MEMORANDUM AND ORDER ON DEFENDANTS’ RENEWED MOTION TO RECUSE
SAYLOR, District Judge.This is a criminal prosecution arising out of an allegedly corrupt hiring scheme at the Massachusetts Office of the Commissioner of Probation between 2000 and *282010. Defendants John J. O’Brien, Elizabeth V. Tavares, and William H. Burke, III, all former Probation officials, are charged with conspiracy to commit racketeering, racketeering, mail fraud, conspiracy to commit bribery, and bribery. Essentially, the government contends that defendants engaged in a scheme to defraud involving the process of hiring probation officers, in which individuals who were “sponsored” by state legislators would be hired in return for favorable appropriations and other legislation. The government does not argue that mere patronage hiring is a crime, but rather that defendants engaged in a rigged hiring process involving, among other things, the use of falsified documents and deliberate misrepresentations.
Defendants John J. O’Brien and William H. Burke, III, now joined by defendant Elizabeth V. Tavares, have filed a renewed motion that I recuse myself in this matter. The motion is, in part, a renewal of defendants’ prior motion, which I denied in a written memorandum and order dated February 6, 2014. Familiarity with that memorandum is presumed.
Defendants now contend that they intend to call one of my colleagues, United States District Judge Timothy S. Hillman, as a witness in this case. Defendants’ renewed motion comes on the eve of trial, and is based on a claim of alleged “newly-discovered” evidence: that Judge Hillman is not merely a colleague, but a friend.
For the following reasons, the motion will be granted.
I. Statement of Facts
1. Timothy S. Hillman is a United States District Judge for the District of Massachusetts.
2. Judge Hillman is assigned to the Central Division, and sits principally in Worcester. He has been' a District Judge since June 2012.
3. There are presently 13 judges in the District of Massachusetts, ten of whom are active and three of whom are senior. Judge Hillman and I are two of the ten active judges.
4. Prior to becoming a District Judge, Judge Hillman was a United States Magistrate Judge, also in the Central Division. He became a Magistrate Judge in February 2006.
5. Between February 2006 and June 2012, Judge Hillman and I served together in Worcester. Other than the Bankruptcy Judge, we were the only two judges in the courthouse.
6. In Worcester, our chambers were on the same floor, and we spoke with one another on a near-daily basis. We have worked together as colleagues for eight years. I consider Judge Hillman to be a friend.
7. Prior to becoming a Magistrate Judge, Judge Hillman was an Associate Justice of the Superior Court. Before that, he was a Justice of the state District Court. He served in the state judiciary from 1991 to 2006.
8. This case was indicted on March 22, 2012. It was initially assigned to me as the District Judge and Judge Hillman as the Magistrate Judge.
9. Then-Magistrate Judge Hillman presided over a number of early proceedings in this case; among other things, he appointed current counsel (that is, the Federal Defender’s Office) to represent defendant O’Brien.
10.At the time of the indictment, Judge Hillman had been nominated *29to a seat on the United States District Court. He received his judicial commission on June 6, 2012. On June 18, the case was reassigned to Magistrate Judge Soro-kin.
11. I was reassigned from Worcester to Boston in July 2012.
12. Between August 23, 2012, and November 20, 2012, the government produced various items of written discovery to the defendants.
13. That written discovery indicated, among other things, that Judge Hillman, while a state court judge, either wrote recommendation letters to various individuals for appointment or promotion as probation officers, or permitted his name to be used as a reference for such persons on the standard court form. According to defendants, Judge Hillman did so 22 times.
14. The government also produced to defendants so-called “sponsor lists” of names of candidates for probation officer positions; Judge Hill-man’s name was on four such lists. There is no evidence that Judge Hillman created those lists, or was even aware they existed.
15. Defense counsel were aware, or reasonably should have been aware, no later than the end of 2012 that Judge Hillman was a possible witness.
16. Defense counsel were aware from the beginning of this case that Judge Hillman and I worked together in Worcester, and that there was only one district judge and one magistrate judge at that location.
17. On May 17, 2013, after consultation with the parties, I set this case for trial beginning on January 6, 2014.
18. On July 23, 2013, defendants jointly moved to continue the trial. I granted that motion.
19. On August 2, 2013, again after consultation with the parties, I set a new trial date of February 24, 2014.
20. Between January 9 and 15, 2014, I issued at least five rulings adverse to the defendants, including significant rulings that denied a motion to dismiss and a second motion to continue the trial.
21. On January 16, 2014, defendants O’Brien and Burke moved for my recusal on a variety of grounds. The principal basis of the motion was the alleged connection of my former law partner, Paul Ware, to this case.
22. Defendants’ initial recusal motion did not mention Judge Hillman, or my possible relationship with him, as a basis for recusal.
23. On January 22, 2014, defendants filed a third motion to continue the trial.
24. On January 24, 2014,1 held a hearing on defendants’ motion for recu-sal. In the course of that hearing, and as part of a comparison for purposes of illustration, I commented that Judge Hillman was a friend as well as a colleague.
25. On January 27, 2014, defendants informed me that they intended to call Judge Hillman as a witness, and that they might seek my recu-sal on that ground.
26. On January 29, 2014, I held an in camera conference (the transcript of which was later unsealed) to address defendants’ claim that they intended to call Judge Hillman as a witness.
*3027. In the course of that conference, counsel for defendant Burke alleged that Judge Hillman had “referred a number of candidates to probation jobs, including his own daughter ... that’s a matter of record.” (Jan. 29 Tr. at 17). He later reiterated that Judge Hillman “called up, he referenced] qualified people, including his daughter, for jobs in the probation department.” (Id. at 20-21).
28. The statements concerning Judge Hillman’s daughter were false.
29. Later that day, counsel for defendant O’Brien filed a letter with the Court withdrawing the allegation as to Judge Hillman’s daughter. Counsel for defendant Burke, who had made the allegation, joined in the letter, although he did not write it.
30. On January 31, 2014, I set a briefing and hearing schedule on the newly raised recusal issue.
31. On January 31, 2014, I continued the trial from February 24 to March 24, 2014, in response to certain late discovery disclosures by the government.
32. On February 6, 2014, I denied the original motion for recusal. I noted the existence of a newly raised issue, and indicated that I would supplement, amend, or modify that decision as might be necessary.
33. On February 13, 2014, defendants filed a renewed motion for recusal.
34. The renewed motion for recusal is in part a motion for reconsideration of the earlier decision, and in part a renewed motion based on the possibility that Judge Hillman might be a witness.
35. On February 24, 2014, I held a hearing on the renewed motion for recusal.
36. At the February 24 hearing, I ordered the parties to file a list of all potential witnesses who were present or former state or federal judges, state or federal public officials, state or federal law enforcement officials, or lawyers.
37. The parties made those filings on February 25, 2014.
38. In their filing, defendants listed as potential witnesses 96 present or former state judges, 64 present or former state legislators, 19 present or former law enforcement officials, and 7 lawyers.
39. On February 25, 2014, defendants O’Brien and Burke filed a fourth motion to continue the trial.
II. Analysis
A. Whether the Renewed Motion to Recuse Was Timely
As with the earlier motion, the first question under § 455(a) is whether the motion to recuse was timely filed. See In re United States, 666 F.2d 690, 694 (1st Cir.1981) (holding that “a motion for disqualification [under § 455 must] be timely filed”). Among other things, the timeliness requirement allows courts to “reject what appear to be strategic motions to recuse a judge whose rulings have gone against the party.” In re United States, 441 F.3d 44, 65 (1st Cir.2006); see also In re Abijoe Realty Corp., 943 F.2d 121, 126 (1st Cir.1991) (denying Rule 60(b) motion “[i]n light of [the] calculated withholding of the disqualification claim”).1
*31Accordingly, “in general, a party must raise the recusal issue ‘at the earliest moment after acquiring knowledge of the relevant facts.’ ” In re United States, 441 F.3d at 65 (quoting In re Abijoe Realty Corp., 943 F.2d at 126) (alterations omitted); United States v. Kelly, 519 F.Supp. 1029, 1050 (D.Mass.1981) (“The law is well settled that one seeking the disqualification of the judge must do so at the earliest moment after knowledge of the facts demonstrating the basis for such disqualification.”).
As noted, this case was indicted, and assigned to me, on March 22, 2012. It was assigned to then-Magistrate Judge Hill-man the same day. Counsel became aware no later than the end of 2012 that Judge Hillman was a potential witness in this case. They did not, however, raise the issue until January 2014, on the eve of trial. Indeed, they did not even raise it in their first motion for recusal.
Counsel acknowledge that they were well aware that Judge Hillman and I are colleagues on a relatively small court, and that we were the only two judges in Worcester (save the bankruptcy judge) for a period of more than six years. They contend, however, that they had no way of knowing that we had a friendly relationship until I specifically mentioned that fact during a hearing on January 24, 2014.
' That claim is scarcely credible. More to the point, it is objectively unreasonable. It is hardly necessary to point out that counsel who intends to call a federal judge as a witness in a case in federal court has an obligation to raise that issue at the earliest possible juncture. And that is obviously true whether the witness judge and the trial judge are friends or not.2
In short, counsel waited at least 15 months before raising the issue of Judge Hillman’s involvement, notwithstanding the fact that they knew he was my colleague on a relatively small court. That was plainly untimely. See, e.g., Lunde v. Helms, 29 F.3d 367, 370 (8th Cir.1994) (motion untimely where seven or eight months elapsed between filing of complaint and motion to recuse); Willner v. Univ. of Kansas, 848 F.2d 1023, 1028-29 (10th Cir.1988) (untimely where ten months elapsed); Chitimacha Tribe of Louisiana v. Harry L. Laws Co., Inc., 690 F.2d 1157, 1164 n. 3 (5th Cir.1982) (untimely where more than two years elapsed). And they have given no valid excuse for the delay. See United States v. Whorley, 550 F.3d 326, 339 (4th Cir.2008) (“The facts underlying [litigant’s] motion were available to him as a matter of public record at the time of his arraignment, and he has shown no cause excusing his failure to seek recu-sal before the eve of trial.”).
*32Furthermore, the renewed motion for recusal, like its predecessor, was filed after a series rulings adverse to defendants, and shortly before the scheduled beginning of the trial. At a minimum, that raises serious concerns that the timing of the motion, and whether in fact it was actually based on “newly discovered” evidence. See Newport News Holdings Corp. v. Virtual City Vision, Inc., 650 F.3d 423, 432 (4th Cir.2011) (waiting seven months to file recusal motion “smack[ed] of gamesmanship” where motion was filed after an adverse ruling); Summers, 119 F.3d at 921 (holding that “where the facts are known before a legal proceeding is held, waiting to file [a recusal] motion until the court has ruled against a party is untimely”).
Nonetheless, because of the relatively high stakes involved, I am reluctant to deny the motion solely on the grounds that it is untimely. I will therefore proceed to consider the merits of the motion. See S.E.C. v. Loving Spirit Found. Inc., 392 F.3d 486, 494 (D.C.Cir.2004) (finding recusal motion untimely but also addressing its merits).
B. The Standard for Recusal Under § 455(a)
As with the previous motion, the principal statute under which defendants seek my recusal is 28 U.S.C. § 455(a). That statute provides that a judge “shall disqualify himself in any proceeding in which his impartiality might reasonably be questioned.” 28 U.S.C. § 455(a). The focus of that provision “is not [the judge’s] actual state of mind at a particular time, but the existence of facts that would prompt a reasonable question in the mind of a well-informed person about the judge’s capacity for impartiality.” In re Bulger, 710 F.3d 42, 46 (1st Cir.2013).
Section 455(a) does not impose “the standard of ‘Caesar’s wife,’ the standard of mere suspicion.” In re Allied-Signal, 891 F.2d 967, 970 (1st Cir.1989). Instead,
[t]he recusal standard must be more demanding because the disqualification decision must reflect not only the need to secure public confidence through proceedings that appear impartial, but also the need to prevent parties from too easily obtaining the disqualification of a judge, thereby potentially manipulating the system for strategic reasons, perhaps to obtain a judge more to their liking.
Bulger, 710 F.3d at 47 (quoting Allied-Signal, 891 F.2d at 970) (emphasis in original) (internal quotations omitted).
The test is an objective one, based on the factual circumstances viewed as a whole; it is not based on the subjective state of mind of the judge, or that of the party seeking recusal. United States v. Voccola, 99 F.3d 37, 42 (1st Cir.1996) (noting that “the charge of lack of impartiality” must be “grounded on facts that would create a reasonable doubt concerning the judge’s impartiality, not in the mind of the judge himself or even necessarily in the mind of the litigant filing the motion under 28 U.S.C. § 455, but rather in the mind of the reasonable man” (quoting United States v. Cowden, 545 F.2d 257, 265 (1st Cir.1976))).
Furthermore, the test is not that of a casual observer, but rather a reasonable and objective observer who is “fully informed of all the relevant facts.” In re United States, 158 F.3d 26, 31 (1st Cir.1998); see Bulger, 710 F.3d at 46 (question depends on “the existence of facts that would prompt a reasonable question in the mind of a well-informed person”); In re Boston’s Children First, 244 F.3d 164, 167 (1st Cir.2001) (issue of recusal must be analyzed from the perspective of “an objective, knowledgeable member of the pub-*33lie”) (quoting In re United States, 666 F.2d 690, 695 (1st Cir.1981)); see also Swan v. Barbadoro, 520 F.3d 24, 26 (1st Cir.2008).
C. Whether I Can Consider the Likelihood That Judge Hillman Will Be a Witness
A threshold issue, as with the earlier motion to recuse, is whether I can consider the potential testimony of the claimed witness as part of the recusal analysis. Defendants continue to contend that I cannot look behind their claims to ascertain whether or not those claims have any substance. They have now refined that argument somewhat; specifically, they now contend that if they make a “non-frivolous showing” that a claimed witness is “materially connected to the case” — and, presumably, to me — no further inquiry may be made, and my recusal at that stage must be automatic. (Def. Mem. at 24).
That position, if adopted, would produce a profound change in the law of recusal. If defendants’ theory were correct, the effective power to decide whether recusal was required would shift dramatically away from the court and to the parties. Consider, for example, the recent filing made by defendants identifying certain expected trial witnesses (which, they take pains to say, may be incomplete). Defendants have identified 96 present or former state and federal judges as witnesses — as well as 64 legislators and 19 law enforcement officials. It is hard to imagine that any judge in this district does not have a “material connection” to multiple individuals on that list. It is also highly doubtful that all 179 individuals actually would testify. And it is likewise doubtful that all 179 individuals actually have testimony that is relevant and admissible; at the very least, much of that evidence surely would be repetitive or cumulative. Nonetheless, under defendants’ view, if I have a connection to any one of those people, and defendants make a “non-frivolous” showing that the person might possibly be a witness, recusal is automatically required. Moreover, according to defendants, I must accept their claim at face value; I cannot even examine whether that person actually has anything relevant to offer, or whether that testimony is even admissible. Defendants could thus effectively veto any judge simply by coming up with a “non-frivolous” basis to list someone as a witness, and could keep a judge by omitting the “witness” or waiving the conflict.
That is not, and never has been, the law. In fact, the principal case cited by defendants in support of their position is to the opposite effect. In United States v. Gordon, 354 F.Supp.2d 524 (D.Del.2005), the defendants in a case charging criminal racketeering and fraud alleged that the extrajudicial activities of a federal judge were relevant to their defense, and proposed to subpoena him as a witness. The court did not simply accept the defendants’ claim on faith, but reviewed the alleged evidence in some detail. Id. at 525-26. After remarking that “[i]t is not enough to simply ask what role may be thrust upon the judicial colleague [who may be called as a witness]; one must also ask what that colleague has at stake,” the judge recused himself. Id. at 528. Among other things, the judge noted that the alleged evidence did not simply involve “taking some factual testimony from a disinterested third party,” but rather involved testimony concerning the reputation of the judicial colleague. Id. In short, the court conducted a sufficient degree of inquiry to satisfy itself that the claim was substantial enough to merit consideration, although of course it did not rule definitively on any of the defendants’ claims.
I will adopt a similarly sensible and practical approach. I will therefore exam*34ine the substance of the testimony defendants seek to elicit from Judge Hillman as part of the analysis of whether recusal is appropriate.
D. Whether Recusal Is Required Based on Judge Hillman’s Likely Testimony
Defendants contend that Judge Hillman has “directly relevant, firsthand knowledge” about the allegations in the indictment. (Jan. 29 Tr. at 9). Although defendants’ contentions are somewhat vague, they have identified at least six subjects of proposed testimony.3
First, defendants contend that over the course of several years Judge Hillman “recommended” 22 candidates for positions in the Probation Department. (Def. Mem. at 14). Defendants characterize Judge Hillman as one of the “most frequent sponsor[s] [sic] of candidates for employment at Probation.” (Id.)
Second, defendants contend that Judge Hillman’s name appears on four so-called “sponsor lists” for different candidates for Probation positions. (Feb. 24 Hrg. Ex. 1). As noted, there is no evidence that Judge Hillman prepared those lists, or that he was even aware that they existed.
Third, defendants contend that Judge Hillman would testify that he “saw nothing wrong” with “recommending qualified candidates.” (Def. Mem. at 14).
Fourth, defendants contend that Judge Hillman “sat on some [hiring] interview panels during the time that defendant O’Brien was Commissioner of Probation.” (Def. Mem. at 15).
Fifth, defendants contend that Judge Hillman was aware of, and can describe, hiring practices in the courts prior to the legislative change in 2001 (and prior to the period charged in the indictment). (Id.)
Sixth, defendants contend that Judge Hillman would testify that he was aware that the state judiciary experienced substantial problems implementing a computer system in the 1990s, and that he had been involved in efforts to remedy those problems. (Id.) Defendants further contend that Judge Hillman may opine that “the legislature was angry” about the computer problems and “there was strong pressure to remedy the problem and that the legislature was treating the judiciary unfavorably as a result.” (Id.)
It appears that Judge Hillman’s principal connection, if any, to this case is the fact that he wrote reference letters, or allowed his name to be used as a reference, for various applicants for positions or promotions. According to the government, hundreds, if not thousands, of individuals gave similar references, and it is a perfectly normal practice to do so. The government contends, however, that the hiring process in the Probation Department was rigged (at least where certain key legislators “sponsored” candidates), and therefore references from individuals such as Judge Hillman were essentially ignored.
If Judge Hillman were simply one of a parade of witnesses who had written reference letters for candidates, or allowed their names to be used as references, it seems unlikely that his testimony on that topic would raise any recusal issues. The mere fact that a witness, especially a mi*35nor witness, might be a Mend or colleague of the judge does not normally require recusal. See, e.g., In re Beyond Innovation Technology Co., Ltd., 166 Fed.Appx. 490, 491 (Fed.Cir.2006) (holding that recu-sal was not required where the expert damages witness in a patent case was “a close personal Mend” of the judge); United States v. Kehlbeck, 766 F.Supp. 707, 712 (S.D.Ind.1990) (noting that “judges may have Mends without having to recuse themselves from every case in which a Mend appears as counsel, party, or witness.”); M.K. Metals, Inc. v. Nat’l Steel Carp., 593 F.Supp. 991, 996 (N.D.Ill.1984) (noting that “Mendship between judges and trial actors other than parties, untainted by any financial interests, should not generally justify recusal”); In re Jim Slemons Hawaii, Inc., 2013 WL 980115 at *12 (9th Cir. BAP Mar. 13, 2013) (noting that “[sjocial acquaintances, Mendships or associational relationships are rarely grounds for recusal”); United States v. Nelson, 2010 WL 2629742 at *7 (E.D.N.Y. June 28, 2010) (noting that “[m]any opinions have discussed the impact of a judge’s Mendship’ with a litigant or ‘interested party’ on the decision to disqualify under § 455(a). In few cases, however, has disqualification been deemed required.”).
Defendants, however, seek to go further. They have argued that Judge Hillman did not simply provide a reference, but “sponsored” candidates; that as a judge he was a powerful figure, like the leaders of the state legislature; and that at least some of his “recommendations” were “followed.” The characterization by defendants of Judge Hillman as a “sponsor” of candidates is by no means accidental. In context, it is a statement fraught with implications. Defendants are clearly seeking to draw explicit parallels between what he did and what they are alleged to have done (and which the indictment charges constitutes a crime).4 The alleged import of this, apparently, is that defendants’ behavior cannot have been criminal because judges were engaged in similar behavior.5 Defendants also seek to elicit Judge Hill-man’s own subjective views that he “saw nothing wrong” with making references— again, presumably, to suggest that defendants likewise did not have the necessary mens rea to commit the crimes charged.6
It is also important to note that defendants apparently intend to make other arguments that may implicate the testimony of Judge Hillman. The contours of their arguments are not always clear, and it is somewhat difficult at this stage to ascertain what actual evidence they will seek to *36elicit. Nonetheless, defendants apparently intend to attack certain individual state judges, and possibly the state judiciary generally, as biased, hypocritical, incompetent, untruthful, and/or vengeful. Defendants apparently intend to try to show, or argue, that the state court did not hire probation officers on the basis of merit; that the hiring process was instead driven by political considerations; that state judges engaged in essentially the same practices as defendants; and that the pending indictment is simply the culmination of a long-standing power struggle between the judiciary and the legislature to control the hiring of probation officers. Some of those arguments, at a minimum, border on improper claims of selective prosecution (“everyone else did the same thing — why are we being prosecuted?”) or jury nullification (“who cares if we did it— it’s just politics”). Implicit throughout is the strong possibility that defendants may try to use the testimony of Judge Hill-man — a federal judge who sits in the very court in which the case will be tried — to try to establish, or at least hint at, some of those points.
Whether any of this has any basis in fact, whether defendants will succeed in getting any of it in evidence, or whether defendants will be permitted to make any of these arguments at trial, is not the key issue at this stage. Rather, it is the fact that the trial judge — whoever that judge may be — will have to sort through those issues, and limit counsel to evidence and arguments that are legally valid. See Gordon, 354 F.Supp.2d at 528 (“Some judicial officer must determine how far to permit the defense to develop evidence about [the involvement of the judge’s colleague] ... and how much of any such evidence can properly be presented at trial.”).
Again, Judge Hillman is both a current colleague and a friend. I have high confidence in his integrity and character, and defendants have proffered nothing to suggest that he did anything improper or illegal. There is certainly nothing wrong with serving as a reference for a qualified applicant. But it seems clear that I would have to rule upon the validity of entire lines of questioning, and that in all likelihood I would limit the testimony in multiple respects. That, in turn, would open me up to the criticism that I was excluding evidence not for proper reasons, but to shield my colleague from embarrassment. That issue at the very least raises a serious possibility that recusal may be appropriate.
E. Whether Other Factors Weigh in Favor of Recusal
Unfortunately, the potential issues do not stop with the testimony of Judge Hill-man. To that must be added the possibility of yet another round of claims by the defense that my recusal is required. As noted, when I required the parties to make a filing identifying all present and former judges, public officials, and lawyers they expected to call as witnesses in this case, defendants produced a list of 179 names. Given that I have been a member of the bar or a public official in Massachusetts for 33 years, it is not surprising that the list included a number of people with whom I have some degree of connection. Among other things, the list includes former colleagues at Goodwin Procter, former colleagues at the U.S. Attorney’s Office, spouses of former colleagues, professional acquaintances, and at least one former client. While my connections with those newly identified individuals are generally tenuous, and the relevance and importance of their testimony is unclear, it is reasonably likely that defendants will contend that those individuals, too, are material witnesses.
*37Defendants have also raised the possibility that United States District Judges Michael Ponsor and Nathaniel Gorton, who are also colleagues and personal friends, may be called to testify. Both apparently gave a single reference on behalf of a single probation applicant. (Def. Mem. at 26 n. 6). While defendants contend that “[g]iven the present state of the evidence,” id., they do not intend to call them as witnesses, presumably defendants are reserving the right to do so should they so choose.
Finally, the issues previously raised by defendants, to the extent that they are not entirely specious, must also be added to the mix. Taken as a whole, those issues add at least some marginal weight to the balance.
F. Whether Recusal Is Appropriate
The question, then, is whether my impartiality might reasonably be questioned if I were to preside over the trial of this case. Again, the test is that of a reasonable and objective observer who is fully informed of all the relevant facts. And I must also be cognizant of “the need to prevent parties from too easily obtaining the disqualification of a judge, thereby potentially manipulating the system for strategic reasons.” See Bulger, 710 F.3d at 47 (quoting Allied-Signal, 891 F.2d at 970).
The recusal question here involves a difficult balancing act. There is certainly a realistic probability, because of my relationship with Judge Hillman, that my impartiality might reasonably be questioned. It will be easy to make the accusation that I am trying to protect Judge Hillman, and difficult to dispel the doubts. Nonetheless, I am very much cognizant of the costs such a recusal will impose. Judicial recusal always imposes some degree of cost; here, the price of recusal will be very steep indeed.
One cost will be disruption and delay. Defendants’ recusal motions could hardly have been more ill-timed; the case was indicted nearly two years ago, and is now less than three weeks away from the start of trial. The government, not just the defendants, has an interest in a speedy trial, and for that matter so does the public. No matter where this case is reassigned, delays — possibly significant ones— are likely to follow.
Another cost is that defendants will inevitably reap the full reward of an untimely tactical maneuver. Whether they deliberately waited to raise the issue, or held back out of inattention or negligence, is unimportant; the result will be the same.
Finally, there is a real danger that recu-sal would set a new and pernicious precedent, emboldening counsel to file untimely recusal motions or otherwise seek to manipulate circumstances to their own advantage. Recusal could thus undermine the very goal of public confidence in the fair and impartial administration of justice that I am seeking to uphold.
There is no completely satisfactory answer to this problem, and whatever choice I make may lead to at least some unfortunate consequences. Under the circumstances, however, I have concluded that the paramount consideration is maintaining public confidence in the fairness and integrity of this proceeding. That consideration outweighs the other factors, however strong those factors may be. Therefore, and with considerable misgivings, I feel duty-bound to recuse myself.
The resulting and inevitable delay in the trial of this matter is unfortunate. But the trial will begin at some point; that day will not be postponed forever. As to the possible precedent, it is perhaps enough to say that each case must rise and fall on its own facts, and that I have great confidence in *38the wisdom and good judgment of my colleagues in making appropriate decisions in future circumstances. In any event, I am persuaded that this step is the correct one to make.
III. Conclusion
For the foregoing reasons, defendants’ renewed motion for recusal is GRANTED.
So Ordered.
. "The policy considerations supporting a timeliness requirement ... [are] to conserve judicial resources and prevent a litigant from waiting until an adverse decision has been *31handed down before moving to disqualify the judge.” Summers v. Singletary, 119 F.3d 917, 921 (11th Cir.1997). Thus, it serves to prevent litigants from taking a "heads-I-win-tails-you-lose position of waiting to see whether they win and if they lose moving to disqualify a judge who voted against them.” Schurz Communications, Inc. v. Federal Communications Comm’n, 982 F.2d 1057, 1060 (7th Cir.1992).
. It is commonplace for judges on the same court to meet with one another, formally and informally, for a variety of reasons. It is likewise commonplace for those judges to socialize on occasion, at court and bar events and privately. A reasonable observer would conclude that judges who "see each other regularly” over time "develop!] professional respect, appreciation, and friendship for one another.” United States v. Gordon, 354 F.Supp.2d 524, 528 (D.Del.2005). While counsel might not be reasonably expected to know the precise parameters of the relationships among the different judges, surely the basic outline would be obvious to any lawyer.
. The truthfulness of defendants' claims as to the substance of Judge Hillman’s testimony is open to some question. At a hearing on January 29, counsel for defendant Burke claimed that Judge Hillman had "referred” his own daughter for a probation job. That statement, as it turns out, was entirely false. Although defense counsel later retracted the claim, the episode does not inspire confidence in counsel’s representations as to the evidence.
. Defendants acknowledge that they are not accusing Judge Hillman of having committed a crime, or having done anything improper at all.
. The government has a different view:
... [T]he defendants have not explained why Judge Hillman is differently situated from the hundreds if not thousands of individuals who appeared as references on applications for probation officers or wrote letters of recommendations. They have not established that a recommendation by Judge Hillman caused the defendants to engage in fraudulent activity to hire such an individual. Therefore, the only logical conclusion to draw is that the defendants have chosen Judge Hillman as a witness not because he has any significant testimony, but as a strategy to force the Court’s recusal and continue the trial....
(Gov’t. Mem. at 3).
.Defendants also seek to elicit testimony about Judge Hillman’s knowledge of certain events prior to 2001, and his opinion that the legislature was "angry” at the judiciary. Some of that evidence appears to be plainly inadmissible — it is unclear how a witness would be permitted to give an opinion as to the purported state of mind of a state legislature — and the relevance of the rest is uncertain.