POSNER, Circuit Judge.

The defendant was ordered reimprisoned for 20 months for violating the terms of his supervised release imposed for a firearms offense. His appointed counsel has filed an Anders brief arguing that there is no nonfrivolous ground for an appeal. The defendant's supervised release was revoked because he committed a felony under Illinois law, and, as counsel acknowledges, there is "plausible evidence" that his client indeed committed the felony. Hence (as explained in an unpublished order also issued today, which considers as well additional issues discussed in the Anders brief) the appeal is frivolous — provided we adhere to this court's position that a sentence imposed after the revocation of supervised release can be set aside only if it is "plainly" unreasonable.

We reserved the question in United States v. Flagg, 481 F.3d 946, 949 (7th Cir.2007). We explained that before the Supreme Court's decision in United States v. Booker, 543 U.S. 220, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005), our court and the other courts of appeals as well had held that the standard of "plainly unreasonable" was the applicable one, e.g., United States v. Carter, 408 F.3d 852, 854 (7th Cir.2005), but that after Booker five of the courts had decided that the proper standard was "unreasonable," see, e.g., United States v. Miqbel, 444 F.3d 1173, 1176 n. 5 (9th Cir.2006); United States v. Cotton, 399 F.3d 913, 916 (8th Cir.2005); United States v. Fleming, 397 F.3d 95, 99 (2d Cir.2005), while two others decided to adhere to "plainly unreasonable," United States v. Crudup, 461 F.3d 433, 437-39 (4th Cir.2006); United States v. Johnson, 403 F.3d 813, 816-17 (6th Cir.2005), although the court in Johnson left open the possibility of reconsidering its position.

Until Booker the federal sentencing guidelines had been mandatory rather than advisory — but not when the sentence was imposed following revocation of supervised release. There were no formal sentencing guidelines for such sentences, but only "policy statements" that explicitly were merely advisory, U.S.S.G. ch. 7, pt. A, Introductory Comments 1, 3(a); United States v. Carter, supra, 408 F.3d at 854, and the Sentencing Reform Act authorizes a court of appeals to set aside a sentence for which there is no guideline only if the sentence is "plainly unreasonable." 18 U.S.C. §§ 3742(a), (b), (e)(4). Subsections (a) and (b) were not invalidated in Booker, but section 3742(e), the section setting forth the standard of appellate review of federal sentences, was. 543 U.S. at 260, 125 S.Ct. 738. (Subsections (a) and (b) are similar to (e), but concern the criteria for filing a notice of appeal rather than the criteria to be used by the appellate court in deciding the appeal.) It is because the "plainly unreasonable" standard for review of nonguideline sentences appears in a subsection of section 3742(e) (subsection (4)) that some courts of appeals think that the "plainly unreasonable" standard no longer governs appellate review of sentences imposed after revocation of supervised release.

This strikes us as an overbroad reading of the Booker decision. Section 3742(e) places tight limits on appellate review of guidelines sentences, limits designed to implement Congress's decision (invalidated in Booker) to make the guidelines mandatory. Nothing in either of the Court's majority opinions in Booker suggests that limiting appellate review of sentences not based on the guidelines is needed to avoid the constitutional problem that required the invalidation of parts of the Sentencing Reform Act in order to save the rest of it. The constitutional problem was that judges were basing sentences on facts that they found — not a jury — and by a preponderance of the evidence rather than by proof beyond a reasonable doubt. The Court held that for sentencing grounded in such a factfinding process to be mandatory violated the Sixth Amendment. Changing the standard of appellate review of guidelines sentences was necessary because the standard made the guidelines mandatory in appellate proceedings, complementing 18 U.S.C. § 3553(b), which made them mandatory at the sentencing stage. Since there are no guidelines for sentences for violating a condition of supervised release, there was no occasion for the Court in Booker to change the standard of appellate review of such sentences. The Court did not mention subsection (e)(4), which governs the appellate review of nonguidelines sentences, because its decision was concerned only with guideline sentences.

We are not disregarding a Supreme Court dictum (though it would not be lèse majesté to do so), for apart from the omission of a reference to subsection (e)(4), there is nothing in the logic or language of the Booker majority opinions to suggest that the Court was altering the statutory standard of appellate review of sentences for violating conditions of supervised release. The Court said that "reasonableness standards [not standard] are not foreign to sentencing law. The Act has long required their [not its] use in important sentencing circumstances — both on review of departures, see 18 USC § 3742(e)(3) (1994 ed.), and on review of sentences imposed where there was no applicable Guideline, see §§ 3742(a)(4), (b)(4), (e)(4)." United States v. Booker, supra, 543 U.S. at 262, 125 S.Ct. 738 (emphasis added). The citations that follow the quoted passage are to both standards, and there is nothing to suggest that the Court meant to merge the two. The only ground for doing so would be that the "plainly" unreasonable standard was unconstitutional, and, to repeat our earlier point, there is nothing in the language or reasoning of the Booker opinions to suggest that. We shall therefore adhere to our ruling in the Carter case requiring that a defendant who challenges his sentence for violating supervised release show that the sentence is plainly unreasonable.

Realism, however, requires acknowledgment that the practical difference between "unreasonable" and "plainly unreasonable" is slight, perhaps even nil, so the tendency to equate them, as in United States v. Sweeting, 437 F.3d 1105, 1106-07 (11th Cir.2006) (per curiam), and United States v. Cotton, supra, 399 F.3d at 916, is understandable. But as pointed out in United States v. Crudup, supra, 461 F.3d at 437-39, the courts must respect Congress's wish to curtail appellate review of non-guidelines sentences particularly sharply, and so must seek to give meaning to the difference between "unreasonable" and "plainly unreasonable." The Crudup opinion gives reasons for why a more limited scope of appellate review is appropriate for such sentences: the Sentencing Commission's decision not to issue guidelines implies that the sentencing court should have more than usual flexibility in sentencing for violations of conditions of supervised release, id. at 438; and the maximum sentence is only five years, id. at 439 n. 8; 18 U.S.C. § 3583(e), and the shorter the sentence, and hence the less there is at stake, the fewer the layers of judicial review necessary to satisfy the requirements of due process of law. Mathews v. Eldridge, 424 U.S. 319, 96 S.Ct. 893, 47 L.Ed.2d 18 (1976).

But still, while appellate courts understand and can implement the difference between deferential and nondeferential review, the making of finer gradations within the category of deferential review strains judicial competence, as we have often noted. School District of Wisconsin Dells v. Littlegeorge, 295 F.3d 671, 674-75 (7th Cir.2002); United States v. Hill, 196 F.3d 806, 808 (7th Cir.1999); Aegerter v. City of Delafield, 174 F.3d 886, 889-90 (7th Cir.1999); Johnson v. Trigg, 28 F.3d 639, 643-44 (7th Cir.1994). The gradations exist formally: there is clear-error review, substantial-evidence review, review for rationality (as of jury verdicts, where the test is whether any rational trier of fact could have arrived at the jury's verdict), arbitrary-and-capricious review, abuse-of-discretion review, ultra-narrow review of credibility determinations based on a witness's demeanor, and more. But in most cases, regardless of the formal gradation of deferential review, the appellate judges are merely giving the benefit of the doubt to the trier of fact or other first-level decision maker — how much benefit of the doubt depends less on the formal standard than on the nature of the issue and the institutional competence of the first-level decision maker relative to that of the appellate court. So while we must do our best to mark any gradations prescribed by Congress, we cannot promise great success in the endeavor.

Perhaps, however, we can borrow for the present class of cases the narrowest judicial review of judgments we know, and that is judicial review of the sanctions imposed by prison disciplinary boards. Superintendent v. Hill, 472 U.S. 445, 455-57, 105 S.Ct. 2768, 86 L.Ed.2d 356 (1985) (the prison disciplinary board's ruling must be supported by a "modicum" of evidence, "some" evidence, "any" evidence, or even just "meager" evidence); Scruggs v. Jordan, 485 F.3d 934, 941 (7th Cir.2007) (must cross a "meager threshold"); Webb v. Anderson, 224 F.3d 649, 652 (7th Cir.2000) ("some" evidence, "any" evidence, "modicum of evidence," "`meager' proof," evidence just needs to `"point to the accused's guilt'"; the standard of review is "lenient"). Such sanctions must indeed be "plainly" unreasonable to be set aside. Under that standard, the appeal in this case is frivolous. The Anders motion is therefore granted and the appeal dismissed.