Wood, Circuit Judge.

Brian J. Kerwin, Attorney, Office of the United States Attorney, Chicago, IL, for Plaintiff-Appellee.

Patrick W. Blegen, Kelsey Hipp Killion, Attorneys, Blegen & Garvey, Chicago, IL, for Defendant-Appellant.

Before Kanne, Rovner, and Wood, Circuit Judges.

Wood, Circuit Judge.

These consolidated appeals come to us from three defendants, Michael Bravo, Thomas Luczak, and Ricardo Denava, each of whom challenges his sentence in connection with his involvement in the Latin Kings street gang. The operative indictments alleged that the Latin Kings had been involved in multiple acts of murder, arson, robbery, extortion, witness tampering, and the illegal distribution of narcotics. All three defendants were charged in 2018 with racketeering conspiracy in violation of 18 U.S.C. § 1962(d). Bravo and Denava pleaded guilty in November 2018, while Luczak was convicted in April 2019 after a six-week jury trial.

Although the factual background is roughly the same for all three, the legal issues they raise on appeal differ. Bravo argues that the district court erred by adding criminal history points for two misdemeanor convictions he had under the Illinois "streetgang contact" statute. Luczak contends that the district court should not have included, as part of his criminal history, points for a murder he allegedly committed. The problem is a familiar one: at trial, the jury acquitted him of that offense, using the reasonable-doubt standard, but the court found at sentencing that his responsibility for the murder was established by a preponderance of the evidence. And Denava claims that the district court failed adequately to consider several mitigating factors under 18 U.S.C. § 3553(a) and therefore abused its discretion.

Because the district court erred in counting Bravo's two misdemeanor offenses toward his criminal history, and that error may have affected his ultimate sentence, we reverse and remand Bravo's case for resentencing. But we see no error in Luczak's and Denava's sentences, and therefore affirm them. We take each defendant's case in turn.

I. Bravo (No. 20-1105)

A

In July 2009, Bravo aided and abetted a drive-by shooting orchestrated by the Bush Chapter of the Latin Kings. For that offense, he was prosecuted in state court and received a sentence of five years' imprisonment. He was paroled in January 2014. A few months later, Bravo and a group of Latin Kings were driving around together in a car, when a Chicago police officer stopped the car for a minor traffic violation. Bravo was arrested for a violation of 720 ILCS § 5/25-5 (since amended), which at the time made it illegal to have "direct or indirect contact with a streetgang member" while on parole. He pleaded guilty and was released from jail the next day, with a sentence of two days' time served. Not long after, in September 2014, Bravo was again arrested and charged with streetgang contact after the police observed him drinking alcohol in an alley with other members of the Latin Kings. For the second time, he pleaded guilty to streetgang contact, was released the next day, and received another sentence of two days' time served.

In the first superseding indictment pertinent to the present case, Bravo was charged with (among other things) racketeering conspiracy under 18 U.S.C. § 1962(d). He pleaded guilty in November 2018 and was sentenced in January 2020 to 108 months' imprisonment and three years of supervised release. At sentencing, the district court added two points to Bravo's criminal history in light of the 2014 misdemeanor convictions; that adjustment bumped him from criminal history category III to category IV. Coupled with his adjusted offense level of 30, he faced an advisory guidelines range of 135 to 168 months. Without the extra points attributable to the misdemeanors, his advisory range would have been 121 to 151 months. Bravo's only argument on appeal is that the court committed reversible legal error by counting the misdemeanor offenses in his criminal history.

B

The pivotal question is whether the district court miscalculated Bravo's criminal history score. This is an issue that we consider de novo. United States v. Wallace , 991 F.3d 810, 814 (7th Cir. 2021). Under the guidelines, all felonies and misdemeanors are presumptively counted in a defendant's score. But there are exceptions:

(1) Sentences for the following prior offenses and offenses similar to them, by whatever name they are known, are counted only if (A) the sentence was a term of probation of more than one year or a term of imprisonment of at least thirty days, or (B) the prior offense was similar to an instant offense.... [enumerated offenses follow].

Guidelines § 4A1.2(c)(1). Each of Bravo's sentences entailed only two days' time served—well under thirty days—and is thus eligible for exclusion. The parties agree that "disorderly conduct or disturbing the peace" is the enumerated exclusion that most closely resembles Illinois's streetgang-contact offense, and so we need not examine the other identified offenses. The question is whether the resemblance between this enumerated offense and Bravo's 2014 misdemeanors is close enough to require their exclusion.

We begin with Application Note 12(A), which we treat as "part of the Guidelines themselves," not just "commentary on them." United States v. Kohl , 910 F.3d 978, 980 (7th Cir. 2018). Note 12(A) calls for a "common sense approach" to determining similarity between enumerated and unenumerated offenses. It offers five relevant factors to consider:

(i) a comparison of punishments imposed for the listed and unlisted offenses; (ii) the perceived seriousness of the offense as indicated by the level of punishment; (iii) the elements of the offense; (iv) the level of culpability involved; and (v) the degree to which the commission of the offense indicates a likelihood of recurring criminal conduct.

Guidelines § 4A1.2, cmt. 12(A). Some of these suggest a categorical or abstract approach to the comparison, while others refer to the actual offense conduct involved in a given case.

Following this roadmap, we first compare the punishments for streetgang contact and disorderly conduct under Illinois law. Streetgang contact is a Class A misdemeanor offense, see 720 ILCS § 5/25-5(b), carrying a maximum imprisonment length of less than one year and a fine range of $75 to $2,500 per offense, see 730 ILCS § 5/5-4.5-55(a), (e). Disorderly conduct, in contrast, takes a number of forms, ranging from a Class C misdemeanor to a Class 3 felony. See 720 ILCS § 5/26-1. For instance, an act done "in such unreasonable manner as to alarm or disturb another and to provoke a breach of the peace," the catch-all provision and least serious form, is a Class C misdemeanor punishable by a maximum of thirty days' imprisonment and at most a $1,500 fine per offense. See 730 ILCS § 5/5-4.5-65. But more specific and serious forms of disorderly conduct include Class A misdemeanors such as "enter[ing] upon the property of another" to "deliberately look[ ] into a dwelling," Class 4 felonies such as false fire alarms, or the Class 3 felony of making a false bomb threat. See 720 ILCS § 5/26-1. Because section 4A1.2(c) says that "sentences for all felony offenses are counted" toward criminal history, we need not consider further the felony variety of disorderly conduct.

The question thus is which misdemeanor (the Class C, Class B, or Class A variant) is the relevant point of comparison? If we look to the Class C disorderly-conduct misdemeanor and compare it to streetgang contact, we might conclude that the streetgang offense is slightly more severe, though the difference is not overwhelming. On the other hand, lenity principles might counsel us to take the opposite approach and assume that the most serious disorderly-conduct misdemeanor—here, Class A—provides the better point of comparison. Without deciding this question, we will assume (favorably to the government) that the Class C misdemeanor provides the correct point of comparison. The other four considerations nevertheless support a finding that Bravo's offenses are substantially equivalent to the disorderly-conduct offense described in section 4A1.2(c)(1).

The second consideration pertains to the relative seriousness of streetgang contact compared to disorderly conduct "as indicated by the level of punishment." Guidelines § 4A1.2, cmt. 12(A). Rather than looking only at the possible punishments available under state law, which would render this consideration redundant of the first, we believe that the guidelines call on us to look at the actual punishment imposed for the streetgang contacts. The guideline refers not to the potential or statutory term, but instead to the "sentence" itself. See § 4A1.2(c)(1)(A); see also United States v. Garrett , 528 F.3d 525, 529 (7th Cir. 2008) (considering both the punishments available under the relevant statutes as well as the sentencing court's decision to impose a fine rather than jail time); United States v. Burge , 683 F.3d 829, 835 (7th Cir. 2012) (same). Other circuits have also looked to the sentence that was imposed. See, e.g., United States v. DeJesus-Concepcion , 607 F.3d 303, 305 (2d Cir. 2010) (per curiam) (describing how a district court may consider "the actual conduct involved and the actual penalty imposed"); United States v. Reyes-Maya , 305 F.3d 362, 367 (5th Cir. 2002) ("More important than the statutory range of punishments is the actual punishment given, as the level of punishment imposed for a particular offense serves as a reasonable proxy for the perceived severity of the crime."); United States v. Grob , 625 F.3d 1209, 1216 (9th Cir. 2010) ("agree[ing] with the Fifth Circuit's observation" in Reyes-Maya and looking to the plaintiff's "actual punishment"). Here, both times Bravo was arrested, released from jail the following day, and received a sentence of two days' time served. This indicates lack of severity.

We next compare the elements of streetgang contact to those of disorderly conduct. The district court did not squarely address this consideration, nor have we previously had occasion to elaborate on it. The government suggests that the sheer number of elements reflects the complexity of an offense, which in turn tracks an offense's overall severity. But complexity and severity are imperfectly related, at best. A cold-blooded murder might not be complex, but it is certainly severe, while a clumsy Ponzi scheme may be highly complex but end up inflicting only minimal damage. The better approach is to compare the core conduct of each offense as indicated by its constituent elements. This offers a better gauge of an offense's severity, and it favors Bravo.

The Class C disorderly-conduct offense has three elements: "a person must engage in conduct that: (1) is unreasonable; (2) alarms or disturbs another; and (3) threatens to provoke or provokes a breach of the peace." Reher v. Vivo , 656 F.3d 772, 775 (7th Cir. 2011) (defining the elements for 720 ILCS § 5/26-1(a)(1) ). The 2014 version of the streetgang-contact statute required only two elements. First, a person had to "knowingly ha[ve] direct or indirect contact with a streetgang member." 720 ILCS § 5/25-5(a) (2014). And second, one of a group of listed predicate conditions had to be met: either that the defendant was sentenced to probation, released on bond, or released from prison with the stipulation that he or she refrain from contact with gang members, or that the defendant was ordered by a judge in a non-criminal proceeding to refrain from contact. Id. § 5/25-5(a)(1)-(4).

Whether we describe those elements concisely, or we subdivide each part, the underlying substance remains the same. And it is that which matters for present purposes. The comparison of elements helps to indicate the relative severity of the conduct generally associated with each offense. While the streetgang-contact offense as of 2014 required only that a parolee be in the presence of a streetgang member, thus bearing the hallmarks of a passive status offense, disorderly conduct requires that a person undertake an action that alarms or disturbs. We are not persuaded that the former is categorically more serious than the latter.

This is also why the fourth consideration—"the level of culpability involved"—cuts in Bravo's favor. We have sometimes viewed this culpability determination through the lens of comparative severity. See United States v. Hagen , 911 F.3d 891, 896 (7th Cir. 2019) ("Refusing to support a child strikes us as a more severe offense than allowing truancy.... We think that one who deserts a child or fails to provide for the child's basic necessities is more culpable than one who permits a child to skip school."). Like the second consideration's contemplation of the actual punishment imposed, the fourth calls on us to evaluate the actual conduct involved in the case. When we do so, we see that neither of Bravo's streetgang contacts appears to be more severe or blameworthy than a typical act of disorderly conduct. On the first occasion, Bravo was just a passenger in a vehicle with gang members when a minor traffic violation was committed. On the second, he was drinking alcohol in an alley with a few gang members. Neither instance involved weapons or drugs, or accusations of disturbing or otherwise harming others.

The last point of comparison Note 12(A) recommends is "the degree to which the commission of the offense indicates a likelihood of future criminal conduct." The district court decided this in favor of the government, emphasizing the factual similarities between being in a car with gang members and Bravo's July 2009 predicate federal conviction involving a drive-by shooting. But the fact that both offenses took place in a car seems at best a superficial connection, especially given the fact that no weapons or drugs were present during the streetgang-contact arrest. The district court reasonably noted that because streetgang contact involves the violation of an express condition of parole, it indicates some likelihood to recidivate. But this is a comparative inquiry. We must ask whether riding in a car or drinking alcohol with gang members after being ordered not to do so indicates a higher likelihood of future criminal wrongdoing than does the hazardous or threatening behavior often involved in disorderly conduct. This is a close call, though one that may marginally favor the government.

But for purposes of the guideline, ties go to the defendant. In order to overcome the exemption for disorderly conduct, Bravo's streetgang contact must be more severe than disorderly conduct. Our review of the considerations identified by Note 12(A) convinces us that Bravo has demonstrated the required equivalence between the two offenses involved in his case. We therefore find that the court erred in adding those two criminal history points, which in turn resulted in a guidelines range of 135 to 168 months.

Although we do not rely on later changes in the relevant Illinois statutes, it is notable that the streetgang contact statute was amended in 2018 in a manner that might even render Bravo's streetgang contacts non-criminal today. When Bravo was convicted in 2014, the statute criminalized "unlawful contact with streetgang members." The amended version of the statute criminalizes "unlawful participation in streetgang related activity" that involves "knowingly commit[ting] any act in furtherance of streetgang related activity. " 720 ILCS § 5/25-5(a) (effective Jan. 1, 2018) (emphasis added). Association or contact alone, without active participation in a designated gang-related activity, is no longer criminalized.

C

Bravo cannot prevail solely on the fact that the district court miscalculated the guidelines range. We must also decide whether the error was prejudicial or harmless. United States v. Shelton , 905 F.3d 1026, 1031 (7th Cir. 2018). As we said in United States v. Corner , 967 F.3d 662 (7th Cir. 2020), a "procedural error (such as a miscalculation of the applicable guideline range) is not reversible if it's clear that the court did not rely on it when selecting the sentence." Id. at 666. But this is a high bar. See Rosales-Mireles v. United States , ––– U.S. ––––, 138 S. Ct. 1897, 1907, 201 L.Ed.2d 376 (2018) (emphasis added) ("When a defendant is sentenced under an incorrect Guidelines range—whether or not the defendant's ultimate sentence falls within the correct range—the error itself can, and most often will , be sufficient to show a reasonable probability of a different outcome absent the error.").

Often a district court will make a statement indicating that its resolution of a disputed guidelines issue ultimately does not affect its choice of a sentence; instead, the judge may say, its sentence results from the entire process of selecting the correct advisory range before weighing the factors identified by section 3553(a). There are benefits from such a statement, and also a few risks. On the benefit side, it can avert the need for resentencing, even if the judge's resolution of the guidelines issue was not correct. On the risk side, the ability simply to say at the end nothing more than "my sentence would be the same no matter how I approach the guidelines" threatens to make the guidelines irrelevant. That would be impermissible, as we know from 18 U.S.C. § 3553(a)(4), as well as Supreme Court decisions such as Gall v. United States , 552 U.S. 38, 128 S.Ct. 586, 169 L.Ed.2d 445 (2007). The Court repeatedly has held that "a district court should begin all sentencing proceedings by correctly calculating the applicable Guidelines range." Id. at 49, 128 S.Ct. 586 ; see also Rosales-Mireles , 138 S. Ct. at 1904 ; Molina-Martinez v. United States , 578 U.S. 189, 193, 136 S.Ct. 1338, 194 L.Ed.2d 444 (2016) ; Peugh v. United States , 569 U.S. 530, 536, 133 S.Ct. 2072, 186 L.Ed.2d 84 (2013).

Judges are, however, entitled to adopt their own sentencing philosophy based in the considerations of section 3553(a), and so they are not compelled to accept the advice that the guidelines offer. See Pepper v. United States , 562 U.S. 476, 490, 131 S.Ct. 1229, 179 L.Ed.2d 196 (2011) (citing Kimbrough v. United States , 552 U.S. 85, 101, 128 S.Ct. 558, 169 L.Ed.2d 481 (2007) ). In a particular case, for example, a judge might conclude that a relatively low drug quantity for the offense of conviction should not drive the result, if that defendant has an especially high criminal history level. Yet if there is a significant dispute about drug quantity, the Gall line of cases requires the judge to resolve it. Only then should the judge explain that even if that resolution were different, other considerations have taken precedence for this defendant's sentence. When the judge goes beyond a flat statement that the guidelines were irrelevant and offers a specific reason why a dispute under the guidelines did not affect the sentence, we are free to accept that explanation. Such an alternative explanation must, however, be tied to the decisions the court made and why they did or did not affect the ultimate outcome.

This is just what the court did in United States v. Abbas , 560 F.3d 660 (7th Cir. 2009). There, the judge "expressly stated that she would have imposed the same sentence even if § 2C1.1 [of the guidelines] did not apply to the defendant's sentence." Id. at 667. The issue was clear, and the judge gave "a detailed explanation of the basis for the parallel result; ... not just a conclusory comment tossed in for good measure." Id. That in turn provided a sound basis for a finding of harmless error.

We cannot say the same about this case. It is not clear to us whether, had the two misdemeanors not been included in Bravo's criminal history, the district court would have imposed 108 months' imprisonment. The court's sentence was firmly anchored in the range recommended by the guidelines. After evaluating the section 3553(a) factors, including the seriousness of the offense, general and specific deterrence, and Bravo's rehabilitation thus far, the court explained its sentence as follows:

I'm going to do a few things. I'm going to start with—if we're looking at the guideline range of 135 to 168, if I go from the 168, I'm going to give him the four years for the RICO time he served for the predicate, which brings you down to the 120 sentence [i.e. , 168–48=120]. And then beyond that, I'm going to give you 12 months for your rehabilitation efforts, which leaves you with a sentence of [108] months, which is nine years.

(The court initially misspoke and said 109 months at that point, but it quickly corrected itself.)

After taking a brief recess and advising Bravo of the conditions of his supervised release, the court offered the following coda upon which the government now relies for its harmless-error argument:

Now, a few things I want to say at the end as well. The two points for criminal history that I added, I would have imposed the same sentence of 108, regardless of what criminal history category. As you know, it's below both the IV and the III categ[ory] level anyways.

This statement, however, tells us nothing about the way in which the court was evaluating Bravo's criminal history, or why a sentence so tied to one guidelines range would have come out the same way with a different starting point. The court seemed to treat the aggravating factors (seriousness of the offense and the need for future deterrence) as the reason to begin with the top end of whatever range was assumed. But that antecedent choice influenced each subsequent step in the arithmetic. Had the court anchored itself in the 121-to-151 range, it may well have begun with the top end of 151 months, subtracted 60 months (4 years for time served and 12 months for rehabilitation), and arrived at a sentence of 91 months. The fact that 108 months is below the 121-to-151 revised guidelines range is neither here nor there. It does not obviate the need for an explanation of why the court found the contested aspect of Bravo's criminal history to be irrelevant.

To be clear, we are not saying that the court could not land on 108 months as the proper sentence; it is just that we cannot tell on this record whether it would have done so had it correctly recognized that Bravo was in criminal history category III, not IV. We conclude, therefore, that Bravo is entitled to resentencing. II. Luczak (No. 20-1484)

A

Thomas Luczak was charged with racketeering conspiracy in the second superseding indictment. The government also filed a Notice of Special Findings alleging that Luczak had murdered a rival gang member named Juan Serratos on June 11, 2000. Luczak, along with three co-defendants who are not involved in this appeal, proceeded to trial before a jury in March 2019.

The jury heard extensive testimony related to the Serratos murder. Alexander Vargas testified that Luczak once admitted to him that he shot Serratos. Jose Zambrano testified that Vargas had told him in 2004 that Luczak had committed the murder. Paulino Salazar testified that he gave Luczak the gun to shoot Serratos on the night of the murder, and that he heard Luczak announce later that night that he shot Serratos in the chest. And Francisco Barajas and Isidro Mendez—two purported eyewitnesses to the shooting—each testified to have been standing next to Serratos when he was shot. Barajas claimed to have heard several gunshots but could not make out the shooter, while Mendez claimed to have seen the silhouette of the shooter in the distance. In his closing argument, Luczak argued that Salazar's testimony, which formed the crux of the prosecution's case, was not credible or corroborated. The jury ultimately found Luczak guilty of racketeering conspiracy but not guilty of the Serratos murder.

At the sentencing hearing on March 9, 2020, the government argued that the trial testimony, even if not enough to prove guilt beyond a reasonable doubt, showed by a preponderance that Luczak shot Serratos. On that basis, it argued that Luczak's base offense level should be elevated from 33 to 43. It relied principally on Salazar's trial testimony, along with a published portion of a February 2015 post-arrest interview with Salazar, in which Salazar again identified Luczak as the shooter. The court ultimately found by a preponderance that Luczak murdered Serratos. Salazar's testimony, it pointed out, was consistent with forensic evidence from the crime scene, including a bullet recovered from the scene that matched the gun Salazar handed to Luczak, and it stated that Salazar's February 2015 interview answers appeared "very spontaneous" and thus credible.

Luczak sought a two-level reduction for acceptance of responsibility under section 3E1.1(a) of the guidelines, arguing that he did not contest his guilt with respect to the racketeering conspiracy at trial but instead focused exclusively on the Serratos murder notice. The district court rejected this argument. It then analyzed the section 3553(a) factors, noting Luczak's strong family ties and likelihood of desisting from crime at this stage in his life, but also emphasizing the seriousness of Luczak's long-term involvement in the gang and the need to deter others. It ultimately imposed a below-guidelines sentence of 210 months' imprisonment.

Luczak now argues that the government failed to prove by a preponderance of the evidence that he killed Serratos. Relatedly, he argues that the use of acquitted conduct in calculating the guidelines range violated his constitutional rights and created an unjustified sentencing disparity. Finally, he complains that the court erred in not awarding the two-point reduction for acceptance of responsibility.

B

We first consider whether the district court erred in determining that Luczak murdered Serratos. We review this decision, which relates to relevant conduct under section 1B1.3 of the guidelines, for clear error. United States v. King , 910 F.3d 320, 329 (7th Cir. 2018). This is a high hurdle to clear, and Luczak has not done so.

While Luczak identifies points of tension in the testimony, such tensions are routine and do not in this case show that the district court clearly erred in its determination. For instance, Luczak argues that the eyewitness testimonies of Mendez and Barajas contradicted Salazar's testimony, as Mendez and Barajas claimed that the shooter was some distance away from Serratos while Salazar reported that Luczak had told him that the shooting took place at point-blank range. But Salazar may have misremembered this detail, or Luczak may have embellished his account of the shooting when describing it to Salazar. To take another example, Luczak argues that Vargas's and Zambrano's stories conflict: Vargas testified that he "didn't ask for specifics" when Luczak initially mentioned the shooting to him, while Zambrano testified that Vargas gave him a detailed version of the shooting. But this disagreement strikes us as minor and does not undercut the thrust of Vargas's and Zambrano's accounts: that Luczak admitted the shooting to Vargas.

Luczak further argues that Salazar, Vargas, and Zambrano—all of whom are gang members—had powerful motivations to lie. But this concern arises every time the government puts on cooperating witnesses. The judge evaluates that factor in the course of assessing witness credibility—an issue on which the district court is owed significant deference. See United States v. Morales , 655 F.3d 608, 646–47 (7th Cir. 2011). Finally, Luczak argues that Salazar's description at trial of how he gave Luczak the gun used in the shooting (a fact confirmed by ballistics evidence) was contradicted by the absence of any mention of a gun in Salazar's 2015 post-arrest statement. But the absence of a gun reference does not impeach Salazar's testimony or render the two statements irreconcilable. It is not as if the 2015 statement affirmatively contradicted the later testimony (say, if the 2015 statement claimed that the gun was handed to someone other than Luczak). We therefore see no reversible error in the district court's finding.

Luczak also argues that the court's consideration of acquitted conduct at sentencing violated his constitutional rights. He is free to do so, as long as he realizes that all we can do is confirm that he has preserved this point for further consideration in the Supreme Court or Congress. We are bound by the Court's decision in United States v. Watts , 519 U.S. 148, 157, 117 S.Ct. 633, 136 L.Ed.2d 554 (1997), which found no such problem.

Next Luczak contends that the district court created an unwarranted sentencing disparity by announcing that it would have imposed the same 210-month sentence even without the preponderance finding. But that is not all that the court did. Had it excluded the murder as relevant conduct, the revised guidelines range would have been 168 to 210 months' imprisonment. Acknowledging this counterfactual possibility, the district court stated the following: "I actually think the 210, that is the high end of the range ... without a supervisory role and without the murder, I actually think that's the appropriate place for you." This is closer to the kind of specific comment on a guidelines argument that we found lacking earlier. And in any event, we have found no error in the court's calculation of Luczak's advisory guidelines range, and so the court's additional comment is of no moment here.

Finally, Luczak argues that the district court erred in refusing to adjust his offense conduct for acceptance of responsibility under section 3E1.1 of the guidelines. The decision to take a case to trial does not automatically disqualify a defendant from this adjustment, Guidelines § 3D1.1 cmt. 2, and Luczak contends that he took the case to trial only to challenge the special finding about the Serratos murder—a point on which he prevailed at the guilt phase. Nonetheless, we give great deference to a trial judge's denial of an acceptance of responsibility reduction, which is "only to be reversed if this Court is left with the definite and firm conviction that a mistake has been committed." United States v. Collins , 796 F.3d 829, 835 (7th Cir. 2015). Here we see no such mistake. While Luczak may not have contested the racketeering conspiracy charge as such, his denials of the related murder allegation—denials the district court ultimately determined to be untrue, albeit under a preponderance standard—provide a sufficient basis for finding that this is not a "rare situation[ ]" where a case brought to trial still warrants the reduction. Guidelines § 3E1.1, cmt. 2.

The district court did not clearly err in attributing the Serratos murder to Luczak and in refusing to grant him an adjustment for acceptance of responsibility, and so we affirm his sentence.

III. Denava (No. 20-3477)

A

Ricardo Denava was charged in the first superseding indictment with one count of racketeering conspiracy, to which he pleaded guilty in November 2018. The court held a sentencing hearing in November 2020. The Pre-Sentence Report set his offense level at 37 and his criminal history at Category III, which resulted in a preliminary sentencing range of 262 to 327 months' imprisonment. The court then lowered the range to a point—240 months—in light of the statutory maximum.

At sentencing, Denava argued that the maximum 240-month sentence was excessive in light of mitigating considerations, including the fact that he was abandoned by his parents and fell under the sway of his uncle, who was a high-ranking Latin King, that he struggles with anxiety and depression and would benefit from drug treatment, that he had a strong employment history, and that he had complied with his pretrial release conditions and accepted responsibility for his role in the gang. After listening to Denava's arguments, the court acknowledged his struggles with alcoholism and the sincerity of his familial support, but emphasized the need to deter others away from entering gang life. It then sentenced Denava to a below-guidelines sentence of 200 months' imprisonment.

On appeal, Denava argues that the 200-month sentence was substantively unreasonable because the court did not afford sufficient weight to his arguments for mitigation. The first of those arguments concerned two misdemeanor marijuana convictions, which Denava says should have been disregarded in his criminal history; the second related to his positive employment history while on bond for this case; and the third accused the court of placing too much weight on his leadership role in the gang.

B

We review the substantive reasonableness of a sentence for abuse of discretion. United States v. Bridgewater , 950 F.3d 928, 933 (7th Cir. 2020). This is a high bar, particularly when a below-guidelines sentence is imposed. See United States v. De La Torre , 940 F.3d 938, 954 (7th Cir. 2019) ("A below-Guidelines sentence will almost never be unreasonable ....").

We see no error, let alone anything unreasonable, in the district court's consideration of Denava's mitigation arguments. Nor did the court commit an obvious oversight when it found that the aggravating considerations—namely, Denava's extensive participation in gang activity and leadership—outweighed any positive factors. The court highlighted Denava's leadership stints in the gang, first as "Inca" and later as "Regional Enforcer" of the 97th Street chapter of the Latin Kings. It noted the violent character of the gang. And while his criminal history score incorporated two points for nonviolent misdemeanor marijuana offenses, the sentencing record does not indicate that the court placed undue weight on criminal history. Deducting those two points would have produced a guidelines range of 210 to 262 months, which is still greater than the 200 months ultimately imposed. The court indicated its awareness of other mitigation considerations, including Denava's acceptance of responsibility, complicated childhood, and ongoing commitment to supporting his family.

While Denava is to be commended for his decision to take responsibility for his involvement in the gang, and for his diligence in securing employment and attempting to provide for his family through legitimate work, it is not our role to reweigh the section 3553(a) factors. We therefore find that the district court did not abuse its discretion and affirm Denava's sentence.

IV

For the foregoing reasons, we REVERSE the judgment of the district court with respect to Michael Bravo and REMAND his case for resentencing. We AFFIRM the judgments with respect to Thomas Luczak and Ricardo Denava.