SUTTON, J., delivered the opinion of the court, in which JORDAN, D. J., joined. SILER, J. (p. 359-60), delivered a separate dissenting opinion.
OPINION
SUTTON, Circuit Judge.
Avery Poynter, 36 years old, pleaded guilty to traveling in interstate commerce for the purpose of engaging in illicit sexual conduct with two minors. After calculating a guidelines range of 188-235 months and considering the § 3553(a) factors, the [350]*350district court imposed a 720-month sentence (the statutory maximum) because Poynter was a repeat child sex offender. Unable to conclude that this variance resulted from a reasonable application of § 3553, we reverse.
I.
On May 4, 1989, Avery Poynter pleaded guilty in Kentucky state court to committing four counts of sodomy in the second degree with an eleven-year-old male. The court sentenced Poynter to a 20-year term of imprisonment, but the State released him on parole four years later after he completed Kentucky’s sex offender treatment program in prison.
In October 2003, Poynter traveled from Kentucky to Tennessee to have sex with a fourteen-year-old male. In December, he met the same minor in Indiana and, after the two traveled to Kentucky, they again had a sexual encounter. In April 2004, Poynter traveled to Indiana with another fourteen-year-old male for another sexual foray. The next month, the two traveled to Florida for the same purpose.
After the police caught Poynter, he pleaded guilty to four counts of traveling in interstate commerce for the purpose of engaging in illicit sexual conduct with a minor. 18 U.S.C. § 2423(b). Accounting for adjustments based on his supervisory control over the victims, U.S.S.G. § 2A3.2(b)(2)(B), and on his multiple counts of conviction, id. § 3D1.4, the district court set Poynter’s initial offense level at 28 and his criminal history category at III, creating a guidelines range of 97-121 months. Because Poynter was convicted of a sex crime and had previously “sustain[ed] at least one sex offense conviction,” the district court applied the mandatory minimums of the “Repeat and Dangerous Sex Offender Against Minors” sentencing guideline. U.S.S.G. § 4B1.5(a). The district court therefore increased Poynter’s final offense level to 32 (including a two-level adjustment for acceptance of responsibility), see id. § 4B1.5(a)(l)(B)(ii), and increased his criminal history to category V, see id. § 4B1.5(a)(2). All of these adjustments considered, Poynter was left with a guidelines range of 188-235 months.
At the sentencing hearing on August 18, 2005, the district court heard testimony from Poynter’s victims and from the mother of the second victim. The district court acknowledged receiving a letter from Poynter’s parents, and Poynter spoke on his own behalf. Poynter “apologize[d] to [his] victims, as well as [his] family”; he recognized that “this has been a problem for [him] for many years” and that he had already gone “back into counseling with the counselors”; he welcomed “anything that [the] Federal [Government] has to offer”; and he said that he did not “intend” to “ever hav[e] another victim.” JA 41-42.
In sentencing Poynter, the district court acknowledged the guidelines range of 188-235 months and the statutory maximum of 720 months. See 18 U.S.C. §§ 2423(b), 2426(a). The court recognized that it must “impose a sentence ... that is sufficient but not greater than necessary to comply with the purposes” of 18 U.S.C. § 3553(a)(2). JA 46. It then analyzed the § 3553(a) factors in exercising its independent judgment about what an appropriate sentence would be. Reflecting on the “seriousness of the offense,” see § 3553(a)(2)(A), the court said that “[t]he protection of children in our society deserves the highest priority,” JA 46, and observed that Poynter “used alcohol and other drugs to seduce these victims” and that he had “victimized multiple children,” JA 49; see § 3553(a)(1). Reflecting on the need to impose a “just punishment” and [351]*351“to promote respect for the - law,” see § 3553(a)(2)(A), the court focused on Poyn-ter’s criminal history: “You did do it once before.... You did not learn your lesson. And so I think that the just punishment component of this has been ratcheted upwards.” JA 46; see § 3553(a)(1). Consistent with the need to “protect the public from further crimes,” see § 3553(a)(2)(C), the district court stated that “[t]he only way I can be sure that [children] are protected is for [Poynter] to be in custody.” JA47.
As for rehabilitation, see § 3553(a)(2)(D), the court said that Poynter should participate in the federal sex offender treatment program but doubted that Poynter would ever recover because sex crimes are “highly recidivistic.” JA 49. The court also noted that the “statutory maximum” protected against “unwarranted sentencing disparities,” JA 47; see § 3553(a)(6), and that “[t]here is no restitution that [Poyn-ter] can pay to these victims that can make them whole,” JA 47; see § 3553(a)(7).
In choosing the statutory maximum (720 months or 60 years), the court emphasized the need for “adequate deterrence,” see § 3553(a)(2)(B), explaining that sex crimes are “highly recidivistic,” that Poynter “forgot the lessons [he] learned in counseling,” that “[o]ne slip is one too many” and that the harm here is so great that it “outweighs in the Court’s judgment [Poynter’s] ten-year compliance.” JA 48. The court added that it was “aware of the guideline range top — capping out at 235 months,” see § 3553(a)(4) — but that it had “chosen the statutory maximum because [it] believe[d] that Mr. Poynter is unsafe to children” as a repeat offender. JA 52. - “It is my intention,” the court concluded, “that he either not ever be released from prison or be released from prison at such an age that it would be highly unlikely that he will ever commit this offense again. If he can’t control himself, I think the Court needs to impose a sentence that will control him.” JA 52-53. The court also imposed a lifetime term of supervised release.
II.
Since United States v. Booker, 543 U.S. 220, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005), we have distinguished between procedural and substantive reasonableness. See United States v. Webb, 403 F.3d 373, 383 (6th Cir.2005). No one questions the district court’s adherence to the procedural requirements of post-Booker review: It properly calculated the guidelines range, recognized that those guidelines were advisory and thoroughly considered the factors listed in § 3553(a). See United States v. Davis, 458 F.3d 491, 495 (6th Cir.2006). The court also stated “the specific reason for the imposition of a sentence” outside the guidelines range, 18 U.S.C. § 3553(c)(2) — “that [Poynter] either not ever be released from prison or be released from prison at such an age that it would be highly,unlikely that he will ever commit this offense again.” JA 52.
The substantive reasonableness of this sentence is another matter — a point that Poynter not only contests but contests with some force. In asking whether a sentence reasonably comports with the substantive demands of § 3553, we face a tug of war between two competing sets of considerations. On one side, we face several limitations in second-guessing the district court’s decision. Unlike the trial court, we did not see the defendant, the victims or family members testify at the sentencing hearing. And. unlike the trial court, most appellate judges have little experience sentencing individuals. While trial judges sentence individuals face to face for a living, we review transcripts for a living. No one sentences transcripts. All of this suggests that we should ac[352]*352knowledge the trial courts’ comparative advantages — its ring-side perspective on the sentencing hearing and its experience over time in sentencing other individuals— and give considerable deference to their sentencing decisions. See Rita v. United States, — U.S. —, 127 S.Ct. 2456, 2469, 168 L.Ed.2d 203 (2007) (“The sentencing judge has access to, and greater familiarity with, the individual ease and the individual defendant before him than the Commission or the appeals court.”); Koon v. United States, 518 U.S. 81, 98, 116 S.Ct. 2035, 135 L.Ed.2d 392 (1996) (“District courts have an institutional advantage over appellate courts in making [sentencing] determinations, especially as they see so many more Guidelines cases than appellate courts do.”).
On the other side, one of the principal functions of the Sentencing Reform Act of 1984 was to eliminate “unwarranted sentencing] disparities among defendants with similar records who have been found guilty of similar conduct.” 18 U.S.C. § 3553(a)(6). District court judges cannot correct that problem within their circuit or even within their own court (so long as two or more judges sit there), much less nationwide, because “different judges (and others) can differ as to how best to reconcile the disparate ends of punishment.” Rita, 127 S.Ct. at 2464. Reasonableness review thus permits “appellate courts to minimize sentencing disparities between and among district courts (and between and among courts of appeals),” Davis, 458 F.3d at 495, or, to use the words of Booker, to “iron out sentencing differences” among trial judges, Booker, 543 U.S. at 263, 125 S.Ct. 738; see also United States v. Sriram, 482 F.3d 956, 962 (7th Cir.2007) (“[Booker] did not authorize sentencing judges to pick any sentence within the applicable statutory sentencing range that strikes their fancy.”).
In trying to account for these competing considerations and in trying to identify principled reasons for upholding some sentences but not others, our circuit has applied a proportionality principle based on at least two of the § 3553(a) factors — the guidelines range, § 3553(a)(4), and the need to avoid unwarranted sentencing disparities, § 3553(a)(6). “[T]he farther the judge’s sentence departs from the guidelines sentence,” we have said, “the more compelling the justification based on factors in section 3553(a) must be.” Davis, 458 F.3d at 496 (internal quotation marks omitted). Many other circuits have done the same or something nearly the same. See United States v. Crisp, 454 F.3d 1285, 1291 (11th Cir.2006); United States v. Cage, 451 F.3d 585, 594 (10th Cir.2006); United States v. Smith, 445 F.3d 1, 4 (1st Cir.2006); United States v. Duhon, 440 F.3d 711, 715 (5th Cir.2006); United States v. Lazenby, 439 F.3d 928, 932 (8th Cir.2006); United States v. Moreland, 437 F.3d 424, 434 (4th Cir.2006); United States v. Dean, 414 F.3d 725, 729 (7th Cir.2005); cf. United States v. King, 454 F.3d 187, 195 (3d Cir.2006); United States v. Simpson, 430 F.3d 1177, 1187 n. 10 (D.C.Cir.2005).
We have applied this principle in several cases over the last year — both to downward and upward variances. Compare, e.g., United States v. Kathman, 490 F.3d 520, 526 (6th Cir.2007) (upholding 61% downward variance based on assessment that defendant was “a law abiding fine young man” who merely used “poor judgment” one evening and was less culpable than similar defendants) (internal quotation marks omitted); United States v. Cherry, 487 F.3d 366, 370-72 (6th Cir.2007) (upholding 43% downward variance because defendant had a “low risk for reoffending,” had taken “extensive” efforts to rehabilitate himself and was not “a run-of-the-mill offender”); United States v. [353]*353Husein, 478 F.3d 318, 333 (6th Cir.2007) (upholding 99.91% downward variance as reasonable in light of defendant’s “extraordinary family circumstances”); United States v. Wells, 473 F.3d 640, 644 (6th Cir.2007) (upholding 14% upward variance based on defendant’s “egregious actions” and “lengthy history of escalating violent behavior”); United States v. Collington, 461 F.3d 805, 809-10 (6th Cir.2006) (upholding 36% downward variance as reasonable in light of defendant’s peculiar criminal history and unique family circumstances); United States v. Fuson, 215 Fed.Appx. 468, 475 (6th Cir.2007) (upholding sentence of probation rather than two years’ imprisonment because felon-in-possession charge arose from incidental possession of an antique handgun); United States v. Williams, 214 Fed.Appx. 552, 556 (6th Cir.2007) (upholding 177% upward variance because sentencing range underrepresented defendant’s “continuing propensity for violence”), with United States v. Borho, 485 F.3d 904, 916 (6th Cir.2007) (vacating 66% downward variance because “the extreme nature of the deviation, without a correspondingly compelling justification, resulted in a substantively unreasonable sentence”) (emphasis omitted); United States v. Funk, 477 F.3d 421, 431 (6th Cir.2007) (vacating 43% downward variance when district court unduly discounted defendant’s criminal history); Davis, 458 F.3d at 499-500 (vacating 99.89% downward variance because “most extreme variance” “leavfes] no room to make reasoned distinctions” between defendant and other “more worthy defendants”).
Gauged by this proportionality principle and by our application of it in these cases, this 60-year sentence, a 206% upward variance from the top of the guidelines range, cannot be sustained. The primary ground for the variance — that the court wanted to prevent Poynter from ever committing this crime again — does not distinguish Poynter from other repeat sex offenders. True enough: “Poynter is unsafe to children,” JA 52; “[o]ne slip is one too many,” JA 48; his crime is “highly recidivistic,” id.; and “[i]f [Poynter] can’t control himself, ... the Court needs to impose a sentence that will control him,” JA 52-53. But, at this level of generality, the same could be said of all repeat sex offenders.
No less importantly, the Sentencing Commission promulgated the “Repeat and Dangerous Sex Offender Against Minors” guideline precisely to address the problem of recidivist sex offenders. Designed to deal with individuals who previously have sustained “at least one sex offense conviction,” U.S.S.G. § 4B1.5(a), and who have exhibited a “pattern of activity involving prohibited sexual conduct,” id. § 4B 1.5(b), the guideline recommends “lengthy incarceration for offenders who commit sex offenses against minors and who present a continuing danger to the public.” Id. § 4B1.5 cmt. background. The guideline protects the public from this “continuing danger” by calling for a “lengthy incarceration” for these individuals — one that significantly bumps up the sentencing range for any sex offender who has committed at least one sex offense before.
In view of his offense and of this guideline, Poynter faced the following enhancements. The underlying crime, together with other adjustments for criminal history, acceptance of responsibility and the like, called for an advisory sentencing range of 97-121 months. The “Repeat and Dangerous Sex Offender Against Minors” guideline increased that range to 188-235 months. And the district court sentenced Poynter to 720 months. What started as a top-end range of roughly 10 years became a top-end range of roughly 20 years and ended as' a 60-year sentence — all because [354]*354he had been convicted once before as a sex offender and all of this even though the increase in the advisory range from roughly 10 to 20 years was meant to account for the problem of recidivism. Absent some explanation for why Poynter’s risk of recidivism differs from that of other repeat sex offenders, we are hard-pressed to understand why Poynter, but apparently not other repeat offenders, needs what amounts to a life sentence — Poynter will be 96 when (or if) he is released from prison.
The other aspects of Poynter’s crimes mentioned by the district court — his multiple counts of conviction, his relationship of trust with the victims, the fact that prior treatment did not prevent these crimes— may well justify an upward variance (even a significant upward variance) but they do not support a 60-year sentence. In explaining why it sentenced Poynter at the top of the statutory range, the district court did not rely on these factors, and it did not explain why a sentence of 60 years was “no greater than necessary” despite the Sentencing Commission’s considered judgment to the contrary — a judgment the government does not contend is generally “unsound” as to repeat offenders, cf. Rita, 127 S.Ct. at 2468.
By relying on a problem common to all repeat sex offenders (recidivism) in increasing Poynter’s sentence and by failing to offer meaningful distinctions between the risk that Poynter posed to the public and the risk that other sex offenders posed to the public, the district court left us little room to distinguish between Poynter and other sex offenders. See Husein, 478 F.3d at 334 (affirming substantial variance and noting that “more worthy defendants than Husein are difficult to imagine, short of those found to be not guilty”) (internal quotation marks omitted); accord Davis, 458 F.3d at 499; United States v. Goldberg, 491 F.3d 668, 673-74 (7th Cir.2007); United States v. Moreland, 437 F.3d 424, 437 (4th Cir.2006); United States v. Haack, 403 F.3d 997, 1005-06 (8th Cir.2005). As utterly depraved as this crime is and as forever scarring as it must be to be victimized by it, not all repeat sex offenders deserve what amounts to a life sentence; otherwise, Congress would not have set a statutory range of 0-60 years. See 18 U.S.C. §§ 2423(b), 2426(a). While there is no yardstick perfectly calibrated to measure one crime and one criminal from the next crime and the next criminal, there are certainly measurable differences between Poynter’s situation and the situation of offenders who might warrant the statutory maximum or something approaching it. Consider those who, unlike Poynter, flee the authorities and seek to obstruct their own conviction; those who do not accept responsibility for their actions; those who employ violence or use weapons to commit the offense; and those who stand convicted of a long list of prior sex offenses with children rather than just one prior conviction. Cf. United States v. Bridgewater, 479 F.3d 439, 440-42 (6th Cir.2007) (affirming 10-year, statutory-maximum sentence for possessing child pornography in part because defendant was more culpable and dangerous than the average offender: the defendant took “photographs of [himself] molesting young girls who were in [his] care” while running a home for abused and neglected children; he successfully concealed these offenses and his criminal past from others “despite [his] continued proximity to youth in church programs”; and his own son “condemned” him and “questioned his remorse and sincerity” in a letter to the court).
Neither can we agree with the district court that its invocation of the 60-year statutory maximum by itself answers § 3553(a)’s concern about “avoiding] unwarranted sentence disparities.” 18 [355]*355U.S.C. § 3553(a)(6); see JA 47 (“I think that the statutory maximum ... take[s] care of [avoiding unwarranted sentence disparities].”). While the statutory maximum eliminates any risk of disparity caused by sentences above 60 years, that maximum does little to diminish disparities for individuals who receive (or should receive) sentences below what amounts to a life sentence for most individuals convicted of this crime. If Poynter’s sentence represents a reasonable application of § 3553(a), it is difficult to understand why any sentence between 188 and 720 months, a nearly 45-year span, would not be reasonable as well. But giving district courts such a sweeping range of discretion does not seem to have been the goal of Booker, much less the continuing goal of the Sentencing Commission. See Booker, 543 U.S. at 253, 125 S.Ct. 738 (“Congress’ basic goal in passing the Sentencing Act was to move the sentencing system in the direction of increased uniformity.”); 28 U.S.C. § 991(b)(1)(B) (One purpose of the Commission is to “avoid[] unwarranted sentencing disparities among defendants with similar records who have been found guilty of similar criminal condüct.”); id. § 994(f) (instructing the Commission that the guidelines should particularly attempt to “reduc[e] unwarranted sentence disparities”); see also 18 U.S.C. § 3553(a)(6).
While Booker empowered district courts, not appellate courts and not the Sentencing Commission, the only way to “avoid unwarranted sentence disparities” is for appellate courts to preserve reasoned distinctions among offenders. In attempting to satisfy the worthy goal of “be[ing] sure that [children] are protected” from sex offenses, we thus do not have available the solution adopted by the district court of simply placing the offender in permanent custody, JA 47, a solution that sweeps all repeat sex offenders within its net. Not only does the 0-60 year range suggest that Congress thought that courts should draw distinctions among sex offenders, but incarceration also is not the sole method Congress prescribed for protecting the public in this area. Life-time supervised release — authorized by Congress, 18 U.S.C. § 3583(k), recommended by the Sentencing Commission, U.S.S.G. § 5D1.2(b), and adopted by the district court, see JA 50 — remains an option. Supervised release permits the district court to prohibit Poynter from “associating unnecessarily with” children, 18 U.S.C. § 3563(b)(6), from volunteering as a youth leader, from becoming a teacher or engaging in any other profession that might place him in a position of trust with children, id. § 3563(b)(5), and from residing near any school, church or day-care center, id. § 3563(b)(13), among other things. See id. § 3583(d).
All of these considerations are for the district court to ponder in the first instance. It is not for us to impose a sentence. It is for us only to say that the justifications offered for this sentence do not support it and do not meaningfully distinguish Poynter from other repeat sex offenders. See Davis, 458 F.3d at 500.
That leaves one issue: Does the Supreme Court’s recent decision in Rita v. United States require us to abandon the proportionality principle that we have applied in this case? The answer would seem to be “no,” on the one hand, because Rita concerned the presumption of reasonableness for within-guidelines sentences, not the proportionality principle, and the Court will consider the validity of proportionality review during its next Term. See United States v. Gall, No. 06-7949. Yet, at the same time, the Court said two things in upholding the presumption of reasonableness that are in some tension with proportionality review: (1) there is no “presumption of unreasonableness” for [356]*356outside-guidelines sentences, Rita, 127 S.Ct. at 2467; and (2) the legitimacy of an appellate presumption of reasonableness for within-guidelines sentences turns in part on the “double determination” that “both the sentencing judge and the Sentencing Commission ... reached the same conclusion as to the proper sentence in the particular case,” id. at 2463.
If, as Rita says, there is no presumption of unreasonableness for outside-guidelines sentences, that suggests a district court has no less discretion to sentence within the guidelines than to sentence outside of them. And if that is so, why should a sentencing court have any greater duty to justify a within-guidelines sentence than an outside-guidelines sentence, much less to strengthen the explanation for its sentence the further the sentence varies from the guidelines’ recommendation? Likewise, if an appellate presumption of reasonableness for within-guidelines sentences turns at least in part on the alignment of the views of the Sentencing Commission with the views of the sentencing court, why should appellate courts give force to the Commission’s recommendations through proportionality review when no such “double determination” exists, when in other words the sentencing judge chose not to follow the Commission’s advice?
These are good questions, but they do not compel us to abandon the proportionality principle. First, that is not all Rita said, and the rest of its reasoning remains consistent with permitting courts to engage in some form of proportionality review. In upholding a presumption of reasonableness for within-guidelines sentences, the Court also reasoned that Congress directed the Sentencing Commission to develop sentencing recommendations based on the same § 3553(a) considerations that district courts must consider in sentencing an individual. Id. at 2463-64. And it explained that the guidelines’ recommendations are based on “empirical” data reflecting the aggregate experiences of sentencing judges across the country and the input of diverse members of the “law enforcement community.” Id. at 2464. If, in a given case, a sentencing court comes to a substantially (more on that adverb later) different conclusion from the Commission about an appropriate sentencing range, it seems quite reasonable to expect the court’s explanation to be commensurate with its variance from the Commission’s empirically grounded and expert advice about how to apply the same § 3553(a) factors to a given crime and a given criminal.
Second, it is not proportionality review, but a congressional directive, that initially requires district courts to treat the explanations they give for outside-guidelines sentences differently from the explanations they give for within-guidelines sentences. While Congress requires sentencing courts to give a statement of reasons for all sentences, it requires courts to give “the specific reason for the imposition of a sentence different from” the guidelines-recommended sentence, 18 U.S.C. § 3553(c)(2) (emphasis added). If it is appropriate to ask sentencing courts to give the “specific reason” for variances, but not within-guidelines sentences, it is a modest additional step — and certainly an extra step that will be useful to the Sentencing Commission in formulating future recommendations — to ask that those “specific reason[s]” have more force the further the court varies from the recommended range.
Third, two of the § 3553(a) provisions that Congress requires district courts to consider — (a)(4) and (a)(6) — suggest that courts must factor proportionality considerations into outside-guidelines sentences. [357]*357Subsection (a)(4) says a sentencing judge “shall consider” the Sentencing Commission’s advice about how to sentence certain types of criminals for certain types of crimes. When a court chooses not to follow that advice, the best way to indicate that the court considered the guidelines range is not merely to say so — “I considered the guidelines range” — but to show that is so by giving a “specific reason” that is commensurate with the degree of the variance. A sentencing system that requires judges to consider the guidelines range and to give a “specific reason” for deviating from that range, yet permits this reasoning to bear no relation to the size of the variance, is not one destined to respect the Commission’s recommendations in the long run.
Any lingering doubt about the point ought to be relieved by subsection (a)(6), which commands courts to “avoid unwarranted sentence disparities among defendants with similar records who have been found guilty of similar conduct.” What would an “unwarranted” sentencing disparity be if not a sentence lacking sufficient justification for its disparity from the sentences of other similarly situated defendants? How else , can a district court judge, or for that matter an appellate judge, tell when such disparities are occurring without consulting the guidelines range? And how else can judges avoid such disparities in the first instance, or correct them on review, without demanding that substantial variances be supported by substantial reasons? There are several questions here, and the practical answer to all of them is to allow courts to factor rough proportionality into the equation.
Fourth, proportionality considerations do not forsake the other § 3553(a) factors — most notably, the nature of the offense and of the defendant, id. § 3553(a)(1), and “the need for the sentence,” id. § 3553(a)(2)(A)-(D). These are precisely the factors that district courts will invoke in giving a “specific reason” for a variance. And in asking that this reasoning be particularly forceful when the district court imposes a particularly large variance, the proportionality principle merely respects the fact that Congress directed the Commission to consider these same factors in recommending the sentencing ranges it did. To permit district courts to' rely on these factors to vary sentences substantially without asking them to give commensurate explanations is' not to respect subsections (a)(1) and (a)(2) but to exalt them — and in the process to make the avoidance of unwarranted sentencing disparities all but impossible.
Fifth, this principle not only respects all of the § 3553(a) factors, but it also in the end is a modest principle, one that is unlikely to make a difference in most cases. It is unlikely to affect the review of sentences that vary slightly or even moderately from the Commission’s recommendations. By its terms, a proportionality principle would ask little in the way of explanation for a sentence that varies little from the guidelines. And the mandate of “avoid[ing] unwarranted sentence disparities,” § 3553(a)(6), will have little weight with modest or even moderate variances. The value of the principle, and the only time it would seem to have an effect on the validity of a sentence, is when the trial court varies substantially from the guidelines.
Sixth, with respect to substantial variances, the proportionality principle would seem to be an indispensable tool of appellate review — if for no other reason than that there is no other benchmark to measure the reasonableness of a sentence at the high or low ends of the sentencing range authorized by Congress. Where else would an appellate court start in gauging [358]*358the reasonableness of such a sentence? We know that the sentencing court must start with the probation department’s pre-sentence report as to an appropriate guidelines’ sentencing range, and the court must subject that report to the adversarial process. See Rita, 127 S.Ct. at 2465. We know that the sentencing court may grant downward or upward departures to individuals deserving of them. See U.S.S.G. § 5K2.0(a); Rita, 127 S.Ct. at 2461-62. And we know that, once the sentencing court accounts for all of the § 3553(a) factors and comes to a reasoned judgment about their application to a given individual, we must give the benefit of the doubt to the district court’s judgment — conducting reasonableness review that comes to nothing more than abuse-of-discretion review. Id. at 2465.
But reasonableness review still amounts to review, and without rough proportionality as an available guide it is difficult to imagine how to conduct it in the setting of extreme variances. Take our task in this case. Congress set a 0-60 year sentencing range for this offense, and the district court gave Poynter a 60-year sentence. Without proportionality review, without the option of measuring the force of the district court’s explanations for exceeding the roughly 20-year guidelines-recommended sentence against the length of that variance, what principle would guide us? The other § 3553(a) considerations, all important to be sure, tell us nothing by themselves about the types of individuals that should be sentenced at the extreme ends of the congressional sentencing range. They are qualitative guides, and without the quantitative recommendations of the Sentencing Commission, appellate courts would have no principled way to distinguish one appealed sentence from another. Neither district courts nor courts of appeals can- use their own experiences in determining appropriate sentences for, say, recidivist sex offenders because that would never generate nationwide consistency. The Supreme Court perhaps could gather nationwide information about sentencing ranges for different crimes, but that prospect, workable or not, seems unlikely to happen any time soon. As a' practical matter, the most meaningful way for appellate courts to “iron out sentencing differences,” Booker, 543 U.S. at 263, 125 S.Ct. 738, and to “avoid excessive sentencing disparities while maintaining flexibility to individualize sentences where necessary,” id. at 264-65, 125 S.Ct. 738, is to permit them to account for the strength of the sentencing court’s explanation in relation to the size of its deviation from the guidelines.
By contrast, it is difficult to see any great risk of harm arising from applying the proportionality principle to extreme variances. Again, take this case. The key reasonableness-review question is why the same factors underlying the “Repeat and Dangerous Sex Offender Against Minors” sentencing guideline, U.S.S.G. § 4B1.5(a), which increased Poynter’s guidelines range from 97-121 months to 188-235 months, do not account for the same concerns that the district court expressed.in raising his sentence to 720 months. Everything the court said in imposing a 720-month sentence could have been said about an individual sentenced within the 188-235 month range. What is missing, in short, is additional reasoning, not additional factfinding. If an appellate court may not ask a district court to explain why the most extreme variance available is justified — -justified in rough relation to the extent of the variance — it is difficult to see why appellate courts would not be required to uphold all such variances. And if one takes the parsimony principle seriously, what better way is there to ensure that a sentence is “sufficient” (in the context of a proposed [359]*359substantial downward variance) and “no greater than necessary” (in the context of a proposed substantial upward variance) than to insist that the district courts explain such variances from the Sentencing Commission’s recommendations — -the only empirical guide to whether sentences are “sufficient” and “no greater than necessary”?
III.
For these reasons, we vacate the sentence imposed below and remand the case to the district court for resentencing.