This case arises out of an award of attorneys’ fees imposed as a sanction on attorneys who brought a voting-rights lawsuit against the Mayor of Albuquerque. After dismissing the case, the district court found the attorneys unreasonably multiplied proceedings in what it called a meritless case and sanctioned them under 28 U.S.C. § 1927. They argue the award was an abuse of discretion. The Mayor cross-appealed, arguing the court abused its discretion by declining to award fees under several other provisions the Mayor raised as grounds for sanctions.
Although most of the attorneys’ arguments lack merit, we vacate the award of fees and remand for the court to consider whether a different trigger for the imposition of sanctions is appropriate. The Mayor dropped his cross-appeal at oral argument. Consequently, exercising jurisdiction under 28 U.S.C. § 1291, we VACATE the award of fees and REMAND for further proceedings consistent with this opinion. We also DENY the motion for sanctions on appeal.
I. Background
Several Albuquerque residents sued Mayor Richard Berry in his official capacity as Mayor of Albuquerque in state court over the City’s redistricting plan enacted after the 2010 census.1 They sought “in-junctive and declaratory relief to achieve a constitutionally acceptable and otherwise lawful redistricting of the Albuquerque City Council.” App. 185. Specifically, they claimed that (1) a newly adopted Albuquerque redistricting map “minimize[d] the opportunities of Latinos to participate in the political process and to elect the representatives of their choice,” in violation of Section Two of the Voting Rights Act of 1965; and (2) the new districts “deviate[d] impermissibly from population equality,” in violation of the Equal Protection Clauses of both the United States and New Mexico Constitutions. Id. at 12, 14. The Mayor removed the case to federal court in January 2013.
After the lawsuit had been commenced, in March 2013, a city-charter amendment passed mandating that no candidate could be elected without receiving a majority of the vote. That abrogated the previous rule, under which the top vote recipient with a plurality of 40% or more would prevail.2 Subsequently, in June 2013, the Mayor provided the voters with a critical expert report. That report, created by Brian Sanderoff, purported to identify flaws in the voters’ theory of the case in general and, more importantly, as described in the reports prepared by the *1267voters’ experts. We discuss those alleged flaws as they become relevant below.
Several weeks later the voters filed a motion for voluntary dismissal without prejudice of their claims. They explained that they did so “to assure that [the lawsuit] would not interfere with the upcoming [Fall] elections in the City of Albuquerque and to ascertain whether the change in the political landscape in the City of Albuquerque [ie., the city-charter amendment] would necessitate further litigation.” Id at 46. The Mayor opposed this motion, arguing that dismissing the case without prejudice would result in legal prejudice to him. Thus, he filed a motion asking the court to instead dismiss the case with prejudice.
On September 30, rather than granting or denying either motion, the court entered a stay in the case and “deferred] ruling until after the upcoming [November] mayoral election.” Id at 190. The court found the facts before it insufficiently clear to justify ruling definitively on either motion at that time. In its order, the court noted that, after “the conclusion of the mayor’s race, the Court will hold a status conference and the parties shall advise the Court how they wish to proceed.” Id at 191. And it instructed the voters to “be prepared to advise the Court whether, given the results of the mayoral election, they still wish to pursue litigation on the allegations raised in the complaint.” Id
Neither party filed anything further from that point forward, and the mayoral election came and went. On November 12, the court held a telephone conference, at which the voters “advised ... that there was also an upcoming election for councilperson, and suggested continuing the stay,” again suggesting that after this election the “issue [might] become moot.” Id at 192, 195. The court understood that as a request to stay the decision until it became clear whether or not “further legal action was necessary, based on the implementation of the [city-charter amendment].” Id at 196. Nothing happened for two months aside from the court vacating a December 17 telephone conference because of its scheduling issues. The record reflects no action by the voters to advise the court of the effect the election had on their claims or whether they wished to proceed. Accordingly, the dueling motions remained pending.
In January 2014, the court revisited the motions. It noted that, although its stay had been based on the voters’ “representations that the outcome of the elections would determine whether the underlying issues had become moot,” they apparently still could not “make a decision about whether they have a meritorious lawsuit or not.” Id The court found that the voters’ failure to take any affirmative action post-election indicated their claims that the results of the Fall elections would let them “determine whether there remained an issue to litigate” had been “disingenuous.” Id In short, the court found their reasons for seeking dismissal without prejudice insufficient. Finding it “apparent that there [was] no longer a case to pursue,” the court denied the voters’ motion to dismiss without prejudice and dismissed the case with prejudice. Id at 195-96.
The Mayor subsequently moved for an award of attorneys’ fees and costs under a host of provisions. After holding a sanctions hearing, the court ruled on the motion in August 2014. The court only awarded sanctions under 28 U.S.C. § 1927, which allows the imposition of fees on lawyers who “unreasonably and vexatiously” multiply proceedings. Although the court found the lawsuit was not filed in bad faith, it found that “at some point during the course of the litigation,” counsel’s conduct “in maintaining [the] case multiplied] the proceedings in an unreasonable and vexa*1268tious manner.” Id. at 405. The court con-eluded that “the magic date that this case was no longer viable and ... counsel unreasonably continued this matter was June 25, 2013, the date [counsel] was provided with [Mr. Sanderoff s] expert report.” Id. at 407. According to the court, “[u]pon reading that report, it would have been clear to a reasonable attorney that this case no longer had merit.” Id. Thus, the court imposed an award of attorneys’ fees, beginning from June 25, 2013, amounting to $48,217.95.
The voters have appealed only that order granting attorneys’ fees.
II. Analysis
Federal law provides that any attorney “who so multiplies the proceedings in any case unreasonably and vexatiously may be required by the court to satisfy personally the excess costs, expenses, and attorney’s fees reasonably incurred because of such conduct.” 28 U.S.C. § 1927. This is an “extreme standard,” and fees should be awarded “only in instances evidencing a serious and standard disregard for the orderly process of justice.” AeroTech, Inc. v. Estes, 110 F.3d 1523, 1528 (10th Cir.1997) (internal quotation marks omitted). Thus, courts must “strictly construe[]” the statute to guard against “dampen[ing] the legitimate zeal of an attorney in representing his client.” Braley v. Campbell, 832 F.2d 1504, 1512 (10th Cir.1987) (en banc).
Courts need not find that an attorney subjectively acted in bad faith. Rather, “any conduct that, viewed objectively, manifests either intentional or reckless disregard of the attorney’s duties to the court[ ] is sanctionable.” Hamilton v. Boise Cascade Express, 519 F.3d 1197, 1202 (10th Cir.2008) (internal quotation marks omitted). The statute makes attorneys potentially liable for harm caused “because of’ unreasonable and vexatious multiplication of proceedings. 28 U.S.C. § 1927. Thus, “there must be a causal connection between the objectionable conduct of counsel and multiplication of the proceedings,” such that the conduct “resulted] in proceedings that would not have been conducted otherwise.” Peterson v. BMI Refractories, 124 F.3d 1386, 1396 (11th Cir.1997); see also Lee v. First Lenders Ins. Servs., Inc., 236 F.3d 443, 445 (8th Cir.2001).
Although we generally review an award of fees under § 1927 for an abuse of discretion, if “the exercise of that discretion depended on the resolution of a purely legal issue,” we review that issue de novo. Hamilton, 519 F.3d at 1202 (emphasis added). We emphasize “depended” because many of the voters’ assertions of legal error on appeal attack legal analysis upon which the imposition of sanctions did not obviously depend, contained in orders not designated in their notice of appeal. Cf. Navani v. Shahani, 496 F.3d 1121, 1133 (10th Cir.2007) (stating we have jurisdiction only over orders appellants designate in their notice of appeal). We address those instances as they become relevant. For now, we only note that barring any actually relevant legal error, a court’s discretion to award fees is broad if it concludes an attorney acted in an objectively unreasonable way that multiplied proceedings.3
*1269The voters’ arguments for reversal fall into two general categories: (1) the court committed legal error in imposing the sanctions, which would be a per se abuse of discretion; and (2) on these facts, imposing § 1927 sanctions is an abuse of discretion. We begin by asking whether the sanctions order “rest[ed] on an erroneous view of the law.” Roth v. Green, 466 F.3d 1179, 1187 (10th Cir.2006). But we stress again that only legal errors upon which the imposition of sanctions actually depended are reversible errors. An otherwise free-floating legal error — e.g., one occurring in an entirely separate decision by the court— cannot support reversal unless it necessarily resulted in the sanctions award.
A. Challenge to Process
The voters initially raise several complaints about the process the court followed in imposing sanctions. The first challenge stems from their July 5, 2013 motion under Federal Rule of Civil Procedure 41(a)(2) that the court dismiss the case without prejudice. As described above, when the Mayor opposed the motion and filed his own motion requesting dismissal with prejudice, the court deferred ruling on either motion in an order staying proceedings until after the upcoming elections.
Although denials of requests under Rule 41(a)(2) to dismiss without prejudice generally receive abuse-of-discretion review, “[ajbsent ‘legal prejudice’ to the defendant, the district court normally should grant such a dismissal.” Ohlander v. Larson, 114 F.3d 1531, 1537 (10th Cir.1997). And, because a “court abuses its discretion when denying a motion to dismiss under Rule 41(a)(2) based on its inconvenience,” the “court’s time or effort spent on the case” is not a proper consideration. See id. The voters complain the court’s order staying the case (1) identified no legal prejudice to the Mayor, and (2) was based solely on its time or effort spent on the case. According to them, this violates Ohlander and amounts to a legal error sufficient to break the causal connection between any objectionable conduct of counsel and multiplication of proceedings. But, for several independent reasons, we disagree.
As an initial matter, denying a motion to dismiss is different from deferring decision on that motion by means of a stay. Ohlander concerned legal error arising when improper considerations enter into a decision to deny a motion to dismiss without prejudice. See id. (explaining we were considering a “district court’s decision to deny a voluntary dismissal”). Here, the voters assert error in the court’s stay order, not in its later order denying their motion and dismissing the case.4 While Ohlander may place some constraints on a court’s ability to deny a Rule 41(a)(2) motion, it says nothing about the district court’s “broad discretion to stay proceedings as an incident to its power to control its own docket.” Clinton v. Jones, 520 U.S. 681, 706, 117 S.Ct. 1636, 137 L.Ed.2d 945 (1997). And “[i]t is well settled that the district court has the power to stay proceedings pending before it and to control its docket for the purpose of economy of time and effort for itself, for *1270counsel, and for litigants.” Pet Milk Co. v. Ritter, 323 F.2d 586, 588 (10th Cir.1963) (internal quotation marks omitted). Thus, the stay order comports with our precedent. And the voters identify no reason the stay, as a factual matter, fell outside the court’s broad discretion to control its docket.
Moreover, even if Ohlander’s rule reached .as far as the voters think, the court did not elevate its convenience over potential legal prejudice to the defendant in staying the case. Proper considerations in the legal-prejudice inquiry include “the opposing party’s effort and expense in preparing for trial; excessive delay and lack of diligence on the part of the movant; insufficient explanation of the need for a dismissal; and the present stage of litigation.” Brown v. Baeke, 413 F.3d 1121, 1124 (10th Cir.2005). But these “factors are neither exhaustive nor conclusive” and courts “should be sensitive to other considerations unique to the circumstances of each case” in determining legal prejudice, including the equities facing both parties. Id.
The voters claim the court “considered and rejected the city’s assertions of legal prejudice,” 1st Cx-App. Br. at 22, implying the court concluded that dismissing the case without prejudice would not prejudice the Mayor. Under the voters’ view of Ohlander, that would have meant the court had no discretion to do anything but dismiss the case. See id. at 23 (asserting the court “answered” the question of “whether [the] case should be dismissed with or without prejudice” against the Mayor). Thus, they conclude, the court’s “sole reason for staying the case rather than granting the motion to dismiss without prejudice was the court’s time or effort spent on the case,” i.e., its convenience. Id. at 22-23 (internal quotation marks omitted). But this argument reads too much into the court’s order.
What the court actually said was that “[a]t [that] point, the record [was] insufficient to warrant dismissal with prejudice” and that staying the case until after the “upcoming mayoral election” would “preju-diee[ ] neither” party and benefit judicial economy. App. 190. This must be understood in the context of the reasons the voters gave for seeking dismissal without prejudice. The asserted reason for seeking a dismissal was “to assure that [the lawsuit] would not interfere with the upcoming elections ... and to ascertain whether” further litigation would be necessary after the city-charter amendment. Id. at 46. The Mayor objected to that request, arguing the legal-prejudice factors counseled against allowing dismissal without prejudice. The sufficiency of the reasons for seeking dismissal was key to the legal-prejudice inquiry and, consequently, key to the resolution of the moJ tion.
Complicating the inquiry, however, was that the voters’ reason for seeking dismissal without prejudice turned on the uncertain future effect of the city-charter amendment on the upcoming elections. The court apparently thought the voters’ explanation sufficiently unpersuasive to make it prudent to wait until the upcoming election actually happened before reaching a final decision.5 See App. 190 (reasoning *1271that if the city-charter amendment indeed “render[ed] further legal action unnecessary,” as the voters conjectured, they “would have no need to continue to prosecute [the] case,” but that if it did not, they would “be free to pursue the litigation without having to re-file the case”).
Thus, we do not read the order as concluding the Mayor would suffer no legal prejudice from a dismissal without prejudice; we read it as expressing uncertainty about whether legal prejudice would attach and staying the case accordingly. For those reasons, the court’s reasoning cannot be fairly characterized as resting solely on its convenience. In fact, the court explicitly enumerated the legal-prejudice factors at the outset of the order and concluded ultimately that its decision to stay the case prejudiced neither party. Far from resting solely on convenience, the court’s decision indicates full awareness of the need to consider legal prejudice. Thus, even if the voters were right that Oklander’s rule restricts the stay-granting power of district courts, the court fully complied with its dictates.
Finally, even assuming the correctness of the two premises just rejected — (1) that Ohlander prevents courts from staying a decision on a motion to dismiss without prejudice without first finding prejudice to the non-movant, and (2) that the court failed to base its decision on potential legal prejudice to the defendant — the voters would still face an insurmountable obstacle. They would have only shown a legal error in the September 2013 stay order. It would remain to be shown that imposing sanctions in August 2014 depended on that incorrect legal conclusion. We do not see how it could have. Perhaps it would have if this supposed error necessarily enabled or led to the conduct later sanctioned as multiplication of proceedings. But that does not follow. As an initial matter, the conduct the court sanctioned as impermissible multiplication of proceedings began in June 2013. It is difficult to see how an error in a September 2013 stay could have caused that conduct; by the time the court entered the stay, the proceedings had already multiplied. Nor can they show — to the extent this is their claim — that absent the legal error the court would have necessarily granted their motion to dismiss without prejudice. The court certainly could have still denied the motion to dismiss even had it applied the law as the voters perceive it.
It is one thing to identify a legal error in a decision one actually appeals, since it makes some sense to assume the decision rested in part on that error. It is another to argue a court order should be reversed based on a legal error in a separate order issued at an entirely different stage of proceedings. The voters fail to show how this error — assuming it occurred — would even be relevant to the issue on appeal. A far tighter connection must be demonstrated. As far as we can tell, their challenge boils down to a backdoor attack on a decision they have not appealed with law that does not apply. Accordingly, the challenge to the court’s order staying the case cannot resolve this appeal.
Two other complaints about the court’s process can be readily dismissed. The voters allege the court “invite[d] [their] actions by statements in its own orders” and thus could not sanction them “for following the court-approved path.” 1st Cx-App. Br. at 24. This gets the chronology backwards. The court found the sanction-able conduct began in June 2013. Statements made in the stay order could not *1272have “invited” sanctionable conduct months earlier. Moreover, the voters cite no Tenth Circuit law in support of their claim, and the out-of-circuit cases cited6 are unpersuasive. Those cases involved denials of motions for summary judgment, which were in essence determinations that the cases contained meritorious, triable issues. Nothing similar happened here.
The voters next claim that the district court committed reversible error in not allowing plaintiffs to “proceed with only the state claim in state court,” id. at 26. That argument also fails. As background, the voters offered, for the first time, at the November telephone conference to dismiss 7 the federal claims and proceed only with state law claims in state court. The Mayor did not acquiesce. In asserting error, the voters rely on an out-of-circuit case stating that in these cases the “mere prospect of the transfer of litigation to state court [is] an insufficient basis for denying [a] motion for voluntary dismissal.” Davis v. USX Corp., 819 F.2d 1270, 1275 (4th Cir.1987). They provide no Tenth Circuit case for this proposition, and because the legal-prejudice factors are “neither exhaustive nor conclusive,” Brown, 413 F.3d at 1124, it is far from certain we would adopt such a bright-line rule. But the insurmountable causal problems the voters would face even if we did mean we need not decide that question.
The Davis rule would only apply to the January 2014 order actually denying plaintiffs’ motion for voluntary dismissal. But nothing in that order indicates the prospect of proceeding in state court influenced the denial.8 And, again,, the voters decided against appealing that order, choosing instead to appeal only the September 2014 order imposing sanctions. As mentioned, errors in separate, not-appealed orders are only pertinent to the degree they necessarily resulted in errors in the order appealed.
We doubt this would be such an error, given that the sanctions rested on the finding that sanctionable conduct began long before January 2014. Perhaps recognizing that causal problem, the voters in places appear to claim the court committed reversible error in the sanctions order by rejecting an argument that offering to proceed in state court should have precluded sanctions. But we see no reason such an offer would bar a court from later concluding that a party improperly multiplied proceedings, and the voters cite no authority to that effect.9 The sanctions order only mentioned the offer in dismissing the voters’ attempt to shift blame for prolonging the case to the Mayor’s rejection of that offer. See App. 410 (“The Court finds no fault with Defendant’s refusal to [return to state court] because [it] would involve the continuation, or possible continuation, of a meritless case.”); id. at 493 (court noting the Mayor “exercised a right [he has] to remove to federal court” and lack of any motion to remand). The voters do not dispute that the Mayor had the right to *1273remove the case or that they never moved to remand. Nor do they claim the Mayor was required to acquiesce in their offer to return to state court. Thus, their argument that this case could have proceeded in state court is beside the point. Of course the state courts could have adjudicated the claim, but so could the federal courts, and this claim was validly in federal court. We detect no error.
Rounding out their process challenges, the voters claim the court improperly ap-. plied a subjective standard by commenting on their “subjective knowledge regarding the merits of [their] case.” Id. at 410. This argument, however, rests on a legal misunderstanding. To be sure, they point to Braley and that case explained that § 1927 allows sanctions “against an attorney personally for conduct that, viewed objectively, manifests either intentional or reckless disregard of the attorney’s duties to the court.” Braley, 832 F.2d at 1512. But that does not mean an attorney’s subjective bad faith is irrelevant, let alone that commenting on apparent bad faith is reversible error. To the contrary, an “attorney’s actions are considered vexatious and unreasonable under § 1927 if the attorney acted in bad faith.” Dreiling v. Peugeot Motors of Am., Inc., 768 F.2d 1159, 1165 (10th Cir.1985); see also Braley, 832 F.2d at 1512 (noting parenthetically that attorneys are “accountable under § 1927 not only for subjective bad faith conduct but also for ‘reckless indifference to the merits of a claim’ ” (emphasis added)); cf. Steinert v. Winn Grp., Inc., 440 F.3d 1214, 1221 (10th Cir.2006) (listing, as an example of sanctionable conduct under § 1927, instances when attorneys are “cavalier or bent on misleading the court”).10
The court committed no legal error in commenting on plaintiffs’ subjective knowledge regarding the merits of their case.11
B. Challenge to the Merits
The voters also contend the court rested its finding that their case ceased to be meritorious after June 25 on a legally erroneous interpretation of their Voting Rights Act and one-person-one-vote claims. We disagree.
We begin with the Voting Rights Act claim. The Supreme Court, in Thornburg v. Gingles, 478 U.S. 30, 50-51, 106 S.Ct. 2752, 92 L.Ed.2d 25 (1986), established a three-part prima facie framework for a vote-dilution claim under Section Two of the Act. First, plaintiffs must prove their minority group is “sufficiently large and geographically compact to constitute a majority in a single-member district.” Sanchez v. State of Colo., 97 F.3d 1303, 1310 (10th Cir.1996). Second, they must *1274show “the minority group is politically cohesive.” Id. (internal quotation marks omitted). Third, they must show “the white majority votes sufficiently as a bloc to enable it&emdash;in the absence of special circumstances, such as the minority candidate running unopposed&emdash;usually to defeat the minority’s preferred candidate.” Id. These factors are necessary, but not sufficient, to establish a vote-dilution claim.
In selecting June 25 as the “magic date,” the district court concluded that the voters’ experts never established the Gingles factors, and that Mr. Sanderoffs report simply confirmed that failure. The Mayor does not dispute that the voters’ expert Dr. Lonna Rae Atkeson purported to identify racially polarized voting. See App. 128 (Dr. Atkeson asserting that “the evidence is clear that racially polarized voting is present,” which she concluded was “indicative of the need for majority-minority districts within the city to provide minorities an equal opportunity to elect candidates of their choice”). The first Gingles prong is not contested; the question is whether Dr. Atkeson’s testimony sufficed to establish the last two prongs.
The voters correctly emphasize that the second and third Gingles factors can be addressed “conjunctively].” Sanchez, 97 F.3d at 1315. But that does not tell us whether Dr. Atkeson actually addressed those factors. The voters breeze over this point, assuming that asserting the existence of racially polarized voting is sufficient shorthand for asserting the final two Gingles factors. That assumption is incorrect. Dr. Atkeson made two statements: (1) an assertion that racially polarized voting exists; and (2) an “assertion” that was really plaintiffs’ desired conclusion in disguise&emdash;i.e., that Latinos were being deprived of an equal opportunity to elect candidates of their choice. This does not satisfy Gingles, because it ignores entirely whether the white majority was actually voting as a bloc to defeat the minority’s preferred candidate. That omission is perhaps unsurprising in light of Mr. San-deroffs finding that in every election selected by Dr. Atkeson “in which ... Hispanic voters had a preferred candidate,” it turns out “the preferred candidate of the Hispanic population won the election.” App. 70; see also id. at 470 (testimony at sanctions hearing reiterating this deficiency in Dr. Atkeson’s report).
Put another way, the district court only erred if a general assertion that racially polarized voting exists suffices to satisfy Gingles’s third prong. The voters provide no cases for that proposition, and we doubt they could. Consider a case where racially polarized voting exists, but a minority is nevertheless electing candidates of its choice. In that case, the requirement that the white majority votes as a bloc to defeat the minority’s preferred candidate would be unsatisfied. While racially polarized voting is necessary to satisfy the third prong, it is not sufficient. We indicated as much in Sanchez, where, albeit in our discussion of the first prong, we noted that “part of the Gingles threshold inquiry” is whether the district court “can fashion a permissible remedy in the particular context of the challenged system.” Sanchez, 97 F.3d at 1311. Naturally, there is no remedy to fashion if a minority group is not actually prevented from electing candidates of choice.
Contrary to the voters’ framing of the issue, the problem was not that their experts collapsed the second and third Gin-gles prongs. The problem was that their experts entirely failed to address whether the white majority was actually voting as a bloc to defeat minority-preferred candidates. It was not error to find this was a fatal flaw in the experts’ analysis. And it was not error to treat Mr. Sanderoffs report as the final straw, because the report revealed why the flaw was there&emdash; *1275minority-preferred candidates had won every one of the plaintiffs’ exemplar races.
In sum, it was insufficient for Dr. Atke-son to simply nod to the desired conclusion by claiming racially polarized voting showed Latinos needed the ability to elect candidates of their choice without asserting the existence of a necessary premise: that the white majority was actually voting as a bloc to defeat the minority’s preferred candidates. Because the voters never even attempted to assert that necessary premise, there was no Gmyfes-related legal error.12
We turn next to the voters’ assertion that the district court erred in finding their one-person-one-vote claim lacked merit. At the outset, both parties treat a district court case summarily affirmed by the Supreme Court, Larios v. Cox, 300 F.Supp.2d 1320 (N.D.Ga.2004), summarily aff'd, 542 U.S. 947, 124 S.Ct. 2806, 159 L.Ed.2d 831 (2004), as if it were binding law. Of course, summary affirmances have limited precedential value. See Plowman v. Massad, 61 F.3d 796, 799 n. 1 (10th Cir.1995) (noting that “the precedential effect of a summary affirmance extends no further than the precise issues presented and necessarily decided by those actions” (internal quotation marks omitted) (quoting Anderson v. Celebrezze, 460 U.S. 780, 784 n. 5, 103 S.Ct. 1564, 75 L.Ed.2d 547 (1983)). Thus, the particularities of Lar-ios’s holding that certain Georgia reapportionment plans violated one-person-one-vote principles, see Larios, 300 F.Supp.2d at 1357-58, have little to say to our inquiry in this case. The question is not whether that district court’s reasoning could support a one-person-one-vote claim here, but whether the court erred in determining such a claim was untenable under governing precedent.
It is difficult to discern the contours of the voters’ position under relevant law, particularly when they focus single-mind-edly on the “merit [of their claims] under Larios,” 1st Cx-App. Br. at 40. Read most charitably, they appear to be claiming they made a meritorious one-person-one-vote claim attacking the constitutionality of voter-population deviations of plus or minus 5% in Albuquerque’s plan.13 But their opening brief identifies no Tenth Circuit or Supreme Court cases showing the district court necessarily erred in finding their claim, so construed, lacked merit.14 In *1276fact, the extent of their analysis is that the constitutionality of such deviations was a question “left open by Larios.” 1st Cx-App. Br. at 41. Again, it is irrelevant whether the Northern District of Georgia opined on that point. The question is whether such an attack could be or was grounded in law we must follow.
The law with which the voters must grapple to show their claim was col-orable is well established. An “apportionment plan with a maximum population deviation [from ideal district size] under 10% falls within” the “category of minor deviations” that are “insufficient to make out a prima facie case of invidious discrimination under the Fourteenth Amendment so as to require justification by the State.” Voinovich v. Quitter, 507 U.S. 146, 161, 118 S.Ct. 1149, 122 L.Ed.2d 500 (1998); see also White v. Regester, 412 U.S. 755, 764, 98 S.Ct. 2332, 37 L.Ed.2d 314 (1973) (observing “relatively minor population deviations” under 10% and noting plaintiffs were thus unable to establish a violation “from population variations alone”). The upshot is that such minor population deviations are “presumed to be constitutionally valid”&emdash;i.e., more than deviation is needed in such cases. League of Women Voters v. City of Chi., 757 F.3d 722, 725 (7th Cir.), cert. denied, - U.S.-, 135 S.Ct. 688, 190 L.Ed.2d 392 (2014). The voters never address these eases or identify any statements by this court or the Supreme Court contradicting these principles. More importantly, they identify no facts or law supporting an argument that the presumption was rebutted (or even rebuttable) on the facts of this case. Nor did they below, which is a significant omission in light of the assertion in Mr. Sanderoffs report that all districts were within 5% of ideal district size. Finally, as the Supreme Court recently reiterated, a 5% deviation is generally permissible in these cases. Ala. Legis. Black Caucus v. Alabama, - U.S. -, 135 S.Ct. 1257, 1263, 191 L.Ed.2d 314 (2015).
A bare statement that someone could have made an argument is not enough. “[C]ursory statements, without supporting analysis and case law, fail to constitute the kind of briefing that is necessary to avoid application of the forfeiture doctrine.” Bronson v. Swensen, 500 F.3d 1099, 1105 (10th Cir.2007); Adler v. Wal-Mart Stores, Inc., 144 F.3d 664, 679 (10th Cir.1998) (“Arguments inadequately briefed in the opening brief are waived .... ”). Because the voters identify no grounds on which we could rest a conclusion that the district court’s decision departed from governing law, we consider this argument waived.
Underlying all these arguments is the voters’ claim that, if they had not chosen to seek dismissal without prejudice, they may have been able to supplement their expert reports or depose Mr. Sanderoff at a later date. Their point seems to be this could have allowed them to build a counterargument to his assertions about the merit of *1277their claims. But what they could have done is largely beside the point, given their failure, even after realizing sanctions were a possibility, to combat those assertions in any convincing way.15 Mr. San-deroffs report purported to show fatal flaws in the case in June 2013. The question here is whether, when the court considered imposing sanctions in August 2014, it had any reason to doubt those flaws existed. If anything, the voters’ failure to take further steps they now highlight as available to them might cut in favor of the court’s ultimate finding that the case became meritless on June 25, 2013. After all, Mr. Sanderoff s report claimed to identify holes in the voters’ theory; declining to take an opportunity to combat the report could support a negative inference about the quality of that theory.
In short, nothing in the record or in the briefing convinces us the court rested its decision on any legal error. ■
C. Factual Propriety of Sanctions
The voters’ inability to show legal error reduces them to arguing that even without legal errors “the record in this case manifestly does not support the imposition of sanctions.” 1st Cx-App. Br. at 29. They emphasize that the only substantive motion they filed was their opposed motion to dismiss on July 5. In large part, this misses the point. Much of this question turns on whether the court abused its discretion in finding that multiplying proceedings after June 25 was sanc-tionable because the case became meritless at that time. We have already explained why the court did not err in finding the Gingles and one-person-one-vote arguments lacked merit, as well as why the voters cannot blame the court’s stay for extending proceedings. Ample evidence supported the court’s decision. When the court looked back in August 2014, it was apparent that the case was substantively weak from the start. Mr. Sanderoffs report hammered home the intractability of those flaws by showing the Gingles claim could not succeed and that district deviations made it unlikely that the one-person-one-vote claim would. It was not an abuse of discretion to find it became objectively unreasonable to pursue the case further on June 25, 2013.
With June 25 established as the critical day, the remaining question is whether the court abused its discretion in finding the voters in fact acted in a way that unreasonably and vexatiously multiplied proceedings. The voters’ main argument here boils down to a comparison of this case to some of our other § 1927 cases they claim exemplify more egregious conduct. Even if that were true, however, none of those cases suggested they were establishing a floor below which a court could not permissibly impose sanctions. In a vacuum, it matters little to our review for abuse of discretion that some lawyers may have acted worse than the voters did. If they acted in a way the court could have justifiably found to be objectively unreasonable, § 1927 requires no more.
That, however, brings us to the reason we cannot affirm the award in its current form. In short, we cannot affirm what appears to be the court’s finding&emdash; implicit in the imposition of fees beginning on June 25&emdash;regarding what the sanctiona-ble conduct actually was. Section 1927 sanctions are for “conduct that, viewed objectively, manifests either intentional or reckless disregard of the attorney’s duties *1278to the court.” Hamilton, 519 F.3d at 1202 (internal quotation marks omitted) (emphasis added). As explained earlier, this means excess costs and fees for which a sanctioned attorney must pay must have been caused by that attorney’s sanctiona-ble action. Here, the court found the voters’ sanctionable conduct was “continuing] to pursue th[e] case” after the receipt of Mr. Sanderoffs report on June 25 would have made “clear to a reasonable attorney that th[e] case no longer had merit.” App. 407.
To be clear, the court did not abuse its discretion by finding June 25 was the date after which it became sanctionable for the voters to multiply proceedings in an objectively unreasonable way. But, by imposing sanctions beginning on June 25, it necessarily found they did something at that point that multiplied proceedings within the meaning of the statute. Because the voters took no affirmative action on that day, the only potentially sanctionable June 25 “conduct” we can see is their failure to immediately cut the case short&emdash;ie., dismiss the case with prejudice&emdash;after receiving Mr. Sanderoffs damaging report.16
We do not discount the possibility that in some cases, failing to act can be sanc-tionable conduct under § 1927. See, e.g., Roadway Express, Inc. v. Piper, 447 U.S. 752, 755-57, 760, 100 S.Ct. 2455, 65 L.Ed.2d 488 (1980) (observing a party’s “uncooperative behavior” and “deliberate inaction in handling” a case leading to sanctions (internal quotation marks omitted)). After all, the statute “authorizfes] the assessment of costs against dilatory attorneys.” Id. at 760, 100 S.Ct. 2455. Surely an attorney’s failure to act could be objectively unreasonable and vexatious and multiply the proceedings in a case by causing the opposing party to file motions to compel action. More pertinent here, in a meritless case, protracted failure to do anything but dismiss the case (or, perhaps, insisting on conditions of dismissal that themselves create further litigation) might be sanctionable.
But the court’s decision below does not appear to rest on this type of reasoning. Rather, the award in effect found the voters’ unreasonable failure to dismiss the case began the very day they received the report. That cannot be reconciled with the need to construe § 1927 to avoid “dampening] the legitimate zeal of an attorney in representing his client,” Braley, 832 F.2d at 1512. When new information appears to make it objectively unreasonable to pursue a case, an attorney must have at least some time&emdash;surely a day or two&emdash;to study that information and make a decision regarding its impact before failing to drop the case becomes sanc-tionable unreasonable pursuit. See Riddle & Assocs., P.C. v. Kelly, 414 F.3d 832, 835 (7th Cir.2005) (“If a lawyer pursues a path that a reasonably careful attorney would have known, after appropriate inquiry, to be unsound, the conduct is objectively unreasonable and vexatious.” (emphasis added)).
Consequently, we vacate the award of fees and remand for further proceedings consistent with this opinion. The court is free to revisit the fees question on remand and certainly may reimpose fees if it finds a more appropriate triggering action. It is possible that an affirmative action creating further litigation taken by the voters after June 25 might suffice, or that a lengthier delay after June 25 will do on its own. We express no opinion on those possibilities here.17 We only note that lawyers and *1279litigants generally should have some reasonable leeway to review ostensibly damaging materials, even if the materials are, ultimately, the last straw proving their case’s weakness.
But they should keep in mind that this time is not unlimited. The necessary time will of course be case-specific and largely within the discretion of the district court. Some relevant, although non-exclusive factors would include the strength of the underlying case and the nature of the new developments affecting the objective reasonableness of further pursuing the case. We leave any further investigation into those questions for the district court on remand and note also that such investigation may be unnecessary if it finds an affirmative action by the voters that multiplied proceedings in an objectively unreasonable way,18
III. Conclusion
For the foregoing reasons, we VACATE the fee award and REMAND for proceedings consistent with this opinion. We DENY the voters’ motions for sanctions against the Mayor for his cross-appeal.19
. As discussed, the district court sanctioned the attorneys rather than the plaintiffs in the underlying case. For easier reading, references in this opinion to “the voters” only refer to plaintiffs’ lawyers, who are the parties appealing the imposition of sanctions.
. Before the charter amendment, a candidate who was not the candidate of choice of the Latino community could win an election with a plurality while two candidates who were both preferred candidates of the community split a majority. According to the voters, this was how Mayor Berry was elected in 2009. Post-amendment, a runoff would be required between the two top candidates. The voters thought this might moot their original redistricting complaints. See App. 112 (explaining the changes might “remedy the constitutionally deficient redistricting map without further Court intervention").
. A non-exhaustive list of sanctionable conduct includes cases where "an attorney acts recklessly or with indifference to the law,” when "an attorney is cavalier or bent on misleading the court,” when he "intentionally acts without a plausible basis,” or "when the entire course of the proceedings was unwarranted.” Steinert v. Winn Grp., Inc., 440 F.3d 1214, 1221 (10th Cir.2006). But sanctions cannot be imposed for the initiation of proceedings — "it is not possible to multiply proceedings until after those proceedings have begun.” Id. at 1225.
. This matters because the voters argue it “was the decision of the court, and not any conduct by Plaintiffs' counsel, that imposed the stay and resulted] in proceedings that would not have been conducted otherwise.” 1st Cx-App. Br. at 20 (internal quotation marks omitted) (alteration and emphasis in original). But they have not appealed the stay order; their claim is essentially that the sanctions were inappropriate because the stay was inappropriate. As discussed below, that argument has an independent causal infirmity — the decision to award sanctions did not depend on the resolution of any legal issue in the stay order.
. Even after that election passed, the voters did nothing beyond requesting an extension of the stay to determine the amendment's effect on yet another election. That was their last action before the court denied the motion to dismiss without prejudice and granted the Mayor’s motion to dismiss with prejudice in January 2014. By that time, the election-based reasons for seeking dismissal without prejudice lacked salience, and the voters had offered no substitute reasons for dismissal. Failing to offer current, relevant reasons for dismissing without prejudice probably offers an “insufficient explanation of the need for a dismissal.” Brown, 413 F.3d at 1124. The court apparently thought so. See App. 196 (noting the election-centric reasons for dis*1271missal “seem to be disingenuous” since after the election’s passage the voters had not yet "ma[d]e a decision about whether they ha[d] a meritorious lawsuit or not”); id. at 410 (stating none of the reasons for dismissing without prejudice were "compelling”).
. Medtronic Navigation, Inc. v. BrainLAB Medizinische Computersysteme GmbH, 603 F.3d 943 (Fed.Cir.2010); Browning v. Kramer, 931 F.2d 340 (5th Cir.1991); In re Ruben, 825 F.2d 977 (6th Cir.1987).
. It is unclear from the record whether this offer was to dismiss with or without prejudice.
. The denial was based entirely on the underlying case’s lack of merit and the court’s finding that the voters’ claimed reasons for seeking dismissal without prejudice were "disingenuous” and insufficient. App. 196.
.This case’s chronology demonstrates why that would be an odd rule. The offer to proceed in state court was made months after the case became meritless and months after the court found the voters began unreasonably multiplying proceedings in a meritless case.
. The voters point to our statement in Miera v. Dairyland Insurance Co., 143 F.3d 1337 (10th Cir.1998), that Braley "rejected a subjective good faith inquiry.” Id. at 1342. But that takes Miera out of context. That case simply described Braley’s rejection of the argument that § 1927 sanctions “should be imposable against an attorney personally only for subjective bad faith." Braley, 832 F.2d at 1512 (emphasis added). To read that as holding that subjective bad faith has no place in this analysis goes too far, and makes scant sense. Has an attorney discovered to be acting in bad faith but who cloaked that bad faith behind objectively reasonable actions insulated himself from § 1927 sanctions? Nothing in the statute requires that result. It would be strange if subjective bad faith did not constitute "conduct that, viewed objectively, manifests ... intentional ... disregard of the attorney’s duties to the court.” Id.
. Moreover, the court plainly rested its sanctions award on counsel's "objectively unreasonable” pursuit of the case after June 25. App. 407. Even on the voters’ erroneous view of the law, a court that expressly bases sanctions on an attorney’s objectively unreasonable conduct surely would not err by commenting on the possibility that the conduct stemmed from subjective bad faith.
.The voters also attack the court's Gingles conclusion on the grounds that the court erroneously based its decision on the fact that "the challenged map resulted in five of the [city’s] nine districts being majority-minority.” 1st Cx-App Br. at 42. That reads too much into the order. To be sure, the court juxtaposed that observation about the number of majority-minority districts with the claim that the plan disadvantaged minority voters. But this was merely a general observation about the factual context of the case, not a reason for rejecting the Gingles claim. It is wrong to assert, as the voters do, that the court found the Gingles claim meritless “because several majority-minority districts were present.” Id. at 43 (emphasis added). That conclusion flowed from the failure of their experts to even attempt to satisfy Gingles’s third prong.
. The Mayor denies the population deviation was even that high. Because we conclude the voters waived this argument, we do not reach that point.
. Although we hold the argument waived on the basis of an insufficient opening brief on appeal, we perhaps could have held it was forfeited below because it was not made there&emdash;at least, not in the form it takes now. The voters certainly made no such argument in their filings opposing sanctions. And, at the hearing, their argument was again based on nothing but Larios and appeared to be that Albuquerque needed to justify any deviation above zero. See, e.g., App. 453 (asking May- or’s expert whether "the Supreme Court case” [sic] of “Cox v. Larios ” “mandated] that deviation needs to be taken down to zero unless you can articulate why you could not do that”). Of course, the Supreme Court's *1276summaiy affirmance in Larios "mandated” no universal principles, and the cases squarely reject the proposition that Albuquerque needed to justify any deviation above zero. See Ala. Legis. Black Caucus v. Alabama, - U.S.-, 135 S.Ct. 1257, 1263, 191 L.Ed.2d 314 (2015) (noting a state plan that attempted to avoid deviating from the ideal "by more than 1%” pursued "a more rigorous deviation standard than our precedents have found necessary under the Constitution”); see also Voinovich v. Quilter, 507 U.S. 146, 161, 113 S.Ct. 1149, 122 L.Ed.2d 500 (1993); White v. Regester, 412 U.S. 755, 764, 93 S.Ct 2332, 37 L.Ed.2d 314 (1973). Moreover, the voters only touched on this point in questions to the Mayor’s expert; we detect no clear one-person-one-vote arguments in their actual oral argument. To now argue their claim was meritorious because it attacked the constitutionality of 5% deviations may be a stretch. Because we conclude the point was in any event waived in the opening brief, we need not decide this question.
. At no point in contesting sanctions did they directly assert Latinos were unable to elect their candidates of choice, which was the key Gingles issue highlighted by Mr. San-deroff's report. Nor did they make any one-person-one-vote argument not premised on the misunderstanding of Larios we discussed above.
. A dismissal with prejudice, of course, was the only type of dismissal to which the Mayor was willing to acquiesce.
. Doing so would be largely pointless. With respect to "a matter committed to the district court’s discretion,” like the propriety of § 1927 sanctions, "we cannot invoke an alternative basis to affirm unless we can say as a *1279matter of law that it would have been an abuse of discretion for the trial court to rule otherwise.” Ashby v. McKenna, 331 F.3d 1148, 1151 (10th Cir.2003) (internal quotation marks omitted). Even if a more appropriate triggering date exists, we cannot say on this record that it would have been an abuse of discretion for the district court to decline to impose sanctions.
. We note, however, that because the Mayor did not contest the voters' argument that attorney Antonio Maestas should not have been sanctioned, § 1927 liability on remand should not extend to Mr. Maestas.
. The Mayor cross-appealed in this matter, asking us to find the district court abused its discretion in failing to award sanctions under a host of other provisions. At oral argument, however, the Mayor conceded that pursuing the cross-appeal further was unwise, and dropped it. After oral argument, the voters moved for sanctions under § 1927 and Federal Rule of Appellate Procedure 38, which allows the award of damages for a frivolous appeal. We deny that motion.