment' to one side of the dispute," including reference to a colloquial comment that the [defendant] "`junked'" some of its computers. Id. at 672-73. Judge Preska responded: "Whether or not my subjective pedagogical intent was successfully conveyed, the remarks complained of could not be interpreted by the objective disinterested observer as conveying the level of personal involvement required for recusal." Id. at 673. Judge Preska concluded:

Id. at 674-75, 676 (citations omitted).

My comments were nowhere near the comments in Metropolitan Opera, which were not a basis for recusal. I did not mention this case by name. (See pages 143-44 above.) While I briefly mentioned that I had a case in front of me where a party proposed using predictive coding (and noted that, on the record, I had made the "died and went to heaven" comment), the case was not yet the subject of publicity (which only resulted after the February 13, 2012 press release by plaintiffs' consultant DOAR) and the objective reasonable observer would not have known that those brief comments referred to this case. (See pages 144, 145-47 above.) Moreover, the comments were minor and fleeting, involved only facts on the public record and, like Metropolitan Opera, were for educational purposes.

To the extent plaintiffs are complaining about my general discussion at these CLE presentations about the use of predictive coding in general, those comments would not cause a reasonable objective observer to believe I was biased in this case. I did not say anything about predictive coding at these LegalTech and other CLE panels that I had not already said in my Search, Forward article, i.e., that lawyers should consider using predictive coding in appropriate cases. My position was the same as plaintiffs' consultant, DOAR CEO Neale. (See pages 145-46 above.) Both plaintiffs and defendants were proposing using predictive coding in this case. (See page 142 above.) I did not determine which party's predictive coding protocol was appropriate in this case until the February 8, 2012 conference, after the panels about which plaintiffs complain.30

In objecting to my February 8 ruling, plaintiffs informed Judge Carter that while predictive coding may be appropriate under certain circumstances, "the devil is in the details." (See page 147 above.)31 I did not discus the "details" of a predictive coding protocol (e.g., number of iterations needed to train the computer, how many "seed" documents would be used, appropriate sample size) at any of the LegalTech or other CLE panels (or in my Search, Forward article), nor did any of the other panelists. Thus, a reasonable objective observer would not think that my comments at these educational panels gives the appearance of bias for MSL or against plaintiffs. My participation in the panels, like Judge Preska's CLE presentation in Metropolitan Opera, does not require recusal.

B. Counsel Losey & Recommind's Participation in LegalTech

While I participated on two panels with defense counsel Losey, we never had any ex parte communication regarding this lawsuit. (See pages 143-44 above.) My preparation for and participation in ediscovery panels involved only ediscovery generally and the general subject of computer-assisted review. (See pages 143-44 above.) Losey's affidavit makes clear that we have never spoken about this case, and I confirm that. (See pages 143-44 above.) During the panel discussions (and preparation sessions), there was absolutely no discussion of the details of the predictive coding protocol involved in this case or with regard to what a predicative coding protocol should look like in any case. (See pages 143-44 above.) Plaintiffs' assertion that speaking on an educational panel with counsel creates an appearance of impropriety is undermined by Canon 4 of the Judicial Code of Conduct, which encourages judges to participate in such activities. See Code of Conduct for United States Judges, Canon 4 ("A judge may engage in extrajudicial activities, including law-related pursuits and ... educational... activities, and may speak, write, lecture, and teach on both law-related and nonlegal subjects."), available at http://www.uscourts.gov/Viewer.aspx?doc=/uscourts/RulesAndPolicies/conduct/Vol02A-Ch02.pdf (last visited June 14, 2012).32

The cases make clear that participation on an educational panel with counsel is not a basis for recusal, for sound policy reasons. See, e.g., Leja v. Schmidt Mfg., Inc., No. Civ. 01-5042, 2010 WL 2571850 at *2 (D.N.J. June 22, 2010) ("In my own case, during the course of 31 years on the bench, I have developed numerous personal friendships with members of the Bar and have participated in many charitable, legal and public service organizations in which lawyers, law firms and other judges have participated. This is probably the experience of most judges. To permit such associations to become grounds for recusal would either push judges towards a hermit like existence or open the floodgates to recusal motions."); In re Wolverine Proctor & Schwartz, LLC, 397 B.R. 179, 183 (Bankr.D.Mass.2008) ("[T]he Court rejects [creditor's] assumption that [the judge's] service on the Financial Literacy Committee with [counsel], and numerous other volunteer lawyers and fellow judges, gives rise to a disqualifying connection and establishes reasonable grounds for doubting this Court's impartiality."); In re Healy, No. 04-28375-D-13L, 2006 WL 3751617 at *4 (Bankr.E.D.Cal. Dec. 18, 2006) ("It is common knowledge, of course, that judges regularly appear on panels and at presentations for members of the bar, and that such events are regularly advertised in various publications that might be viewed by both the public and the bar. But it is not reasonable to conclude that the participation of a judge with members of the bar who appear before the judge's court would create a predisposition, or an appearance of a predisposition, to favor the members of the bar who participate over those who do not. The Debtor's assertion of an appearance of impropriety is undermined by Canon 4 of the Code of Conduct, which not only permits judges to participate in such activities, but encourages judges to do so."); Moran v. Clarke, 213 F.Supp.2d 1067, 1073 (E.D.Mo.2002) ("A judge's involvement with other attorneys in bar association activities is not a basis for recusal. Indeed, the commentary to Canon 4 ... encourages judges to `contribute to the improvement of the law, the legal system, and the administration of justice.... [T]he judge is encouraged to do so, either independently or through a bar association, judicial conference, or other organization dedicated to the improvement of the law.' A judge should not be required to withdraw from all social relationships and live in seclusion.'"); Bailey v. Broder, 94 Civ. 2394, 1997 WL 73717 at *2 (S.D.N.Y. Feb. 20, 1997) ("If my `relationship' with [counsel (because of bar association and court-related social activities)] were to require recusal under the instant circumstances, I daresay recusal would be required in so many other cases because of my acquaintance with an attorney that I — and probably most other judges — would be unable to function in our jobs."); Rosado v. Bridgeport Roman Catholic Diocesan Corp., 292 Conn. 1, 970 A.2d 656, 671 (Court "can conceive of no reason to depart from the rule that [a judge's] membership in a task force concerning a particular legal issue does not justify disqualification of a judge simply because the judge's service happens to be coincident with his participation in a case dealing with the same issue."), cert. denied, ___ U.S. ___, 130 S.Ct. 500, 175 L.Ed.2d 348 (2009).33

Plaintiffs further assert that I failed to notify them of my ex parte communications with defense counsel Losey. (Dkt. No. 170: Pls. Br. at 7-8, 14-15, 19-21.) Canon 3(A)(4) provides: "a judge should not initiate, permit, or consider ex parte communications or consider other communications concerning a pending or impending matter that are made outside the presence of the parties or their lawyers." Code of Conduct for United States Judges, Canon 3(A)(4), available at http://www.uscourts.gov/Viewer.aspx?doc=/uscourts/RulesAndPolicies/conduct/Vo102A-Ch02.pdf (last visited June 14, 2012). Here, I had no ex parte communications with Losey "concerning a pending or impending matter," and thus nothing to disclose to plaintiffs. (See pages 143-44 above.) Moreover, I never tried to hide the fact that I knew Losey. I specifically advised counsel that I knew Losey "very well" and also alluded to my engagement with LegalTech during the January 4, 2012 conference. (See page 143 above.) Had plaintiffs been concerned, they could have followed up with the Court. Thus, I had no duty to notify the parties.

Plaintiffs cite to Pfizer Inc. v. Kelly (In re School Asbestos Litig.), 977 F.2d 764, 782 (3d Cir.1992), in furtherance of their argument that my "participation on pro-predictive coding panels with defense counsel Ralph Losey while presiding over the parties' dispute on predictive coding requires [my] recusal." (Pls. Br. at 14-15.) In re School Asbestos Litig. is inapposite because in that case the judge:

Pfizer Inc. v. Kelly (In re School Asbestos Litig.), 977 F.2d at 782. The panels in which I participated are distinguishable. First, I was a speaker at educational conferences, not an audience member. Second, the conferences were not one-sided, but concerned ediscovery issues including search methods in general. Third, while Recommind was one of thirty-nine sponsors and one of 186 exhibitors contributing to LegalTech's revenue (see page 143 above), I had no part in approving the sponsors or exhibitors (i.e., funding for LegalTech) and received no expense reimbursement or teaching fees from Recommind or LegalTech, as opposed to those companies that sponsored the panels on which I spoke. Fourth, there was no "pre-screening" of MSL's case or ediscovery protocol; the panel discussions only covered the subject of computer-assisted review in general.34

Plaintiffs assert that my "numerous speaking engagements in favor of predictive coding, which were at least indirectly sponsored and funded by Recommind and other e-discovery vendors, mandate recusal." (Pls. Br. at 15-17.) Plaintiffs correctly assert that:

(Pls. Br. at 16 (quoting Aguinda v. Texaco, Inc. (In re Aguinda), 241 F.3d 194, 206 (2d Cir.2001)).) Plaintiffs, however, fail to appreciate the difference between the facts in In re Aguinda and this case. Plaintiffs assert that Recommind indirectly sponsored me at LegalTech. (Pls. Br. at 15-17, 23, 24-25; Pls. Reply Br. at 7.) LegalTech New York 2012 had thirty-nine sponsors, 186 exhibitors, and hundreds of paying attendees. (See page 143 above.) Unlike In re Aguinda, LegalTech did not receive a "significant portion of [its] general funding from litigants or counsel" or Recommind. Compare Aguinda v. Texaco, Inc. (In re Aguinda), 241 F.3d at 206. Recommind sponsored the Emerging Technology and Information Governance tracks; my panels were on tracks sponsored by other companies. (See page 143 n. 9 above.) Thus, "the source of the payments [do] not give the appearance of influencing the judge in the judge's judicial duties or otherwise give the appearance of impropriety," Code of Conduct for United States Judges, Canon 4(H), available at http://www.uscourts.gov/Viewer.aspx?doc=/uscourts/RulesAndPolicies/conduct/Vo102A-Ch02.pdf (last visited June 14, 2012), because Recommind's support of LegalTech was one minor source of Legal Tech's funding and was not affiliated with the panels

My comments at LegalTech (and all other ediscovery conferences at which I recently was a speaker) addressed my interest in and support for computer-assisted review in general in cases where it is appropriate. (See pages 143-44 above.) My comments were in-line with my Search, Forward article also discussing the use of predictive coding in appropriate cases. (See pages 141-42 above.) "Engaging in such law-related activities — including speeches that comment on current events and legal developments-is permitted not only because judges are citizens, but because they are particularly knowledgeable on such topics." In re Judicial Misconduct, 632 F.3d 1289, 1289 (9th Cir. 2011). My comments about computer-assisted review are supported by Canon 4 of the Code of Conduct that encourages judges to "speak, write, lecture, teach, and participate in other activities concerning the law, the legal system, and the administration of justice." Code of Conduct for United States Judges Canon 4(A)(1), available at http://www.uscourts.gov/Viewer.aspx?doc=/uscourts/RulesAndPolicies/conduct/Vol02A-Ch02.pdf (last visited June 14, 2012).

There are probably fewer than a dozen federal judges nationally who regularly speak at ediscovery conferences. Plaintiffs' argument that a judge's public support for computer-assisted review is a recusable offense (Pls. Br. at 12-17, 23; Pls. Reply Br. at 4-6) would preclude judges who know the most about ediscovery in general (and computer-assisted review in particular) from presiding over any case where the use of predictive coding was an option, or would preclude those judges from speaking at CLE programs. Plaintiffs' position also would discourage lawyers from participating in CLE programs with judges about ediscovery issues, for fear of subsequent motions to recuse the judge (or disqualify counsel). Taken further, it would preclude any judge who speaks at a CLE conference about any ediscovery subject from handling future cases involving ediscovery. Such a position defies logic and is inconsistent with the Code of Conduct for United States Judges. See, e.g., Wilborn v. Wells Fargo Bank, N.A. (In re Wilborn), 401 B.R. at 871-72 ("The Committee on Codes of Conduct of the Judicial Conference of the United States clearly desires that federal judges be free to convey their legal opinions at CLE presentations so that practitioners may benefit from their experience and perspective. As such, the Code of Conduct for United States Judges expressly allows for, and encourages, federal judges to engage in `scholarly presentation[s] made for the purposes of legal education.' At both the Dallas Seminar and the LSU Seminar, the undersigned judge did exactly that. To recuse himself in this suit would therefore be improper." (citations omitted)).

Moreover, in Samuel v. Univ. of Pittsburgh, the court distinguished the difference between a judge's personal philosophy and discussions of opinion regarding a particular case:

Samuel v. Univ. of Pittsburgh, 395 F.Supp. at 1278. Here, my comments at ediscovery conferences related to the general use of predictive coding in appropriate cases, and I did not express any opinion regarding the specific issues in this case. Consequently, neither my comments nor the fact that Losey was on some panels with me, nor the fact that MSL's vendor Recommind sponsored different panels at LegalTech, separately or collectively, are a basis for recusal.

IV. RECUSAL IS NOT REQUIRED BASED ON MY IN-COURT COMMENTS

A. Legal Standard

"Generally, claims of judicial bias must be based on extrajudicial matters, and adverse rulings, without more, will rarely suffice to provide a reasonable basis for questioning a judge's impartiality." Chen v. Chen Qualified Settlement Fund, 552 F.3d 218, 227 (2d Cir.2009).35 As the Supreme Court has noted:

Liteky v. United States, 510 U.S. 540, 555, 114 S.Ct. 1147, 1157, 127 L.Ed.2d 474 (1994) (citation omitted).36 The Supreme Court stressed that "[n]ot establishing bias or partiality ... are expressions of impatience, dissatisfaction, annoyance, and even anger, that are within the bounds of what imperfect men and women, even after having been confirmed as federal judges, sometimes display." Liteky v. United States, 510 U.S. at 555-56, 114 S.Ct. at 1157.37 Additionally, a "judge's ordinary efforts at courtroom administration-even a stern and short-tempered judge's ordinary efforts at courtroom administration-remain immune." Liteky v. United States, 510 U.S. at 556, 114 S.Ct. at 1157.38

"A judge is not to be faulted as biased or prejudiced because he has considered that the effective discharge of his responsibility over proceedings before him... has demanded the consistent rejection of an attorney's contentions or strong measures to prevent what he regards as inexcusable waste of time." Rosen v. Sugarman, 357 F.2d 794, 798 (2d Cir.1966) (Friendly, C.J.).39 "Moreover, an occasional display of irritation ... does not suffice to show personal bias or prejudice, whether the irritation was justified or not." Rosen v. Sugarman, 357 F.2d at 798.40

B. Application

Plaintiffs argue for recusal alleging that I have taken "personal offense" to plaintiffs' filing of objections to my rulings and plaintiffs' filing of this recusal motion (Dkt. No. 170: Pls. Br. at 21-22), and that I have "chastised and yelled at Plaintiffs' counsel" and "intimidated Plaintiffs for disagreeing with rulings" (Dkt. No. 192: Pls. Reply Br. at 1-2). Plaintiffs' arguments lack merit.

Plaintiffs assert that my "repeated comments towards Plaintiffs may create an appearance of such partiality and — collectively with [my] other conduct — warrant recusal." (Pls. Br. at 21.)41 "A judge's comments that form the basis of a recusal motion should not be viewed in isolation, but rather must be viewed in context." Wilborn v. Wells Fargo Bank, N.A. (In re Wilborn), 401 B.R. 848, 860 (Bankr.S.D.Tex.2009).42 While I have been critical of plaintiffs' counsel at times, my criticism and resulting frustration are solely due to counsels' performance, and are not a basis for recusal.

Plaintiffs assert that I have "characterized Plaintiffs' objections as `whining' and [have] warned Plaintiffs against filing" objections. (Pls. Br. at 22, citing Dkt. No. 97: 2/8/12 Conf. Tr. at 20-22.)43 The referenced colloquy follows:

(2/8/12 Conf. Tr. at 20-21.) While I could have chosen my words more carefully, I was merely expressing my frustration with counsel Wipper's failure to answer my question. Expressions of impatience, dissatisfaction, annoyance, and even anger are not grounds for recusal. See, e.g., United States v. English, 629 F.3d 311, 321 (2d Cir.2011) (Judge's statement that "`I must tell you in a 27-kilo case I don't think I've ever let anybody out'" was not evidence of bias when the judge's bail denial decision was "explicitly tied to the facts before the court and w[as] fully explained on the record."); Gottlieb v. SEC, 310 Fed. Appx. 424, 425 (2d Cir.) (Recusal motion was without merit where the judge and plaintiff "clashed and had words in open court to the effect that Judge Preska said that `I am God in my courtroom' and [plaintiff] defiantly responded with `You are not God anywhere.'" (quotations omitted)), cert. denied, ___ U.S. ___, 130 S.Ct. 511, 175 L.Ed.2d 377 (2009); SEC v. Razmilovic, No. CV-04-2276, 2010 WL 2540762 at *5 (E.D.N.Y. June 14, 2010) ("Moreover, my characterization of [defendant]'s conduct in this action to date as `arrogance' does not demonstrate such a `deep-seated favoritism' of the SEC or `unequivocal antagonism' towards [defendant] so as to evidence my bias or partiality in this action."); cf. Metro. Opera Ass'n, Inc. v. Local 100, Hotel Emps. Int'l Union, 332 F.Supp.2d at 672 ("The fact that I used the colloquial term junked' instead of the word `dismantled' ... is hardly a basis for recusal.").

Moreover, I have expressed frustration with both sides in this case. Earlier in the same conference, I stated: "I've seen many a big case in this court go a lot more smoothly than this. As I say, I cannot speak to what happened before I inherited the case, but I expect cooperation. Stop the whining and stop the sandbagging. This goes for both sides. Get along." (2/8/12 Conf. Tr. at 13.)

Plaintiffs' assertion that I have sought to dissuade them from objecting to my rulings (Pls. Reply Br. at 1-2) is false. I have reminded the parties at nearly every conference, usually more than once, that they have the right to take their objections to Judge Carter. For example, at the April 25, 2012 conference I said, "And you have the right to take objections to Judge Carter, which you're not shy about, so take your objections. Stop arguing with me." (Dkt. No. 180: 4/25/12 Conf. Tr. at 34.)44

Plaintiffs argue for recusal alleging that the Court has labeled "Plaintiffs intransigent `scorched earth' litigants." (Pls. Br. at 22, citing Dkt. No. 158: 4/2/12 Order at 2.) Plaintiffs take this out of context. In response to plaintiffs' recusal letter, I instructed plaintiffs to "advise as to whether they wish to file a formal motion or for the Court to consider the letter as the motion" and concluded by stating that "If plaintiffs were to prevail [on recusal], it would serve to discourage judges ... from speaking on educational panels about ediscovery (or any other subject for that matter). The Court suspects this will fall on deaf ears, but I strongly suggest that plaintiffs rethink their `scorched earth' approach to this litigation." (4/2/12 Order at 1-2, quoted more fully on pages 147-48 above.) Courts often use the phrase "scorched earth" to describe hardball litigation tactics designed to complicate and prolong litigation and drive up litigation costs.45 My "scorched earth" comment provides no basis for recusal.

Plaintiffs further claim that I called their discovery requests "blackmail to convince the defendants to settle," and that I "intimated" that I "was waiting for Plaintiffs' `funding source' to `run out,' implying that [I] desired they drop the case and that [I] would put them through the ringer on discovery until they do so." (Pls. Reply Br. at 2.) Plaintiffs again take what I actually said out of context. At the April 25, 2012 conference, the following colloquy occurred:

(4/25/12 Conf. Tr. at 8-9.)46 The Court was reminding plaintiffs that they are only entitled to discovery related to either the named plaintiffs or to company policies to support a motion (not yet filed, much less granted) for class certification; plaintiffs are not entitled to discovery about individual potential class members until plaintiffs have moved for and been granted class certification. Plaintiffs cannot take class action discovery, at great expense to defendants, as if their class motion already was granted. The costs of extensive discovery have long been recognized as a factor forcing defendants to settle even meritless cases.47

Plaintiffs further assert that I "tried to bait and intimidate counsels' female lawyers into waiving Plaintiffs' legal rights on the proportionality issue." (Pls. Reply Br. at 2.) Plaintiffs allege that I "agreed that a particular `hot' document was relevant and should be coded as responsive, but that ruling came with a quid-pro-quo: Plaintiffs must waive their right to object to his ruling setting a hard cap on the number of documents they will get in discovery." (Pls. Reply Br. at 2.) Plaintiffs once again mis-characterize what occurred. During the May 7, 2012 conference (which lasted all day), plaintiffs' lead counsel Steven Wittels asked for permission to leave the conference. (Dkt. No. 194: 5/7/12 Conf. Tr. at 77-78.) The Court permitted him to leave after being assured that he was comfortable with his two associates handling the remainder of the conference in his absence. (5/7/12 Conf. Tr. at 78.) The Court and counsel later discussed whether an email that did not involve any named plaintiff or centralized decision-making, but rather another employee asking when she would be considered for a raise, should be coded as "relevant" for purposes of the "seed set" to train the predictive coding computer algorithm. I warned plaintiffs that if marginally-relevant documents about raises for people other than the named plaintiffs were coded as relevant, the predictive coding software would code similar documents (about individual, non-plaintiff employees) as relevant, possibly distorting what would be classified as the top-ranked documents. I stated that the email could be coded as relevant if plaintiffs were willing to take this risk, in terms of the possibility that the Court, applying Rule 26(b)(2)(C) proportionality, would agree with MSL to cut-off production at some point. (5/7/12 Conf. Tr. at 80-88.) In this context, the following colloquy occurred:

(5/7/12 Conf. Tr. at 86-88.)

In sum, my comments regarding plaintiffs' counsel Sanford Wittels & Heisler's handling of the case were neither unfounded48 nor so extreme to suggest bias against plaintiffs. See, e.g., Matrix Essentials, Inc. v. Quality King Distribs., Inc., 324 Fed.Appx. 22, 25 (2d Cir.2009) ("That the district judge was critical of some aspects of [plaintiff]'s presentation is not improper and does not justify reassignment."); Chen v. Chen Qualified Settlement Fund, 552 F.3d 218, 227-28 (2d Cir.2009) ("Here, the record demonstrates that Judge Korman was critical of the quality of [counsel's] representation in this case. On one occasion, Judge Korman used derogatory language in referring to [counsel's] affidavit.... Given that the quality of [counsel's] work was inextricably intertwined with the court's consideration of the fees application, it was not inappropriate for Judge Korman to express an opinion regarding [counsel's] handling of the case."); Charron v. United States, 200 F.3d at 789 ("The judicial comments and actions upon which the [plaintiffs] rely, however, merely reflect Judge Horn's evaluation and criticism of [plaintiffs' attorney's] handling of the cases and her perception that his professional performance was severely deficient.... The judge's comments and actions, however, do not establish either personal bias and prejudice or the appearance of partiality."); United States v. IBM Corp. (In re IBM Corp.), 618 F.2d 923, 932 (2d Cir. 1980) ("Our reading of the record does disclose that on several occasions Chief Judge Edelstein has expressed his dissatisfaction with counsel for [defendant]. There have been exchanges in the courtroom and in robing room conferences which indicate that [the judge] has, whether justifiably or not, reached the conclusion that he has been `baited' by counsel by their persistence in raising points which he believed had already been determined by previous rulings."); Chevron Corp. v. Donziger, 783 F.Supp.2d at 723 ("Here, there is no allegation of extra judicial source. And the rulings complained of, which in some cases are rather different than the [plaintiffs'] Representatives' distorted and misleading accounts of them, plainly do not fall within the rarest circumstances in which they could evidence the requisite bias or appearance of partiality.... Disagreement or dissatisfaction with the Court's rulings is not enough to succeed on this [recusal] motion. An adversary system inherently has one side that wins and another that loses. If losses compromised the appearance of justice, this system would grind to a halt." (quotations & fns. omitted)); Armatullo v. Taylor, 04 Civ. 5357, 2005 WL 2386093 at *19 (S.D.N.Y. Sept. 28, 2005) ("[W]hile Justice Berkman expressed great impatience with [petitioner], it cannot be said that she exhibited `a deep-seated favoritism or antagonism that would make fair judgment impossible.'").49

For all of these reasons, recusal is not warranted based on my in-court comments.

CONCLUSION

For the reasons discussed above, plaintiffs' recusal motion (Dkt. No. 169) is DENIED as untimely and in any event as without merit.

SO ORDERED.