dissenting.
I agree that the district court erred in the guideline loss calculation, but I respectfully dissent because that error was harmless. The record shows an experienced district judge considering a difficult *685case thoroughly and exercising his discretion reasonably under 18 U.S.C. § 3553(a) to craft a sentence to fit both the crime and the criminal. The error in the guideline calculation did not affect the final decision, and I find no abuse of discretion in the final decision about the sentence.
As the majority explains, given the nature of Vrdolyak’s and Levine’s crime, it is more accurate to say that there was a loss that cannot be determined reliably than to say that there was no loss to the victim of the crime. Where the crime makes it so difficult to determine with confidence the amount of the loss, the guidelines and cases from this and other circuits establish that the intended gain for the conspirators is a useful substitute for loss in applying the guidelines to gauge the severity of the crime. Application Note 3(B) to U.S.S.G. § 2B1.1 provides: “The court shall use the gain that resulted from the offense as an alternative measure of loss only if there is a loss but it reasonably cannot be determined.” That approach has been used, for example, in United States v. Briscoe, 65 F.3d 576, 589-90 & n. 16 (7th Cir.1995) (affirming sentence of union president who defrauded union through kickback scheme with lender to union members; loss calculation based on defendant’s gain), as well as in United States v. Vinyard, 266 F.3d 320, 332 (4th Cir.2001) (affirming sentence in kickback scheme based on defendant’s gain), and United States v. Yeager, 331 F.3d 1216, 1224-26 (11th Cir.2003) (affirming sentence based on defendant’s gain through scheme to divert drug sales through unauthorized dealers at lower prices).
But the guideline calculation is only the beginning of the story, for both the district court and this court. Pursuant to United States v. Booker, 543 U.S. 220, 259-60, 125 S.Ct. 738, 160 L.Ed.2d 621 (2005), after calculating the applicable sentencing guideline range, the district court was required to look beyond the guidelines and to consider the case under 18 U.S.C. § 3553(a). Congress has instructed:
The court shall impose a sentence sufficient, but not greater than necessary, to comply with the purposes set forth in paragraph (2) of this subsection.
Paragraph (2) of subsection (a) requires the court to consider:
the need for the sentence imposed—
(A) to reflect the seriousness of the offense, to promote respect for the law, and to provide just punishment for the offense;
(B) to afford adequate deterrence to criminal conduct;
(C) to protect the public from further crimes of the defendant; and
(D) to provide the defendant with needed educational or vocational training, medical care, or other correctional treatment in the most effective manner....
18 U.S.C. § 3553(a). The district court was required to consider the guidelines, but it was prohibited from presuming that a guideline sentence would be a reasonable sentence. Rita v. United States, 551 U.S. 338, 351, 127 S.Ct. 2456, 168 L.Ed.2d 203 (2007).
At the end of the lengthy sentencing hearing in this case, the district judge addressed the factors and purposes under section 3553(a). He explained how and why he had concluded that a sentence of five years’ probation (which was above the guideline range he calculated), plus 2500 hours of community service and a fine, was sufficient but not greater than necessary to serve those purposes. The defendant committed a serious crime, but there were a number of factors that the district court could and did consider in mitigation. The defendant is 71 years old, had no prior *686criminal record, and posed little risk of repeat offenses. He had given up his law license. The crime of fraud did not involve violence, and there was no element of public corruption. The defendant had agreed to help a friend by committing the crime, but he was not the instigator of the crime and did not actually benefit from it. The district court was also impressed by a surprising volume of information showing the defendant’s character was very different from his public image in the media. That information showed generosity with time, money, and influence to help people in need, especially where the defendant had no moral or other obligation to help them and where he received no publicity or recognition for his kindnesses. That is not the entire picture, of course, but those are all factors that could reasonably lead the district court to exercise its discretion under section 3553(a) to impose the sentence that it did. See Gall v. United States, 552 U.S. 38, 48-49, 128 S.Ct. 586, 169 L.Ed.2d 445 (2007) (affirming below-guideline sentence of probation and recognizing substantial restrictions on liberty imposed by sentence of probation).
Under Booker and Gall the district court is required to calculate the applicable sentencing guidelines for the crime and the criminal, and an error in the calculation is a procedural error in sentencing that may require a remand. Booker, 543 U.S. at 259-60, 125 S.Ct. 738; Gall, 552 U.S. at 49-51, 128 S.Ct. 586. At the same time, however, it is clear that errors in calculating the advisory guideline calculations are subject to harmless error analysis. E.g., United States v. Abbas, 560 F.3d 660, 666-67 (7th Cir.2009) (holding guideline error was harmless); United States v. Anderson, 517 F.3d 953, 965-66 (7th Cir.2008) (same); see generally Williams v. United States, 503 U.S. 193, 203, 112 S.Ct. 1112, 117 L.Ed.2d 341 (1992) (stating before Booker that guideline errors were subject to harmless error analysis). In both Abbas and Anderson, the district courts recognized the disputed guideline issues, stated that their sentences would be the same regardless of how the guideline issues were decided, and provided thoughtful explanations of their reasoning. In such cases, because the sentencing guidelines are no longer mandatory, appellate courts should readily find that guideline errors are harmless.
Correct application of the guidelines can present many difficult or esoteric questions, including many that have little to do with the ultimate legal and moral judgment about an appropriate sentence. Since Booker, this court has often recognized that the sentencing judge may impose a reasonable sentence under section 3553(a) regardless of how a difficult guideline issue might be resolved. “When a judge proceeds in this manner, she must make clear that the § 3553(a) factors drive the sentence without regard as to how the prior conviction fits under a particular guideline. Doing so will make the often nit-picking review of issues like this under our now advisory guideline scheme unnecessary.” United States v. Sanner, 565 F.3d 400, 406 (7th Cir.2009) (affirming above-guideline sentence without regard for correct resolution of guideline issue); Abbas, 560 F.3d at 666-67 (finding that district court erred in guideline calculation but holding error was harmless based on judge’s explanation of alternative basis for same sentence). In this case, the judge considered the relevant factors thoughtfully and made his intentions and reasons clear. The precise level of loss in the judge’s alternative guideline calculation did not drive the final decision.
As Abbas and Anderson make clear, this is not to say that a district court can insulate any sentence from appellate review by saying a few magic words about *687section 3553(a). Abbas, 560 F.3d at 666-67; Anderson, 517 F.3d at 965; accord, United States v. Peña-Hermosillo, 522 F.3d 1108, 1118 (10th Cir.2008) (finding guideline error was not harmless where district court provided only “perfunctory” explanation for alternative rationale); see generally United States v. Williams, 431 F.3d 767, 773-76 (11th Cir.2005) (Carnes, J., concurring) (encouraging district courts to provide alternative sentencing rationales where resolution of disputed guideline issues would not affect sentences). But where the record shows that the district court considered the disputed guideline issue, considered the prospect that its decision on the issue might be wrong, and provided a thoughtful explanation of its reasons under section 3553(a), it should be relatively easy to find that an error in calculating an advisory guideline was harmless, as I believe this one was. In some important respects, this case provides a mirror-image of United States v. Spano, 476 F.3d 476 (7th Cir.2007), a public corruption case in which the district court imposed a sentence on one defendant that was substantially higher than the proper guideline range. The district court imposed an upward departure of four levels for extraordinary abuse of trust, and this court held that the decision was an error. We found that the error was harmless, however, because the judge explained why he thought a guideline sentence that did not take into account the egregious abuse of trust would not be adequate. Despite the guideline error, we upheld the above-guideline sentence as a proper and sensible exercise of the district court’s discretion under section 3553(a). Id. at 480-81. In this case involving private corruption, the district court imposed a sentence below the correct guideline range, but with a reasonable exercise of that same discretion.
The majority identifies three reasons why the loss calculation error should not be deemed harmless and concludes further that the case should be remanded to another district judge under Circuit Rule 36 to ensure the government a fair hearing. In my view, the criticisms here are not warranted, and the government received a fair hearing before the district court.
First, the majority criticizes the district court for basing the alternative sentencing rationale on the assumption that the guideline loss could have been no higher than $500,000, which would put the sentencing guideline range two offense levels lower than the correct range here.1 The district court’s thorough discussion of the factors relevant to sentencing under section 3553(a) showed that the two-level difference would not have made any difference in the court’s ultimate decision, so the error in calculating the loss under the guidelines had no effect. The sentencing range that is 14 levels below level 20 (the correct level) is the same as the range that is 14 levels below the level 18 the district court considered as its alternative—zero to six months in prison. And in fact, though we reach the conclusion from opposite directions, the majority and I agree that there is no point in remanding the case to Judge Shadur for a new guideline calculation and imposition of the same sentence as before. Where the majority sees a *688“cascade of errors and omissions,” I see just one harmless error in the advisory guideline calculation.
Second, the majority criticizes the district court for believing the defendant intended no harm to the medical school and actually intended to benefit it by arranging for the highest bid from Smithfield. This criticism does not accurately reflect the record. On the first point, the district court pointed out correctly that there was no evidence that the defendant intended to hurt the medical school, see Gov’t App. at 25, and the government itself had submitted evidence showing that the defendant had wanted Smithfield to pay “top dollar” for the property. See Def. Reply Mem. at 8 (quoting recorded conversation on March 31, 2006). The district court necessarily recognized, however, that the defendant must have realized that his help for his friend Levine would hurt the medical school by distorting the sale process in favor of the corrupt side deal to benefit Levine and the defendant. That’s why the defendant is guilty. That’s why the district court accepted his plea of guilty. On the second point, the furthest the district judge went was to note several times that the defendant’s “finder’s fee” (10 percent of the purchase price, to be split with Levine) would go up as Smithfield’s purchase offer went up. Gov’t App. 21-22, 25. I do not find in the record any indication that the district judge thought that the defendant, who had pled guilty to a serious crime and whom he was sentencing for that crime, acted with the intent to benefit the medical school. The district judge understood how Levine and Vrdolyak had corrupted the sale process—Vrdolyak had pled guilty, after all.
Third, the majority criticizes the district court for giving too much weight to numerous letters urging leniency in sentencing while giving too little weight to information that hurt the defendant’s cause. On this topic, the majority identifies three more specific errors: failing to discuss adequately the information weighing against the defendant, failing to consider the defendant’s wealth and its effect on his ability to show mitigating good works, and ignoring the defendant’s earlier work as an influential Chicago alderman.
These criticisms are not warranted. The record shows that the district court gave careful and discriminating consideration to the mitigating and aggravating information. The letters in mitigation came from the defendant’s family and friends, and from others who described ways in which the defendant had helped them over the years. The district judge said that those letters were an “extraordinary outpouring that’s not matched—at least in my recollection—in any other case that I have had coming up to be 29 years on the bench.” Gov’t App. 76.
Like victim impact statements, such letters are entirely appropriate in a sentencing hearing. Cf. Gall, 552 U.S. at 43, 128 S.Ct. 586 (affirming below-guideline sentence of probation where district court relied in part on “small flood” of letters from family, friends, neighbors, and business associates). The district court has an obligation to consider such letters when considering the history and characteristics of the defendant. See 18 U.S.C. § 3553(a)(1). Such letters often add little to the relevant picture of the defendant and his crime, but sometimes they can provide unexpected information and add new insights into the defendant’s character.
This was such a case. The veteran district judge was surprised by what he learned. Like almost everyone who has lived in Chicago over the past three or four decades, the judge had been generally aware of the reputation of “Fast Eddie” Vrdolyak, leader of the opposition to May- *689or Harold Washington and power broker in Chicago politics. The judge did his best to put aside those preconceptions and to approach the sentencing decision with an open mind. He explained candidly:
When I first encountered the case, and throughout its pendency, I never expected that I was going to reach the destination that I find called for here. And as I have indicated, I had (as I suspect anybody who has seen the political environment in Chicago over a long time frame probably shared) a perception of our defendant today that I suppose is epitomized by the moniker “Fast Eddie,” indeed from the so-called Council Wars that existed during the Harold Washington mayoralty.
Gov’t App. 135-36. In explaining that he would have reached the same result regardless of the guideline loss calculation, the judge said:
In any event, the point I think that is most important is that when I applied 3553(a) I would reach the same result whichever of those views [about loss under the guidelines] is taken. And that’s because I again surprised myself in terms of how I looked at this thing coming in, with the absence of a full appreciation or full understanding or disclosure of what the thing involved. I would not have dreamed of imposing a noncustodial sentence. But I have got to tell you that when you look at the 3553(a) factors, it seems to me that the reasonable result, the one that is called for taking all of the considerations into account, is just what I have indicated.
Id. at 138. These frank observations of the judge—about the ways in which the facts overcame the defendant’s public reputation and persuaded the judge to reach a result he had not expected to reach—deserve substantial weight.
Contrary to the majority’s criticisms about failing to consider the defendant’s wealth and influence, the district judge was discriminating in weighing the letters about the defendant’s character. During the defense presentation, the judge commented:
I am not sure that things that are done for family members carry—or for that matter for what’s called a public persona carries—much weight. At least as I read these letters, the thing that I found frankly most persuasive on his part were the things that were not visible, and things in which at least according to these people he reached out in situations where he need not have done so. He had no obligation to do that either morally or otherwise and did it anyway. And that’s frankly the reason that I characterized the letters that I received—not simply in terms of volume, but in terms of impact—as giving the kind of astonishingly different portrayal than what you have characterized as public persona.
Gov’t App. 93 (emphasis added).
The most important mitigating information here involved not the “checkbook charity” that can be easy for the wealthy, but many instances in which the defendant provided hands-on help in long-term relationships with people in need, or where he provided generous help anonymously. The district judge’s comments show that he was much more impressed by these many instances of the defendant’s generosity and kindness with his time and influence in situations where there was no visibility or public reward for his actions.2
*690Even with the appropriate discounting desired by the majority, the letters still have unusual persuasive weight, and the district court was not required to ignore or discount the evidence of past good works. The majority’s observation that wealthy defendants should not get a break compared to poor ones merely because they have given away some of their wealth is certainly true but misses the district court’s real point in weighing this mitigating information.
The majority writes that the district court “ignored the fact that the defendant was for many years an influential Chicago alderman.” op. at 683. This fact is relevant, the majority asserts, because politicians are in the business of dispensing favors, so that later expressions of gratitude for politicians’ largesse should not be given weight in a sentencing decision. This criticism is not warranted. The defendant left public office more than 20 years ago, long before most of the events cited in mitigation. His history in Chicago politics and public life certainly did not escape the district judge’s notice. The district judge’s comment quoted above— discounting letters about things the defendant had done for his “public persona”— shows that the district court was quite conscious of precisely the point the majority faults him for ignoring. See also Gov’t App. 137 (court referring to “the extraordinary volume and character of the things that I received in support of Mr. Vrdolyak, the kinds of thing that the public persona would never have dreamed existed, and I certainly not”). In other words, the letters that impressed the court most were those addressing private actions that were not designed to help the defendant himself, in his private or public life.3
The majority criticizes the district court for not saying more about evidence putting the defendant’s character in a bad light. The district court showed that it was familiar with the evidence the government had submitted, including the incidents cited from the defendant’s legal career. The court acknowledged the “substantial information that’s unfavorable to Vrdolyak” and said it was reminded of Dr. Jekyll and Mr. Hyde, as defense counsel urged a sentence for Dr. Jekyll and the government urged a sentence for Mr. Hyde. Gov’t App. 76. The court specifically asked the defense to address the lawyer disciplinary matters, id. at 127, and the defense provided explanations that offered at least some mitigating effect. Id. at 128-29. Criticism of the district court for having failed to say more about these events, which had been the subject of two censures and one short suspension from practice, demands too much from an oral sentencing explanation.
Based on what is described as a “cascade of errors and omissions,” the majority *691orders a remand under Circuit Rule 36 to another district judge for a fresh look at the sentence. I see instead just one harmless error, and I do not believe the government was denied a fair hearing. The majority concludes that Judge Shadur “had committed himself irrevocably to a noncustodial sentence for the defendant.” op. at 683. The majority criticizes the judge for having announced this view near the outset of the sentencing hearing, before the government had a chance to argue for a custodial sentence. This criticism is off target for two reasons.
First, it fails to acknowledge how much work the parties and the district court had put into the sentencing decision before the hearing itself. In detailed written briefs, the parties had set forth their positions on the guideline issues, the section 3553(a) factors, and the sentence that would be appropriate. The district judge’s comments throughout the sentencing hearing show that he fully understood the parties’ positions from the outset. Having done such extensive preparation, it would have been unusual for the judge not to have had a good idea how he intended to sentence the defendant, and why, at the beginning of the sentencing hearing.
Second, the fact that the judge signaled his informed inclinations early in the hearing does not call for criticism. Most lawyers appreciate knowing more rather than less about the judge’s thinking while there is still an opportunity to persuade the court. This criticism is no more warranted than would be criticism that an appellate judge had a view of the case at the beginning of an oral argument. Such views do not mean that minds are closed to further persuasion by probative evidence and legal argument. See also Gov’t App. 76-77 (district court inviting views and stating “I really have not reached a conclusion”). Judges must keep their minds open to new information that will change their thinking, as the district judge did in this case. And where the majority finds in the record a district judge who was without justification impatient and angry with the government, I find a frank and wide-ranging discussion of the issues in the case, and some reasonable annoyance with overly aggressive arguments and invitations to speculate.4
Perhaps the most remarkable thing about the way this case has proceeded on appeal is that the government’s briefs did not challenge the reasonableness of the sentence or ask for remand to a different judge. In its opening brief, the government argued only the guideline loss error. The government did not even bother to mention the district court’s alternative guideline calculation and analysis under 18 U.S.C. § 3553(a). Where the district court has stated such an alternative basis, we should treat the appellant’s silence as at least a forfeiture of the issue. And it is hard to believe that the government’s approach to this appeal was not carefully considered in every respect. We would be justified in finding a waiver based on the government’s failure to address the alternative calculation and its failure to challenge the reasonableness of the sentence.
Finally, although it is evident that the majority views the sentence here as simply not heavy enough to punish this crime adequately, we should not overlook an important dimension of this sentence that may in the end be more powerful than a *692shorter prison term. The idea of a person as prominent as Ed Vrdolyak doing 2500 hours of community service in Chicago has elements of public shaming and service that were well within the district court’s discretion in deciding how best to accomplish the purposes of sentencing with a sentence “sufficient, but not greater than necessary,” as section 3553(a) directs. It is hard to imagine that this defendant’s community service could be completed without considerable news media attention, which would add to the deterrent effect of that portion of the sentence.
In sum, the record here shows that an experienced judge considered the case thoughtfully and learned information that overcame his initial inclinations in the case (and many public reactions to the case). The judge exercised his discretion under section 3553(a) and imposed a sentence reasonably tailored to fit both the crime and the criminal. Though the majority and I disagree with the district judge on the loss calculation under the advisory sentencing guidelines, that error was harmless because it did not drive the final sentencing decision under section 3553(a). I would affirm the district court’s judgment.
. The correct guideline calculation starts with a base offense level of 6 under U.S.S.G. § 2B 1.1 (a)(2), adds 16 levels under § 2B 1.1 (b)(1)® for a loss of more than one million dollars, and subtracts two levels for acceptance of responsibility under § 3El. 1(a), for a total offense level of 20. With Vrdolyak's criminal history category of I, the sentencing guideline range is 33 to 41 months in prison. Using a loss of $500,000, the district court assumed an upward adjustment of 14 and a total offense level of 18, with a guideline range of 27 to 33 months in prison.
. "But, surely, if ever a man is to receive credit for the good he has done, and his immediate misconduct assessed in the context of his overall life hitherto, it should be at the moment of his sentencing, when his very future hangs in the balance. This elementary *690principle of weighing the good with the bad, which is basic to all the great religions, moral philosophies, and systems of justice, was plainly part of what Congress had in mind when it directed courts to consider, as a necessary sentencing factor, ‘the history and characteristics of the defendant.’ " United States v. Adelson, 441 F.Supp.2d 506, 513-14 (S.D.N.Y.2006), aff'd mem., 301 Fed.Appx. 93 (2d Cir.2008).
. The letters of Dr. Mark Siegler and others described several instances in which the defendant intervened to enable needy people to obtain life-saving medical care that was not otherwise available to them. See Def. Ex. N (Siegler). The letter of Jonathan Kleinbard, a former vice president of the University of Chicago, told of a case in which Vrdolyak represented a plaintiff suing the University of Chicago Medical Center for medical malpractice. Vrdolyak failed in his effort to settle the case with Kleinbard’s help, and he eventually won a judgment "in the millions of dollars.” Kleinbard reported that Vrdolyak then donated his entire legal fee to the University of Chicago Medical Center. Def. Ex. Z.
. The majority criticizes the district court for refusing to consider after-the-fact evidence from other interested buyers about how much more they would have been willing to offer for the property. I see no abuse of discretion in the district court's decision not to spend time on the government’s speculative effort to show the likely result of an honest effort to have sold the property for the best available price.