Defendant-Appellee Guadalupe Garcia-Lara pleaded guilty to one count of possession with intent to distribute 500 grams or more of methamphetamine in violation of 21 U.S.C. § 841(a)(1). Because Mr. Garcia-Lara had two prior convictions for controlled substance offenses, the “career offender” enhancement applied to his advisory sentence under the U.S. Sentencing Guidelines (“U.S.S.G.” or “Guidelines”). See U.S.S.G. § 4B1.1. Believing the career offender enhancement overstated Mr. Garcia-Lara’s criminal history, the District Court sentenced him to a below-Guidelines sentence of 140 months’ imprisonment. The Government appeals that sentence as substantively unreasonable. Exercising jurisdiction pursuant to 18 U.S.C. § 3731, we vacate Mr. Garcia-Lara’s sentence and remand for resentencing.

*1135I. BACKGROUND

On November 20, 2004, a Kansas Highway Patrol trooper stopped Mr. Garcia-Lara for speeding as he was driving on Interstate 35 near Emporia, Kansas. The trooper received consent to search the vehicle from Mr. Garcia-Lara, the sole occupant of the vehicle, and found approximately 18 pounds of marijuana and 557 grams of methamphetamine hidden inside two spare tires found in the trunk of the vehicle. A grand jury returned an indictment against Mr. Garcia-Lara, charging him with one count of possession with intent to distribute 500 grams or more of methamphetamine in violation of 21 U.S.C. § 841(a)(1). Mr. Garcia-Lara pleaded guilty to the charge without a plea agreement on February 17, 2005.

The U.S. Probation Office prepared a Presentence Investigation Report (“PSR”) in anticipation of Mr. Garcia-Lara’s sentencing. The PSR reported a criminal history category of V and an initial base offense level of 32. Because two of Mr. Garcia-Lara’s prior convictions were for controlled substance offenses as defined in U.S.S.G. § 4B 1.2(b), the PSR applied the “career offender” provision of the Guidelines, U.S.S.G. § 4Bl.l(b)(A), raising his criminal history category to VI and his base offense level to 37. After applying a three-level reduction to the offense level for acceptance of responsibility, the PSR concluded Mr. Garcia-Lara had a total offense level of 34 and a criminal history category of VI, resulting in an advisory Guidelines sentence of 262 to 327 months’ imprisonment.

Applying 18 U.S.C. § 3553(a), the District Court concluded that a sentence of 262 months, at the bottom of the advisory Guidelines range, over-represented Mr. Garcia-Lara’s criminal history, resulting in a sentence greater than necessary to accomplish the goals of § 3553(a). Accordingly, the court sentenced Mr. Garcia-Lara as if the career offender enhancement did not apply. Noting that the advisory Guidelines sentence for a non-career offender would be 140 to 175 months’ imprisonment, the District Court sentenced Mr. Garcia-Lara to 140 months.

II. DISCUSSION

A. Post-Rita Sentencing Review

Since United States v. Booker, 543 U.S. 220, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005), this Court has repeatedly stated that we review a district court’s sentencing determination for reasonableness, which is guided by the statutory factors delineated in 18 U.S.C. § 3553(a). See, e.g., United States v. Kristl, 437 F.3d 1050, 1053 (10th Cir. 2006). In the Supreme Court’s recently issued opinion Rita v. United States, — U.S.-, 127 S.Ct. 2456, 168 L.Ed.2d 203 (2007), the majority referred to reasonableness review as an “abuse of discretion” standard. See id. at 2465 (“Given our explanation in Booker that appellate ‘reasonableness’ review merely asks whether the trial court abused its discretion, the presumption [of reasonableness applied to a within-Guidelines sentence] applies only on appellate review.”). We write now, in some detail, to make clear that the Supreme Court’s reference in Rita to reasonableness review as an abuse-of-discretion standard does nothing to change the appellate reasonableness standard this Circuit has applied since Booker. As we explain, our case law post-Booker has consistently applied the principle that reasonableness, as defined by § 3553(a), is the measure of the appropriate use of a district court’s discretion; or, put differently, we review a district court’s sentence for abuse of discretion, asking whether it is reasonable under the § 3553(a) factors.

Our application of the reasonableness standard of review since Booker has necessarily been a review of a district court’s *1136decision for abuse of discretion. For example, because a legal standard based on reasonableness is inherently fact dependent, see Cooter & Gell v. Hartmarx Corp., 496 U.S. 384, 401-05, 110 S.Ct. 2447, 110 L.Ed.2d 359 (1990), we have implicitly acknowledged that we employ an abuse-of-discretion standard by reviewing a district court’s factual findings for clear error, United States v. Valtierra-Rojas, 468 F.3d 1235, 1241 n. 8 (10th Cir. 2006). Moreover, as the dissent recognizes, Booker itself implicitly equates reasonableness review with review for abuse of discretion, and Rita simply makes that equivalence explicit. In short, Rita says nothing new about the standard of review.

But even though we accord deference to a district court’s sentencing decision, it is clear that district courts must apply, and our appellate review is guided by, the factors set forth in 18 U.S.C. § 3553(a). See United States v. Pruitt, 487 F.3d 1298, 1303 (10th Cir. 2007) (“The district court must consider the sentencing factors set forth in 18 U.S.C. § 3553(a) when imposing a sentence.”); Kristl, 437 F.3d at 1053 (noting that appellate reasonableness review is “guided by the factors set forth in 18 U.S.C. § 3553(a)”). Because reasonableness in sentencing is defined by statutory factors, it necessarily involves legal determinations. Consequently, even if we defer to lower courts’ application of the factors, we cannot do so when they commit legal errors. See, e.g., United States v. Soussi 316 F.3d 1095, 1108 (10th Cir. 2002) (“A district court abuses its discretion if it makes an error of law.”). Thus, a district court may abuse its discretion when, for example, it gives one statutory factor too much weight, see, e.g., United States v. Cage, 451 F.3d 585, 595 (10th Cir. 2006), or expressly disregards another, see, e.g., United States v. Hildreth, 485 F.3d 1120, 1130 (10th Cir. 2007). That is, sentencing courts may not ignore applicable case law interpreting reasonableness under § 3553(a), and this Court may not ignore a district court’s errors in interpretation. Cooter & Gell, 496 U.S. at 402, 110 S.Ct. 2447 (noting that an abuse-of-discretion standard does not “preclude the appellate court’s correction of a district court’s legal errors”).

In short, in determining whether a sentence is reasonable, we defer to the district court’s exercise of discretion within the bounds of reasonableness: “[T]he fact that there is inevitably a range of sentences that could be held reasonable means that our affirmance of a sentence will necessarily defer, in effect, to the district court’s exercise of discretion in choosing a particular sentence within that range.” United States v. Sanchez-Juarez, 446 F.3d 1109, 1117 (10th Cir. 2006) (emphasis added). The fact that a range of reasonable sentences exists, beyond which we will not defer to a district court, makes clear that there are definite limits on a sentencing court’s exercise of discretion. See Rita, 127 S.Ct. at 2466-67 (“In sentencing, as in other areas, district judges at times make mistakes that are substantive.... Circuit courts exist to correct such mistakes when they occur.”). Our appellate review cannot merely rubber stamp the district court’s decision.

Rather than announcing a new or revised standard of review, Rita simply affirmed the appellate courts’ application of a “presumption” of reasonableness to within-Guidelines sentences on appellate review. Id. at 2462; see also id. at 2463 (“[T]he presumption reflects the fact that, by the time an appeals court is considering a within-Guidelines sentence on review, both the sentencing judge and the Sentencing Commission will have reached the same conclusion as to the proper sentence in the particular case.”). Indeed, the Court’s holding in Rita was simply that appellate courts may employ a presump*1137tion of reasonableness when reviewing proeedurally proper within-Guidelines sentences. The purpose of the Court’s reference in Rita to reasonableness review as an abuse-of-discretion standard was to emphasize that the presumption of reasonableness is an appellate presumption (i.e., sentencing courts may not apply it): “We repeat that the presumption before us is an appellate court presumption. Given our emphasis in Booker that appellate ‘reasonableness’ review merely asks whether the trial court abused its discretion, the presumption applies only on appellate review.” Id. at 2465. In other words, the Court noted a settled proposition (i.e., “that appellate ‘reasonableness’ review merely asks whether the trial court abused its discretion”) in order to emphasize a logical conclusion that may be inferred from it (i.e., that “the presumption applies only on appellate review”).

To turn the Court’s statement regarding the appellate nature of the presumption into a holding regarding the standard of review, the dissent takes the statement out of context and expands upon it by relying on Justice Stevens’s concurring opinion, in which Justice Ginsburg joined. In relying on Justice Stevens’s concurrence, the dissent misunderstands our duty to apply Supreme Court precedent. Justices Stevens and Ginsburg fully joined the majority opinion (not merely the judgment), and the concurrence garnered only two votes. Whatever compromises may have been required to bring Justices Stevens and Ginsburg on board the majority were made and are reflected in the text of the majority opinion. As the concurrence has no legal weight, we will not read it as a supplement to the Supreme Court’s clear majority opinion.

Finally, we note that, while the Rita Court also made clear that a sentence falling outside the Guidelines is not entitled to a “presumption of unreasonableness,” id. at 2467, the applicable Guidelines sentence and the policy statements of the Sentencing Commission nevertheless remain statutory factors that the district court must consider. See 18 U.S.C. § 3553(a)(4), (5). As the Supreme Court explicitly recognized in Booker, the sentencing factors contained in § 3553(a) delimit a district court’s discretion. 543 U.S. at 264, 125 S.Ct. 738 (disagreeing with the view that, after Booker, a sentencing court has the same discretion it possessed prior to the Sentencing Act). Because the Guidelines are a factor under § 3553(a)(4) and (5), a sentencing court abuses its discretion when it does not consider them. Booker, 543 U.S. at 264, 125 S.Ct. 738 (“The district courts, while not bound to apply the Guidelines, must consult those Guidelines and take them into account when sentencing.”); see also Cage, 451 F.3d at 594 (“Booker does not place original sentencing decisions entirely in the discretion of trial judges; the Guidelines— as an expression of the political will of Congress — continue to assert advisory influence on those decisions.”).

In recognizing the continued importance of the Guidelines, we do not treat them as nearly mandatory, as the dissent suggests. Indeed, we emphasize that a court may, in its discretion, conclude that a non-Guidelines sentence best serves the purposes of sentencing under § 3553(a). But if a court exercises its discretion to disregard, or give little weight to, the Guideline factors, § 3553(a)(4), (5), in crafting a sentence, it must find reasonable justification for doing so in the remaining § 3553(a) factors. A court’s conclusion that the Guidelines are simply “wrong” or an inadequate reflection of the statutory sentencing purposes is an unreasonable application of the § 3553(a) factors unless the court can justify the sentence imposed *1138in light of the facts of the particular case considered under § 3553(a).

B. Mr. Garciar-Lara’s Sentence

We now turn specifically to Mr. Garcia-Lara’s sentence. Applying the “career offender” provision of the Guidelines, U.S.S.G. § 4Bl.l(b)(A), and an adjustment for acceptance of responsibility, the District Court correctly calculated an advisory Guidelines range of 262 to 327 months’ imprisonment. The court then sentenced Mr. Garcia-Lara to 140 months’ imprisonment — a 47%, 122-month decrease from the 262-month Guidelines minimum.

We note at the outset that the problem here does not he in the manner in which the District Court set forth its reasoning concerning the sentence imposed. In other words, the sentence is reasonable in a procedural sense. See Cage, 451 F.3d at 591 (explaining “[rjeasonableness has both procedural and substantive components”). Indeed, the District Court correctly calculated the advisory Guidelines sentence and then noted the pertinent § 3553(a) factors it considered in varying from that advisory sentence. See United States v. Sanchez-Juarez, 446 F.3d 1109, 1117 (10th Cir. 2006) (holding district courts must “provide sufficient reasons to allow meaningful appellate review of their discretionary sentencing decisions”). The issue before us is therefore whether the District Court’s reasons for the sentence imposed support the conclusion that the sentence is substantively reasonable.

We have held that, “[t]he farther the [trial] court diverges from the advisory guideline range, the more compelling the reasons for the divergence must be.”1 Valtierra-Rojas, 468 F.3d at 1239 (quotation omitted) (alterations in original). Specifically, an “extreme” divergence is reasonable given “dramatic” facts. Cage, 451 F.3d at 594. A “substantial” variance requires “compelling reasons, although they need not be as dramatic as the reasons supporting an extreme divergence.” Hildreth, 485 F.3d at 1128 (quotations omitted). A lesser but still “significant” variance needs only “sufficient explanation and justification” to be within the range of reasonableness. United States v. Bishop, 469 F.3d 896, 908 (10th Cir. 2006).

To assess the magnitude of a variance, we look to the difference between the advisory Guidelines range and the sentence imposed in terms of both percentage and absolute number of months. Valtierra-Rojas, 468 F.3d at 1240. In terms of the absolute number of months, the 122-month divergence in this case more closely resembles variances we have characterized as “extreme,” see, e.g., United States v. Mateo, 471 F.3d 1162, 1170 (10th Cir. 2006) (99 months), than those we have character*1139ized as “substantial,” see, e.g., Hildreth, 485 F.3d at 1128 (27 months). The 47%-percent decrease, however, more closely resembles divergences we have characterized as “substantial” or “significant.” See id. at 1127-28 (reviewing and categorizing variance cases). But we need not decide whether the divergence is “extreme” or “substantial” because, even under the level of scrutiny we apply to substantial variances, the sentence imposed is not reasonable. As we explain below, the District Court did not justify the variance with “compelling reasons.”

The District Court based the 140-month sentence on its conclusion that the “defendant’s criminal history category and the overall impact of the career offender adjustment overstates the seriousness of [the] defendant’s criminal history and produces a sentence which is greater than necessary to accomplish the goals of 18 U.S.C. § 3553.” In reaching this conclusion, the court specifically noted that none of Mr. Garcia-Lara’s convictions were for crimes of violence, his prior drug convictions did not involve “large quantities” of drugs, his last drug conviction was at the age of 22 (Mr. Garcia-Lara was 30 years old at the time of sentencing), and his “longest term in prison has been about 2]é years.” Based on these facts, the court concluded that a sentence of 140 months was “proper and reasonable” to “reflect the seriousness of the offense, to provide just punishment, to protect the public from further crimes of the defendant, and to avoid unwarranted sentence disparities among defendants with similar records who have been found guilty of similar conduct.” See 18 U.S.C. § 3553(a)(2)(A), (2)(C), (6).

First, the fact that Mr. Garcia-Lara’s crimes were nonviolent and involved moderate quantities of drugs does not support the District Court’s conclusion that the career offender enhancement overstates his criminal history. The Guidelines do not condition application of the career offender provision on prior commission of crimes of violence; indeed, § 4B1.1 requires only “two prior felony convictions of either a crime of violence or a controlled substance offense.” U.S.S.G. § 4Bl.l(a) (emphasis added). Nor does the career offender provision set a threshold quantity requirement in order for a controlled substance conviction to qualify. Indeed, the purpose behind the career offender guideline is to carry out congressional intent to target specific recidivism, including repeat drug traffickers. See U.S.S.G. § 4B1.1 cmt. background.

As reflected in the PSR, there is no question that Mr. Garcia-Lara is a recidivist drug dealer. Since Mr. Garcia-Lara entered the United States from Mexico in 1990, he has spent most of his time in prison or on probation, largely for controlled substance offenses. The PSR lists arrests beginning in August 1990 for narcotics offenses and indicates that Mr. Garcia-Lara’s deportation was ordered in 1990. Shortly thereafter, in March 1991, he was convicted for selling cocaine. Apparently while Mr. Garcia-Lara was serving probation for this offense, he was arrested for the transport or sale of a controlled substance; was arrested, convicted, and spent ten days in jail for criminal trespass; and was arrested for the sale of cocaine. Less than a week after this last arrest for selling cocaine, Mr. Garcia-Lara’s probation for the March 1991 conviction was revoked. It appears that he then served over one year of his three-year prison term and was paroled in January 1994. Approximately two months after being released on parole, Mr. Garcia-Lara was arrested and subsequently convicted for possession of cocaine and was sentenced to two years in prison. After serving over one year in prison, he was paroled in May *11401995. Just short of two months after being released from prison, Mr. Garcia-Lara was arrested and subsequently convicted of possession of cocaine base for sale. He received a five-year prison sentence and served over two-and-a-half years in prison before being released on parole in June 1998 and subsequently deported. Less than two years after he was deported, Mr. Garcia-Lara was again arrested — this time for re-entry after deportation — and was sentenced to 48 months in custody followed by three years of supervised release. He served three-and-a-half years in prison and was released from custody in November 2003. Less than one year later, on November 20, 2004, he committed the instant offense involving 18 pounds of marijuana and over 500 grams of methamphetamine. For every separate criminal conviction, Mr. Garcia-Lara gave a different alias.

We recount Mr. Garcia-Lara’s criminal history in such detail to illustrate his demonstrated propensity to break the law and, in particular, to commit drug offenses. Thus, to the extent the District Court believed the career offender enhancement over-represented Mr. Garcia-Lara’s prior criminal history, it ignored Congress’s policy of targeting recidivist drug offenders for more severe punishment.2 Such an exercise of the District Court’s discretion may be reasonable if there are sufficiently compelling reasons, after considering the § 3553(a) factors, that justify the imposition of a non-Guidelines sentence. Here, the District Court has provided none.

In particular, the court’s decision not to apply the career offender guideline is not justified by “particular characteristics of the defendant” that are “sufficiently uncommon.” Mateo, 471 F.3d at 1169. Rather, Mr. Garcia-Lara’s criminal history satisfies the requirement for the guideline’s application, namely two prior controlled substance offenses. In addition to these offenses, the PSR notes several controlled substance arrests that did not result in conviction or for which the disposition is unknown. Further, the fact that Mr. Garcia-Lara’s last controlled substance offense occurred when he was 22 years old ignores the fact that he spent much of the intervening time in prison or, presumably, outside the country after he was deported. Similarly, even the fact that Mr. Garcia-Lara’s longest prison term was three-and-a-half years (not two and a half, as the District Court found) ignores the reality that he had several such sentences, many following just a few months after a prior release from prison. In short, Mr. Garcia-Lara’s characteristics and criminal history are not “out of the ordinary.”3 Cage, 451 F.3d at 596. As our case law makes clear, a sentencing court may not accord ordinary facts extraordinary weight. Id. at 595; see also Hildreth, 485 F.3d at 1129 (holding that sentence was unreasonable because the court did not distinguish the *1141defendant or his offense “from the ordinary defendant upon which the Guidelines sentence is calculated”). That is, a sentencing court may not ignore the Guidelines calculation for the ordinary defendant and instead adopt its own sentencing philosophy.4 See Hildreth, 485 F.3d at 1129.

Moreover, in the present case, because the District Court failed to distinguish Mr. Garcia-Lara from other career offenders, it incorrectly applied the § 3553(a) factor requiring consideration of “the need to avoid unwarranted sentencing disparities among [similar] defendants.”5 § 3553(a)(6). The District Court explained that the 140-month sentence did not create “disparity” among defendants with similar records because the sentence “would be within the Guidelines range of a defendant under the same facts who was not treated as a career offender.” But this logic depends on the court’s unreasonable determination that Mr. Garcia-Lara should not be treated as a career offender. The statute specifically requires that a sentencing court consider “the need to avoid unwarranted sentence disparities among defendants with similar records who have been found guilty of similar conduct.” § 3553(a)(6) (emphasis added). Because the court did not distinguish Mr. Garcia-Lara from the ordinary career offender, the court should have considered how the chosen sentence would further the goal of sentencing uniformity with respect to other career offenders with similar records convicted of the same offense. The court’s substantial divergence from the applicable Guidelines range clearly does not further that goal.

In sum, sentencing a career offender as if he did not have a career offender’s criminal record, absent sufficiently compelling circumstances, does not serve the statutory purposes noted by the District Court. The District Court’s 140-month sentence is therefore unreasonable.

III. CONCLUSION

For the foregoing reasons, we VACATE Mr. Garcia-Lara’s sentence and REMAND for resentencing.

dissenting.

Because the panel majority fails to accord the district court the deference to which it is entitled under an abuse of discretion standard, as announced in Unit*1142ed States v. Booker, 543 U.S. 220, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005), and fails to follow the letter and spirit of the Court’s mandate in Rita v. United States, — U.S. -, 127 S.Ct. 2456, 168 L.Ed.2d 203 (2007), I respectfully dissent.

In my opinion, the decision announced today stands as exhibit A on two points: First, it shows that notwithstanding the repeated reaffirmation and clarification of an appellate abuse of discretion standard by the majority and concurrence in Rita, the Newtonian pull of the Guidelines toward a near-mandatory center remains. Second, it demonstrates that the “trust that those Judges who had treated the Guidelines as virtually mandatory during the post-Booker interregnum will now recognize the Guidelines are truly advisory,” id. at 2474 (Stevens, J., concurring), is misplaced.

I

In Rita, the Court addressed whether the Courts of Appeals may apply a presumption of reasonableness to review within-Guidelines sentences, as employed by our Circuit and many others. See, e.g., United States v. Kristl, 437 F.3d 1050, 1053-54 (10th Cir. 2006) (adopting a presumption of reasonableness); United States v. Williams, 436 F.3d 706, 708 (6th Cir. 2006) (same); United States v. Mykytiuk, 415 F.3d 606, 608 (7th Cir. 2005) (same). In discussing the role of appellate courts in sentencing following Booker, the Court emphatically reaffirmed the proper standard of review applicable to sentencing challenges: We ask merely whether a district court has abused its discretion in applying the § 3553(a) factors.

While Booker itself implicitly equates reasonableness review of sentences with review for abuse of discretion, see 543 U.S. at 261-62, 125 S.Ct. 738, the Court made this equivalence explicit in Rita. “Given our explanation in Booker,” the Rita majority instructs, “appellate ‘reasonableness’ review merely asks whether the trial court abused its discretion.” Rita, 127 S.Ct. at 2465 (emphasis added). I disagree with the majority that in “mak[ing] that equivalence explicit,” the Court “says nothing new about the standard of review.” (Maj. Op. at 1136). Rather, the Court’s holding in Rita, while primarily addressed to the propriety of an appellate presumption of reasonableness when reviewing within-Guidelines sentences, nonetheless demands that we revisit our post -Booker sentencing jurisprudence.

Victor Rita argued that an appellate presumption of reasonableness for within-Guidelines sentences would impair the exercise of district court discretion and thus raise the same Sixth Amendment concerns that Booker sought to remedy — the use of judge-found facts to increase sentences beyond the range permitted by jury-found facts alone. See Rita, 127 S.Ct. at 2465. Justices Souter, Scalia, and Thomas espoused the view that the Sixth Amendment would indeed be violated either by the use of the presumption, id. at 2488 (Souter, J. dissenting), or by any form of substantive appellate review, id. at 2476 (Scalia, J. dissenting, joined by Thomas, J.).1 In rejecting Rita’s challenge and the *1143positions of those justices, the majority gave clear instruction that, notwithstanding an appellate presumption of reasonableness, district courts must have discretion to vary above or below the Guidelines range without being bound or even swayed by a similar presumption at sentencing. See id. at 2466 (“A nonbinding appellate presumption that a Guidelines sentence is reasonable does not require the sentencing judge to impose that sentence. Still less does it forbid the sentencing judge from imposing a sentence higher than the Guidelines provide for the jury-determined facts standing alone.”). Accordingly, circuit courts must review the individualized decisions of district courts with significant deference, allowing sentencing courts to exercise discretion in practice and not just in theory. As the Court implicitly acknowledges, imposing undue restrictions on the exercise of discretion would run the risk of undermining defendants’ Sixth Amendment right to a jury trial. See id. For this reason, among others, I read the Court’s statement that “appellate ‘reasonableness’ review merely asks whether the trial court abused its discretion,” id. at 2465, to require that we grant district courts greater discretion at sentencing than we have, in practice, provided over two eventful years of post-Booker review.

The panel’s discussion of Rita shrugs off the concurring opinion of Justices Stevens and Ginsburg, whose votes also underpin the majority holding.2 That concurrence traces the evolution of sentencing review, explaining in detail how the Court’s jurisprudence has evolved from de novo review to abuse of discretion, a standard that Booker “called ‘reasonableness’ review.” Id. at 2470-71 (Stevens, J., concurring). Justice Stevens notes that the bases for using an abuse of discretion standard-“a district judge [is] better positioned than an appellate judge to decide the issue,” “may have insights not conveyed by the record,” and may consider “special, narrow facts that utterly resist generalization” — apply in full force in the sentencing context. See id. at 2471-72 (Stevens, J., concurring) (citing Koon v. United States, 518 U.S. 81, 116 S.Ct. 2035, 135 L.Ed.2d 392 (1996); Cooter & Gell v. Hartmarx Corp., 496 U.S. 384, 403-05, 110 S.Ct. 2447, 110 L.Ed.2d 359 (1990); and Pierce v. Underwood, 487 U.S. 552, 558-560, 108 S.Ct. 2541, 101 L.Ed.2d 490 (1988)). By eliminating the de novo standard of review required under 18 U.S.C. § 3742(e), Booker “restored the abuse-of-discretion standard identified in [these] earlier cases.” Id. at 2471 (Stevens, J., concurring). Under that standard, “except to the extent specifically directed by statute, ‘it is not the role of an appellate court to substitute its judgment for that of the sentencing court as to the appropriateness of a particular sentence.’ ” Id. at 2472 (Stevens, J., concurring) (quoting Williams v. United States, 503 U.S. 193, 205, 112 S.Ct. 1112, 117 L.Ed.2d 341 (1992)) (emphasis added).

The concurrence accordingly instructs circuit courts: “It should also be clear that appellate courts must review sentences individually and deferentially whether they are inside the Guidelines range (and thus potentially subject to a formal ‘presumption’ of reasonableness) or outside that range.” Id. at 2474 (Stevens, J., concurring) (emphasis added). Yet, notwithstanding the direction of the Rita majority, *1144the clarification provided by the Stevens concurrence, and the fact that all three justices who declined to join Rita espouse an even more deferential view of appellate review than the majority, the panel dismissively insists on applying pre-i?iia case law in reversing the individualized and reasoned determination of the district court in this case.

II

The panel’s rejection of the district court’s particularized and reasoned determination is accomplished in part by a neat semantic tap dance over the nature of “reasonableness” review post-Booker. Contrary to the majority’s position, we have been anything but consistent about the amount of daylight, if any, between reasonableness and abuse of discretion review.3 For example, we frankly admitted in United States v. Rodriguez-Quintanilla that “the relationship between the abuse of discretion standard, the ‘plainly unreasonable’ standard and the post -Booker ‘reasonableness’ standard is less than crystal clear.” 442 F.3d 1254, 1258 (10th Cir. 2006). Rather than address this question head-on, we have typically held that our determination would be the same under either standard. See, e.g., United States v. Cordova, 461 F.3d 1184, 1188 (10th Cir. 2006) (“We need not explore the exact contours of our post-Booker standard of review, however, because we conclude that the District Court did not abuse its discretion and that the sentence was both procedurally and substantively reasonable.”).

Even if the majority were correct that we have, sotto voce, equated reasonableness review with review for abuse of discretion, it is the nature of our reasonableness review that must ultimately square with the Court’s mandate in Rita. And it is my view that the nature of our review has come off the rails by privileging certain elements of substantive reasonableness over others. Since the Court’s decision in Booker, we have woven a rich tapestry of case law assessing the propriety of district courts’ applications of various § 3553(a) factors to individual defendants. While some of that case law is new, much of it extends precedent that predates Booker. To take two examples, we have (1) reaffirmed our pre-Booker holding that, except in limited circumstances, co-defendant disparity does not warrant a variance under § 3553(6), see United States v. Davis, 437 F.3d 989, 997 (10th Cir. 2006), and (2) sanctioned district courts’ consideration of the characteristics of a defendant’s prior offense conduct (as well as his arrest record) under § 3553(a)(2)(B) and (C), see United States v. Mateo, 471 F.3d 1162, 1170 (10th Cir. 2006).

Armed with this precedent, we have then typically taken a divide and conquer approach on review, examining the district *1145court’s application of each relied-upon factor to the facts of the case. Correct as our legal constructions are, the net effect of this exercise is to require the district court to jump through a series of hoops, any one of which might require reversal as legal error — a reality which somewhat undermines the “sliding scale” approach we have taken to variance post-Booker. See Maj. Op. at 1138. Although we often plow through detailed factual histories related to the defendant’s background and offense conduct, we generally review for legal error-namely, the district court’s misapplication of a particular Guidelines factor. A typical example is our recent holding in United States v. Hildreth, 485 F.3d 1120 (10th Cir. 2007), in which we reversed the district court’s substantial downward variance due to its misapplication of the Guidelines factors. Although the district court in that case carefully considered the § 3553(a) factors singularly and in concert, and articulated its rationale for varying downward at some length, we held that “the District Court essentially ignored the recommendation of the sentencing Guidelines.” Id. at 1129. We then went through each individual justification in some detail, explaining why none sufficiently “distinguishfed the defendant] from defendants with similar histories convicted of the same crime” to support a variance. Id.

This type of appellate review is grounded in reason and serves a not ignoble purpose — the elimination, or at least mitigation, of sentencing disparity between similarly situated offenders. See Rita, 127 S.Ct. at 2467 (“Congress sought to diminish unwarranted sentencing disparity. It sought a Guidelines system that would bring about greater fairness in sentencing through increased uniformity.”); Booker, 543 U.S. at 253, 125 S.Ct. 738 (“This point is critically important. Congress’ basic goal in passing the Sentencing Act was to move the sentencing system in the direction of increased uniformity.”); United States v. Gonzalez-Huerta, 403 F.3d 727, 738 (10th Cir. 2005) (“[T]he purpose of the Guidelines was to promote uniformity in sentencing so as to prevent vastly divergent sentences for offenders with similar criminal histories and offenses. The federal courts have been striving towards this worthy goal since 1987.”). And so we have structured our substantive reasonableness review to hunt the white whale of disparity, a hunt which has required us to exercise a degree of scrutiny closer to de novo than abuse of discretion, to the exclusion of other elements that contribute to a substantively reasonable sentence.4

Notwithstanding its nobility of purpose, this type of review simply does not square with Rita’s mandate that we locate primary authority for sentencing at the retail level-the district court. See 127 S.Ct. at 2463 (“The upshot is that the sentencing statutes envision both the sentencing judge and Commission as carrying out the same basic § 3553(a) objectives, the one, at retail, the other at wholesale.”). More importantly, our focus on disparity has occasioned serious neglect of the procedural values of locating authority with the district court, values which were important to the Booker Court and have been reaffirmed in Rita. The Rita majority reemphasized the worth of the very process of *1146adversarial testing that is accomplished by the district court’s consideration of the presentence report, the arguments of the parties, and the § 3553(a) factors. See id. at 2465. There is independent value not just in the district court’s ability to consider sentencing facts itself, and to hear the arguments of the parties, but also to visibly play the role of setting a sentence within the statutory range and to state its reasons to the defendant face to face. See id. at 2468 (“Judicial decisions are reasoned decisions. Confidence in a judge’s use of reason underlies the public’s trust in the judicial institution. A public statement of those reasons helps provide the public with the assurance that creates that trust.”), 2469 (“By articulating reasons, even if brief, the sentencing judge not only assures reviewing courts (and the public) that the sentencing process is a reasoned process but also helps that process evolve.”).

Justices Stevens and Ginsburg firmly defend the primacy of this individualized, retail-level process:

While reviewing courts may presume that a sentence within the advisory Guidelines is reasonable, appellate judges must still always defer to the sentencing judge’s individualized sentencing determination. As we stated in Koon, ‘[i]t has been uniform and constant in the federal judicial tradition for the sentencing judge to consider every convicted person as an individual and every case as a unique study in the human failings that sometimes mitigate, sometimes magnify, the crime and the punishment to ensue.’

Id. at 2472-73 (Stevens, J., concurring) (citation omitted). This more process-minded view of federal sentencing, sensitive to the interplay between the trial judge, jury, prosecutor, and defendant, and skeptical of efforts to undermine that interplay by either legislative or appellate interference, recognizes that allocution plays an inherent role in sentencing and that the district court is not a mere abacus. Until this very opinion, we have not equated reasonableness with abuse of discretion, and the present semantic effort to do so is but a neat elision of the trial court’s Rita discretion. Our present practice substitutes that discretion with a supposed mathematical certitude that rejects trial court judgment regarding proportionality and repentance, thereby accommodating neither trial court discretion nor Rita.

My view is that we have lost sight of one of the broader instructions of Booker. While applying Booker's, mandate that the Guidelines henceforth play only an (important) advisory role in sentencing, we have paid too little attention to its implication that the benefits of placing primary authority for sentencing in district courts extend beyond their traditional role in fact-finding. For example, in this case the majority takes issue with the district court’s application of the § 3553(a) factors, but accords no weight to the thoroughness of its consideration. This lack of attention to the district court’s satisfaction of its procedural duties underweights the process values discussed supra.

Read in its entirety, the district court explained its choice of sentence in an expansive and well-reasoned statement:

The court believes that defendant’s criminal history category and the overall impact of the career offender adjustment overstates the seriousness of defendant’s criminal history and produces a sentence which is greater than necessary to accomplish the goals of 18 U.S.C. § 3553. The career offender adjustment places the defendant at a Category VI in criminal history. But, defendant has no conviction for a crime of violence, there is no indication that his prior convictions involved large quantities of *1147drugs, his last drug conviction was at the age of 22, and his longest term in prison has been about 2/á years. Although defendant has a serious criminal history, it does not merit a Category VI, and defendant’s next sentence should not escalate to 22 years.

Courts and commentators have noted that the career offender provisions of the Guidelines sometimes lead to extraordinary and inappropriate increases in sentences. See U.S. v. Phelps, 366 F.Supp.2d 580, 590 (E.D.Tenn. 2005); U.S. v. Woodley, 844 F.Supp.2d 274, 277 (D.Mass. 2004). The Guidelines, themselves, encourage departures where the criminal history category significantly over-represents the seriousness of a defendant’s criminal history or the likelihood that the defendant will commit future crimes. U.S.S.G. § 4A1.3(b)(l).

We believe this case is one where a departure under the Guidelines would be justified under the pre-Booker system or a non-Guidelines sentence is justified under the post-Booker system to produce a sentence less than the Guidelines range in this case....

We have examined the nature and circumstances of the offense and the history and characteristics of defendant. We find that the proposed guideline sentence of 262 months or more is greater than necessary to afford adequate deterrence and promote respect for the law. In order to reflect the seriousness of the offense, to provide just punishment, to protect the public from further crimes of the defendant, and to avoid unwarranted sentence disparities among defendants with similar records who have been found guilty of similar conduct, the court finds that a sentence of 140 months is proper and reasonable and no greater than necessary to comply with the purposes of 18 U.S.C. § 3553. We note with regard to the issue of disparity among defendants with similar records that a sentence of 140 months would be within the Guidelines range of a defendant under the same facts who was not treated as a career offender under the Guidelines.

Crediting the court’s finding that a sentence longer than set would be greater than necessary to accomplish the goals of § 3553(a), I fail to understand how the majority can insist that the defendant be resentenced. By “resentencing,” the majority assuredly does not mean a shorter or equal sentence. But in insisting on a longer sentence, we necessarily nullify the trial court’s § 3553(a) finding. I am hard-pressed to square today’s decision with our duty to treat the Guidelines as “truly advisory.” In particular, the panel’s treatment of the career offender provision of the Guidelines leaves little room for district courts to vary downward, and effectively, if not intentionally, treats the Guidelines as mandatory.

Although the reasons given by the district court for its variance-that Gareia-Lara “has no conviction for a crime of violence, there is no indication that his prior convictions involved large quantities of drugs, his last drug conviction was at the age of 22, and his longest term in prison has been about 2% years” — would not warrant a departure under the Guidelines, they reflect the history and characteristics of the defendant and are properly considered under § 3553(a)(1). See Rita, 127 S.Ct. at 2473 (Stevens, J., concurring) (listing numerous examples of factors “not ordinarily considered under the Guidelines” that “ § 3553(a) authorizes the sentencing judge to consider”).

In my opinion, the district court exercised its discretion precisely in the manner directed by Rita: It “beg[a]n by considering the presentence report and its interpretation of the Guidelines,” “subjected] the defendant’s sentence to the thorough *1148adversarial testing contemplated by federal sentencing procedure,” and selected a reasonable sentence without “enjoy[ing] the benefit of a legal presumption that the Guidelines sentence should apply.” Id. at 2465. Garcia-Lara’s 140-month sentence reasonably addresses the need to protect the public and rehabilitate the defendant and reflects the seriousness of his crime. § 3553(a). On those findings, I cannot say that the trial court’s selection of an eleven and one half year sentence was an abuse of discretion.

Ill

Instead of according substantial weight to evidence that the district court’s sentencing determination was thorough and reasoned, we have instituted a baroquely complicated “sliding scale” approach to substantive reasonableness review, which subjects the district court’s justifications to ever higher hurdles depending on the degree of the variance. Born in this court’s holding in United States v. Cage, 451 F.3d 585, 594 (10th Cir. 2006), that involved an imposed sentence of 6 days compared to an advisory Guidelines minimum of 46 months, the “sliding scale” has now evolved by such repeated rococo refinement of species as to do Charles Darwin and Louis XIV proud. See Hildreth, 485 F.3d at 1127-28 (detailing the standards applied to different degrees of variance). It seems that every problem has led to a new mutation with different, sometimes conflicting, rhetoric and outcome resulting in a sliding scale that is primarily notable for its preference for scaling and aversion to sliding. By the majority’s auto de fe, this system is preserved, rendering the Guidelines virtually mandatory. Thus, we have assumed a role in which we imper-missibly substitute our judgment as to the appropriateness of a particular sentence for that of the sentencing court.

Not only does this system impermissibly intrude upon the authority of the district court, but it also produces inconsistent results. Our case law dictates that we treat Garcia-Lara’s 47%, 122-month variance as “substantial” and uphold it if the facts of his case provide compelling justification. See Hildreth, 485 F.3d at 1128. When assessing the magnitude of a variance, we look to the discrepancy between the sentence imposed and the advisory Guidelines range in both percentage terms and absolute number of months. United States v. Valtierra-Rojas, 468 F.3d 1235, 1240 (10th Cir. 2006). Although “there is no formula into which we input the degree of divergence,” “comparison with other cases is a useful tool.” Id. In United States v. Allen, 488 F.3d 1244, 1253 (10th Cir. 2007), we required only “compelling justification” for a 167%, 225-month upward variance.5 Accordingly, under our pre-Rita jurisprudence, I would consider Garcia-Lara’s 47%, 122-month downward variance as “substantial,” requiring no more than a “compelling justification.”

In United States v. Mateo, we approved a district court’s use of the armed career criminal enhancement as a “guidepost” in sentencing a defendant who did not meet the criteria for that enhancement. 471 F.3d 1162, 1170 (10th Cir. 2006); see also U.S.S.G. § 4B1.4. Although Mateo’s PSR properly calculated a Guidelines range of 15 to 21 months’ imprisonment, the district court determined that this range underrepresented Mateo’s criminal history. 471 F.3d at 1166. The court referred to the *1149armed career criminal enhancement to account for the discrepancy, and imposed a sentence of 120 months’ incarceration.6 In reviewing this variance, we held that “whether the rationale provided by the sentencing court for a non-Guidelines sentence is sufficiently compelling is determined by considering whether the particular characteristics of the defendant the court relied upon in fashioning the sentence are commonplace ... or are sufficiently uncommon to justify a divergence from the presumptively reasonable Guidelines sentence.” Id. at 1169. Although we termed the 471% increase in Mateo’s sentence “extreme,” we concluded that his “significant contact with the criminal justice systems in three different states over a relatively short period of time” were dramatic facts that justified the variance. Id.7 “Based on [Mateo’s] specific circumstances and the District Court’s use of the armed career criminal provision as a guidepost to gauge the length of the sentence, the District Court imposed a reasonable sentence.” Id.

Garcia-Lara’s case presents a comparable scenario. Here, the district court properly calculated a Guidelines range that reflected the career offender enhancement. Having concluded that such a sentence would over-represent Garcia-Lara’s criminal history, considering the § 3553(a) factors, the court imposed a lesser sentence after again consulting the Guidelines to determine what Garcia-Lara’s advisory range would be absent the career criminal enhancement. The court’s reasons were not as “dramatic” as those in Mateo, but under our pre-Rita jurisprudence we do not require “dramatic facts” when the vari-anee at issue is not “extreme.” Cf. Mateo, 471 F.3d at 1170. Instead, Garcia-Lara’s substantial variance represents a proper exercise of district court discretion if the individualized factors of his case — with reference to § 3553(a) — provide compelling reasons that distinguish him from other career offenders.

As the district court persuasively explained, Garcia-Lara’s criminal history meets the bare minimum for application of the career offender enhancement; he has only two prior drug-related convictions, neither of which resulted in long prison terms or were shown to involve large quantities of drugs. Given these facts, and Garcia-Lara’s relatively young age at the time he committed his prior drug crimes, the district court’s decision to vary downward to offset the impact of the career offender enhancement was within the court’s discretion. Like the sentencing court in Mateo, the court carefully considered the defendant’s specific circumstances and used an alternative Guidelines range as a “guidepost” to gauge the length of a reasonable sentence.

In imposing Garcia-Lara’s sentence, the court did not show “express disregard” for § 3553(a)(6), which requires a sentencing court to consider “the need to avoid unwarranted sentencing disparities among [similar] defendants.” Cf. Hildreth, 485 F.3d at 1130. Instead, the court sentenced Garcia-Lara only after reasoning “that a sentence of 140 months would be within the Guidelines range of a defendant under the same facts who was not treated as a career offender under the Guidelines.” In faulting the district court for having incor*1150rectly applied § 3553(a)(6), Maj. Op. 1140, the panel majority fails to credit our prior case law on point.

In United States v. Shaw, 471 F.3d 1136, 1141 (10th Cir. 2006), we held:

The district court concluded that Shaw’s conduct was more serious than his code-fendant’s because Shaw was the one who actually assaulted the bank manager. See § 3553(a)(1) (“the nature and circumstances of the offense”). The Guidelines do not explicitly distinguish between principals and accessories for purposes of the “serious bodily injury” enhancement at issue here. USSG § lB1.3(a)(l)(A). But § 3553(a)(2)(A) does authorize a sentencing court to impose a nonguideline sentence if the court concludes the guideline range does not adequately “reflect the seriousness of the offense.” While an adjustment based on a factor that was already built into the guideline calculation may challenge the overall uniformity of sentences under § 3556(a)(6), any tension between subsection (a)(2)(A) and subsection (a)(6) can be resolved by the district court in light of all the facts before it, as long as it does so reasonably. See Cage, 451 F.3d at 595 (“The problem with the sentencing decision, however, is not in the consideration of these factors; it is in the weight the district court placed on them.”).

Shaw’s reasoning applies equally to this case. Section 3553(a)(1) permits a sentencing court to impose a non-Guidelines sentence if the Guidelines range does not adequately reflect “the history and characteristics of the defendant.” The court’s sentencing decision, made after careful consideration of the facts before it, reasonably resolved any tension between subsections (a)(1) and (a)(6). And, as the district court pointed out, “[t]he Guidelines, themselves, encourage departures where the criminal history category significantly over-represents the seriousness of a defendant’s criminal history or the likelihood that the defendant will commit future crimes.”

We recently recognized, “In any given case there could be a range of reasonable sentences that includes sentences both within and outside the Guidelines range- [C]ourt[s] may impose a non-Guidelines sentence if the sentencing factors set forth in § 3553(a) warrant it, even if a Guidelines sentence might also be reasonable.” United States v. Begay, 470 F.3d 964, 975-76 (10th Cir. 2006). Accordingly, I And ample support for the district court’s variance even under our pre-Rita case law. This is not to say that there may not also be support in our case law for a reversal on these facts, as argued by the majority, but to emphasize that our existing precedent neither gives clear direction to district courts nor allows the discretion afforded to them by Rita.

IV

The United States Supreme Court unambiguously pronounces the advisory nature of the Guidelines and reemphasizes district court discretion and deferential review solely for abuse of discretion. Given today’s rejection of this clear mandate, it is evident that if the Court is to move the universe of sentencing review away from the status quo of micro-management at the appellate level, the Court is going to need a longer Archimedean lever than Rita to accomplish the task.