concurring:
Dwayne Ingram is serving a twelve year sentence for selling less than one gram of crack cocaine. More astonishingly still, his twelve year sentence is considerably less than the fifteen to nineteen year sentence *38recommended under the guidelines promulgated by the United States Sentencing Commission.
Ordinarily, someone who committed the crime which Ingram committed, and who had accrued a criminal history like his, would be subject to a guidelines range of 21 to 27 months. And even this, lesser recommended jail range derives from narcotics guidelines that “no one has ever accused ... of being anything other than stringent.” United States v. Preacely, 628 F.3d 72, 84 (2d Cir.2010) (Lynch, J., concurring).
Why, then, was Ingram’s guidelines range more than eight times higher than this already “stringent” recommendation? Because in a previous brush with the law, Ingram was caught selling 13.6 grams of crack; in another, as here, he sold less than one gram. Together, these drug felonies triggered the three-strikes provision known as the career offender guideline, U.S.S.G. § 4B1.1.
Given the great deference owed to district judges’ sentencing decisions, see United States v. Cavera, 550 F.3d 180, 189 (2d Cir.2008) (en banc), and, of course, to the judgments our elected representatives have made about the proper punishment for drug crimes, I am required to join the panel in its holding that Ingram’s sentence is not substantively unreasonable.
I write separately, however, to call attention to a procedural challenge that has been strangely absent from this case. The fact that Ingram never called the district court’s attention — or our own — to United States v. Preacely, 628 F.3d 72 (2d Cir.2010), the case that most closely corresponds to Ingram’s situation and that would have afforded him his best support both below and on appeal, leads me to believe that Preacely’s holding has not been sufficiently understood. My comments here are directed, then, to those in the bar and on the bench who may not have recognized the subtle but important contribution Preacely has made to our Circuit’s law regarding career offenders.
I. LEGAL BACKGROUND
Under the career offender provision of the sentencing guidelines, a defendant, eighteen years or older, who has at least two prior felony convictions for crimes of violence or controlled substance offenses and is charged with a third, becomes subject to a heightened offense level and an automatic assignment to criminal history category VI. In Ingram’s case, the career offender enhancement raised his offense level from 10 to 31 and his criminal history category from V to VI.
Ingram’s two previous drug felonies triggered the career offender guideline, so the district court was correct to include the relevant enhancements when it calculated the applicable guidelines range. In fact, it probably would have committed procedural error had it not done so. See Cavera, 550 F.3d at 189 (“A district court should normally begin all sentencing proceedings by calculating, with the assistance of the Presentence Report, the applicable Guidelines range.”)
The district court’s guidelines determination need not have ended there, however. The fact that a given guidelines range may apply does not always mean that it should. As U.S.S.G. § 4A1.3(b) cautions sentencing courts, “a downward departure may be warranted” if “reliable information indicates that the defendant’s criminal history category substantially over-represents the seriousness of the defendant’s criminal history or the likelihood that the defendant will commit other crimes.” We have called this a “horizontal departure,” United States v. Mishoe, 241 F.3d 214, 217 (2d *39Cir.2001), as it results in a move from right to left on the sentencing grid.
Horizontal departures are not without constraints. The guidelines require sentencing judges to put “specific reasons” for their departures in writing. U.S.S.G. § 4A1.3(e). More importantly, at least in the present context, career offenders can only receive a one-level departure, from criminal history category VI to V. Id. § 4A1.3(b)(3). We have further held that horizontal departures under § 4A1.3(b) can only be based on “individualized consideration of the circumstances of a defendant’s case.” Mishoe, 241 F.3d at 218. Thus, general rules cannot be invoked to authorize departures whenever, for example, a defendant’s criminal history involved only street-level sales or the current offense involves only a small quantity of drugs. See id. at 219. But a sentencing judge can consider factors such as “the amount of drugs involved in [the defendant’s] prior offenses, his role in those offenses, the sentences previously imposed, and the amount of time previously served compared to the sentencing range called for by placement in [criminal history category] VI.” Mishoe, 241 F.3d at 219.
In United States v. Preacely, we made clear — though perhaps not clear enough for the parties in the instant case — that a court’s discretion to make a horizontal departure under § 4A1.3(b) differs from its discretion (under United States v. Booker, 543 U.S. 220, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005), and its heirs) to choose a sentence either above or below the guidelines’ recommended range. Simply put, a horizontal departure is a decision to shift to a new guidelines range. Thereafter, a court must choose whether to depart from that new range in light of the statutory sentencing factors outlined at 18 U.S.C. § 3553(a). The first move is what Judge Raggi, at oral argument, helpfully termed a “pre-Booker departure”: a recalibration of the applicable guidelines range in order to reflect the defendant’s individualized circumstances. But these recalibrated guidelines remain advisory only, and Booker permits a second departure, either upward or downward, to what would then constitute a non-guideline sentence.
The district court in Preacely clearly understood its ability to give a non-guideline sentence under Booker. Not only did the sentencing judge state that he was “empowered to go to nothing,” Preacely, 628 F.3d at 88 (Raggi, J., dissenting), but he ultimately imposed a sentence of 94 months, exactly half the low end of the career offender guideline range. What concerned us in Preacely was the possibility that the district court did not realize it could make a pr e-Booker, § 4A1.3(b) departure and adopt a criminal history category of V rather than VI. Id. at 81 (“It is ambiguous whether the sentencing judge understood that he was not required to treat Preacely as a ‘sixth degree offender’.... ”). Though we did not decide whether a pr e-Booker departure was merited, id. at 82, we remanded for resentenc-ing due to statements in the record suggesting that the district court thought the career offender guideline had to be employed, even if a sentence within that guideline did not have to be given.
I recognize that the distinction drawn in Preacely might strike some as unnecessary, or futile. See, e.g., id. at 89-90 and n. 4 (Raggi, J., dissenting). What does the discretion afforded by § 4A1.3(b) add to the discretion already guaranteed under Booker? Adjusting a defendant’s criminal history and resulting guidelines range according to the Mishoe factors might seem to make little difference in a sentencing regime in which the guidelines are merely advisory, and judges must ultimately “impose a sentence sufficient, but not greater *40than necessary, to comply” with the sentencing factors mandated by statute. 18 U.S.C. § 3553(a).
There are two responses to this concern. The first is that the doubled exercise of discretion which the Preacely majority described — and worried the district court might not have understood — was not our invention; it is envisioned by the sentencing guidelines themselves. The Sentencing Commission is hardly unaware of Booker, of the advisory nature of the guidelines, or of the discretion district judges have to depart from them. Despite this, U.S.S.G. § 4A1.3 gives district judges the additional, constrained discretion to make the kind of horizontal departures I have described. If there is a redundancy here, it is one that our sentencing guidelines system explicitly establishes.
A second response suggests why this system makes sense. The so-called “anchoring effects” long described by cogni-five scientists and behavioral economists, see Amos Tversky & Daniel Kahneman, Judgment Under Uncertainty: Heuristics and Biases, 185 Sci. 1124 (1974), show why the starting, guidelines-departure point matters, even when courts know they are not bound to that point. When people are given an initial numerical reference, even one they know is random, they tend (perhaps unwittingly) to “anchor” their subsequent judgments — as to someone’s age, a house’s worth, how many cans of soup to buy, or even what sentence a defendant deserves — to the initial number given.1 Because anchoring effects influence our judgments, we cannot be confident that judges who begin at criminal category VI and thence depart to whatever below-guidelines sentence they think appropriate would end up reaching the same “appropriate” sentence they would have reached had they, instead, started from the category V guideline range and departed from there.2
*41Hence the advantage of a two-step process. Applying the Mishoe factors, courts can choose the more appropriate criminal history category without being influenced by anchoring effects. Its discretion at this stage is directed towards choosing a category, not picking the number of months the defendant will serve in prison. When the court turns to the latter task3 and anchoring effects potentially come into play, the anchor being used — the guidelines range — will at least be one that the court had some discretion in determining. If the court decides to make a pre-Booker horizontal departure, any Booker departure that the court subsequently chooses will be influenced not by the category VI career offender range, but instead by the somewhat lower minimum starting point found one box to the left, under category V.
For all of these reasons, I believe our Court was right in Preacely — both in its reading of the guidelines and, given what we know about human decision-making, as a normative matter.
II. APPLICATION
A. Procedural Reasonableness
None of this is to insist that Preacely squarely governs the case currently before us. I make no claim that the facts of Ingram’s case match those of Preacely’s. By the time he was sentenced, Preacely had become a particularly strong candidate for a criminal history departure under U.S.S.G. § 4A1.3(b). Preacely’s rehabilitation in the five years between his arrest and sentencing was remarkable: he never failed a drug test; “rendered significant assistance to the government”; completed a workforce development program sponsored by the district attorney; became one of his employer’s most trusted workers; and, after attending voluntary counseling, married his longtime girlfriend, supported their daughter, and became a “youth advis- or for a gang prevention program.” Preacely, 628 F.3d at 76-77. Ingram has not, or not yet, matched this extraordinary record, although the BOCES certifications he has recently received are a commendable step in the right direction.
On the other hand, Ingram is serving a far longer sentence than Preacely, despite having been caught selling a significantly smaller quantity of drugs (less than one gram to Preacely’s 21 grams) and having fewer prior drug convictions. Though both defendants were calculated to have a career offender guidelines range of 188-235 months, Preacely was sentenced to 94 months — reduced to 72 on remand — while Ingram, as noted earlier, is serving 144 months. The 96-month sentence that Ingram requested at sentencing was itself *42longer than that which prompted Preacely to appeal. All of this is relevant under Mishoe since “a comparison of the time served for prior offenses with the time to be served” is one of the factors to be considered when determining what criminal history category to apply. 241 F.3d at 219.
That said, Preacely does not require our Court to decide whether a defendant such as Ingram should have received a downward departure under § 4A1.3(b). When considering a Preacely challenge, we ask only whether the district court realized that it had the authority to make such a departure in the first place. See Preacely, 628 F.3d at 81. In the case before us, it may be that the court below understood its authority to depart horizontally but simply decided that a departure was unwarranted. The district judge, in fact, described Ingram as a “veteran” dealer, a “menace” who “made a career selling drugs to a vulnerable population in the public housing project” where he lived. Sent. Tr. at 11, 15 (emphasis added).4 At the same time, the district court also made statements suggesting that it was evaluating Ingram’s culpability and risk of recidivism on the basis of his career offender status. See, e.g., Sent. Tr. at 5 (Defense counsel: “I believe that Mr. Ingram has a substantial ability to be rehabilitated.” The court: “With a criminal history Category of VI?”); Sent. Tr. at 8 (The court: “When you get to criminal history Category VI and you’re 30 years old, that’s — you’ve been involved with the law a lot.”). While not dispositive, it was this kind of “repeated emphasis on [the defendant’s] status as a Category VI career offender” that led the Preacely majority to question whether the district court thought that criminal history category VI was mandatory. Preacely, 628 F.3d at 80.
In the end, however, these questions are not properly before us. Neither party has so much as mentioned Preacely in its briefs on appeal. And we can hardly fault the district court for failing to discuss its power to depart from category VI when no one requested that it do so. In briefing and at the sentencing hearing, Ingram’s counsel offered a full-throated, well-argued plea for a non-guidelines sentence. What Ingram did not request was a pre-Booker departure to a different guidelines range reflecting a more appropriate criminal history. To the contrary, Ingram’s pre-sen-tencing memorandum explicitly agreed that the 188-235 month range was proee-durally correct. Defs Sent. Mem. at 2. Never having been made, any procedural argument possibly open to Ingram under Preacely must be deemed forfeited.5
*43B. Substantive Reasonableness
Turning, then, to the argument Ingram did make on appeal: as I have already indicated, I agree with the Court’s holding that Ingram’s sentence is not substantively unreasonable under our precedents. But I do so with three qualifications.
First, when our Court says that a sentence like Ingram’s is “substantively reasonable,” we are using a legal term of art. Speaking in ordinary language — that is, stepping back from our labyrinthine and often draconian sentencing law — I believe that there is nothing “reasonable” about sending a man to prison for twelve years to punish him for a non-violent, $80 drug sale. Whether judged in terms of the money that will be spent to house and feed Ingram over the next twelve years,6 the disruption that will be inflicted on the lives of Ingram’s children, the missed opportunity for meaningful rehabilitation,7 or the expressive condemnation that Ingram’s crime deserves,8 a prison term of this duration is simply not a reasonable thing to impose.
But of course, that is not what we mean when we say that Ingram’s sentence is “substantively reasonable.” We mean not that it makes sense, but — as the Court’s opinion today rightly notes — that it is “located within the range of permissible decisions,” Cavera, 550 F.3d at 191, a range that is informed but not determined by the sentencing guidelines, which themselves quantify past practice.9
*44This brings me to my second qualification. Judge Raggi may have been right to say in Preacely that finding the defendant’s 94-month sentence substantively unreasonable “would be difficult ... under our precedent.” Preacely, 628 F.3d at 86 (Raggi, /., dissenting). Still, without reaching that question, Judge Lynch’s concurrence in Preacely clearly expressed his substantive concern about the length of sentences recommended under the career offender guidelines. See id. at 83 (Lynch, /., concurring) (“[E]ven for a man with a history of multiple (if mostly minor) criminal convictions (almost exclusively tied to the possession and sale of narcotics), a sentence of nearly sixteen years in prison for the possession of a few thousand dollars worth of cocaine seems remarkably severe.”) Yet even assuming the substantive reasonableness of Preacely’s 94-month sentence for 21 grams of crack, it obviously does not follow that a 144-month sentence for one gram of crack is similarly reasonable.
Nevertheless, Ingram’s sentence also is hard to find unreasonable under our precedents, given the fact that our cases provide no precedent at all for striking down a career offender sentence on substantive grounds. Our substantive unreasonableness cases have something of a self-fulfilling character: never having said that n months is too many, we have no basis in precedent for saying that n + 1 months is beyond the pale.
To be clear, I am not saying that Ingram’s sentence is substantively unreasonable. Were I to do so, this would be a dissent, not a concurring opinion. But I don’t hesitate to say that a sentence of this length, for this crime, is headed towards unreasonableness — and is, in fact, well along that road.10
*45This leads, finally, to my third qualification. The closer a sentence comes to the boundary of the substantively reasonable, the more attentive will (and should) our procedural scrutiny be. I read Preacely to exemplify this approach: the record was at best ambiguous as to whether the district court had committed procedural error, yet substantive worries about Preacely’s sentence made the very possibility of procedural error all the more salient. The wisdom of the Preacely majority’s decision to remand for resentencing was borne out by the subsequent proceedings in that case: the district court — perhaps after taking what it may have believed to be a substantive hint — reduced Preacely’s sentence from 94 to 72 months.
We should not follow the same course in the present case only, and I repeat only, because the potential procedural error in Preacely — the district court’s possible misunderstanding of the defendant’s request for a pre-Booker departure — was not argued to us in this case. My hope is that by reiterating Preacely’s distinction between pre-Booker adjustments to the guidelines range that is applied, and Booker departures from the applied guidelines range, this opinion will help make somewhat more effective the assistance given career offenders who might not fully deserve that label or the dramatic sentencing consequences it brings.
. See, e.g., Daniel Kahneman, Thinking, Fast and Slow 119-28 (2011) (describing various studies of anchoring effects); Birte Englich, Thomas Mussweiler, & Fritz Strack, Playing Dice with Criminal Sentences: The Influence of hrelevant Anchors on Experts’ Judicial Decision Making, 32 Personality & Soc. Psychol. Bull. 188 (2006) (showing the effect of rolled dice on German judges’ sentencing choices); Daniel M. Isaacs, Note, Baseline Framing in Sentencing, 121 Yale LJ. 426 (2011) (discussing the anchoring effects of sentencing baselines established under the U.S.S.G.); Nancy Gertner, What Yogi Berra Teaches About Post-Booker Sentencing, 115 Yale L.J. Pocket Part 137, 138 (2006), http://yalelawjournal.org/ images/pdfs/50.pdf (“In effect, the 300-odd page Guideline Manual provides ready-made anchors.”).
. It is important to distinguish the guidelines’ intended, salutary effect — promoting consistency and proportionality in sentencing'— from the unintended anchoring effect that the guidelines can exert. Proper reliance on the guidelines is not only rational, but legally compelled. As our court has stated, en banc, "sentencing judges, certainly, are not free to ignore the Guidelines.... The Guidelines provide the starting point and the initial benchmark for sentencing, and district courts must remain cognizant of them throughout the sentencing process.” Cavera, 550 F.3d at 189 (quotation marks and citations omitted). Anchoring leads to cognitive error not insofar as judges intentionally use the guidelines in an advisory fashion, but instead when “judges irrationally assign too much weight to the guidelines range, just because it offers some initial numbers.” Ryan W. Scott, Inter-Judge Sentencing Disparity After Booker: A First Look, 63 Stan. L. Rev. 1, 45 (2010); see also Isaacs, Baseline Framing, supra, at 442 (“[Guidelines are supposed to be actively considered in determining the appropriate sentence; they are not supposed to encourage irrational decisionmaking.”).
Examples from the empirical literature illuminate this point. Studies have shown that judges weigh prosecution requests heavily when making decisions about sentencing and bail. See Englich et al., Irrelevant Anchors, supra, at 189. Such weight may well be rational in situations where the judge trusts a familiar prosecutor’s experience and commit*41ment to justice (over mere “winning”). The problem is that judges have also been shown — in fictionalized situations that psychologists have constructed to test for unconscious anchoring effects — to be influenced by prosecution requests which they know were randomly generated. See id. at 192-95. It is these latter, irrational effects that that make it important to pick the most appropriate guideline starting point even (or especially) when judges have the discretion to depart from that point.
. Significantly, a sentencing court might depart for different reasons during its pre-Booker and Booker exercises of discretion — that is, under U.S.S.G. § 4A1.3 and 18 U.S.C. § 3553(a). The grounds for a pre-Booker departure are narrower than those under § 3553(a). A court, for example, which has a policy disagreement with the crack sentences recommended under the guidelines would be justified in departing below the guidelines for that reason. See Spears v. United States, 555 U.S. 261, 264, 129 S.Ct. 840, 172 L.Ed.2d 596 (2009) (per curiam). It would not, however, be justified in horizontally departing to a lower guidelines range. See Mishoe, 241 F.3d at 218.
. The 44-month downward departure which the district court did give Ingram seems to have been motivated by a point Ingram himself made at sentencing: namely, that he was a drug-addicted member of the very same “vulnerable population” the district court blamed him for targeting. See Sent. Tr. at 21 (Ingram: “[T]he people you talking about is me, Your Honor, that you looking at is me. I'm part of them — them addicts.”); 25 (The Court: "I’m going to give him the benefit of some of the things he said and give him a little bit more leniency th[a]n maybe I otherwise would have done.”).
. A few further comments about procedural unreasonableness are made necessary by Judge Raggi's powerful concurring opinion. Judge Raggi's concurrence seems to me mostly to show her deep and continuing disagreement with Preacely. See post at 45-51. She does argue that the sentencing judge here, more so than the judge in Preacely, clearly knew that he could make a horizontal departure. On that we simply disagree. While I acknowledge that the question is closer in this case than in Preacely, I believe that, even here, raising a Preacely challenge on appeal might not have been useless, had the issue been properly presented to the district court. While I do not disagree with Judge *43Raggi that the facts in this case may have argued less powerfully for a horizontal departure than did those in Preacely, my point is simply that raising the issue to the district court would not have been fatuous. Putting aside Judge Raggi’s post hoc rationalizations for counsel’s failure to mention Preacely, see post at 50-51, I worry that the case was not raised by either side in this litigation simply because they had not grasped its import — as defense counsel admitted at oral argument.
Putting these points aside, the thrust of Judge Raggi’s concurrence — her critique of sequential departures, or what I have called the doubled exercise of discretion under § 4A1.3(b) and Booker — applies every bit as much to Preacely. This does not mean that her argument has no force, but only that what force it has proved insufficient to persuade the previous panel.
. At current rates, the cost of Ingram’s incarceration will range somewhere between $252,000 and $407,000. See U.S. Gov’t Accountability Office, Bureau of Prisons: Eligibility and Capacity Impact Use of Flexibilities to Reduce Inmates' Time in Prison 19 fig. 3 (2012).
. See United States v. Leitch, No. 11-CR-609 (JG), 2013 WL 753445 (E.D.N.Y. Feb.28, 2013); Mosi Secret, Outside Box, Federal Judges Offer Addicts a Free Path, N.Y. Times, Mar. 1, 2013 at Al.
. If the length of sentence is meant, at least in part, as a measure of society's condemnation of a given act, what are we to make of a sentence which punishes a one-gram sale of crack more severely than New York’s minimum sentences for first degree rape, kidnapping, or assault with a deadly weapon? See N.Y. Penal Law § 70.02.
. Judge Raggi worries that labeling a sentence substantively reasonable yet absurd "risks calling the law into disrepute.” Post at 51. Here Judge Raggi conflates what is reasonable under current law with the reasonableness of that law itself.
It would certainly be strange for us to say that something within a judge's complete discretion is at once substantively reasonable and absurd. Sentencing, however, is not defined by complete judicial discretion. The legislature tells judges that we must do certain things in regard to sentencing, and the legislature's judgment thereby defines what sentences lie within the pale. Such sentences cannot be found substantively unreasonable without infringing on the prerogative of the legislature.
And yet, we judges have a right — a duty even — to express criticism of legislative judgments that require us to uphold results we think are wrong. We may alert Congress to mistakes or gaps in its legislation. See Robert A. Katzmann, Courts and Congress (1997). *44We may tell the legislature that we think a judgment it has made is mistaken, even absurd, see United States v. Dorvee, 616 F.3d 174, 184-87 (2d Cir.2010), and urge Congress to reconsider its judgment, see United States v. Douglas, 713 F.3d 694, 702-03 (2d Cir.2013); United States v. Acoff, 634 F.3d 200, 205 (2d Cir.2011) (Lynch, /., concurring). We may even go further and suggest that a judgment made by the legislature is headed towards unconstitutionality. See United States v. Then, 56 F.3d 464, 468-69 (2d Cir.1995) (Calabresi, concurring); infra note 10. To do these things is not to "call[ ] the law into disrepute,” but rather to work with coordinate branches of government to prevent disreputable laws from enduring.
Thus while Judge Raggi's point — that an act within a judge’s own discretion cannot be both absurd and reasonable — is well-taken, it does not speak to the problem raised by Ingram’s sentence, or by the career offender guidelines more broadly. When the legislature tells judges that repeated small-scale drug transactions should be punished more severely than rape, see supra note 8, it may well be “substantively reasonable” for judges to impose such disproportionate sentences, as instructed, and yet appropriate for them to decry the instruction itself as absurd. (Judge Raggi, in her concurrence, takes issue with this description of Congress's instructions. See post at 51 n. 9. But this is a career offender case, and in regard to such cases, Congress has decreed that the Sentencing Commission "shall assure that the guidelines specify a sentence ... at or near the maximum term authorized." 28 U.S.C. § 994(h) (emphasis added).)
. We might compare this to laws which, due to changed circumstances, can be described as “heading toward unconstitutionality.” United States v. Then, 56 F.3d 464, 469 (2d Cir.1995) (Calabresi, /., concurring). As I have previously urged, rather than reaching to strike down a law or, alternatively, "allowing legislative inertia to dominate,” id., courts should instead initiate a dialogue with lawmakers, putting them “on notice that a serious and thoughtful legislative review and reconsideration [is] in order and that failure to undertake such a review might in time result injudicial action...." Id. The present case, of course, involves not unconstitutionality but our capacity to find a sentence substantively unreasonable. And I speak here not (or not just) to Congress, but also to the Sentencing Commission and the district judges responsible for applying its guidelines.