In the
 United States Court of Appeals
              For the Seventh Circuit
                        ____________

No. 07-1266
UNITED STATES OF AMERICA,
                                               Plaintiff-Appellee,
                               v.

CHRISTOPHER MCILRATH,
                                           Defendant-Appellant.
                        ____________
           Appeal from the United States District Court
              for the Eastern District of Wisconsin.
            No. 06-CR-182—J. P. Stadtmueller, Judge.
                        ____________
   ARGUED NOVEMBER 14, 2007—DECIDED JANUARY 10, 2008
                        ____________


  Before POSNER, WOOD, and WILLIAMS, Circuit Judges.
  POSNER, Circuit Judge. The defendant pleaded guilty to
the crime of traveling across state lines to have sex with a
minor, in violation of 18 U.S.C. § 2423(b), and was sen-
tenced to 46 months in prison. That is the bottom of the
applicable sentencing guidelines range, but he chal-
lenges the sentence as too severe, arguing that he should
have been sentenced just to home confinement.
  The defendant was caught in a standard Internet sting. A
detective entered a chat room pretending to be a 15-year-
old girl. She received a message from the defendant ask-
2                                              No. 07-1266

ing her whether she wanted to chat. When she replied
affirmatively he engaged her in sex talk that led eventually
to his proposing that he travel to her city (which was in a
different state) to have sex with her. She agreed; he
went there and was arrested. In another chat, a copy
of which was found on his computer, he had persuaded a
12-year-old girl to agree to have sex with him, although
apparently they never did.
  The defendant was 31 years old when he committed
the crime and had no criminal record. He was a loner
who had not had sex until the previous year, with a
woman who then rejected him, breaking his heart and
(he claimed) precipitating the incidents with the 12- and
(supposed) 15-year-old girls. A forensic psychologist
named Eric Ostrov opined that the defendant had been
driven to commit the crime after suffering for years
from low self-esteem and poor body image. Dr. Ostrov
estimated that there was a 9-to-13 percent risk of the
defendant’s repeating his crime, although he thought the
risk could be reduced by counseling and psychotropic
medication. He further opined that prison would be
devastating for the defendant; he “would have almost no
resources for coping with prison life” and would be a
“target for predators.” The defendant’s father asked for
leniency for his son because of both parents’ ill health. He
also noted the defendant’s cooperation with the gov-
ernment, solid record of employment, and charitable
work. The defendant echoed his father and added that
he had lost his job and his savings (plus his car, forfeited
as an instrumentality of the crime) as a result of his con-
viction.
  The judge gave extended consideration to the mitigat-
ing factors urged by the defendant’s lawyer. But he did
No. 07-1266                                               3

not think they justified a below-guidelines sentence. He
was rather dismissive of psychological evidence, remark-
ing that persons in the defendant’s position “hire very
expensive psychotherapists, psychologists, psychiatrists,
counselors and come into court and plead with my col-
leagues in other parts of the country to have a sentence of
probation.” He stated that the kind of preying on minors,
with the aid of the Internet, that the defendant engaged in
“is a deadly, deadly, serious problem that we have as a
society,” and “there are many, many more Marias [the
detective’s stage name was Maria] out there who are
actual victims, who have the same emotional scars, per-
haps even deeper than the scars that are confronting you
on a day like today; and we as a society have found,
unfortunately, no better way to deal with this serious,
serious problem other than to incapacitate or remove
individuals from the community for a period of time
not only as a form of punishment but a form of deterrence
not only to you personally but deterrence to others who
may find themselves similarly situated.”
  These and similar remarks of the judge at sentencing
discharged his duty to consider not only the sentencing
guidelines, but also the sentencing factors set forth in
18 U.S.C. § 3553(a), the most important of which we
quote once again:
   (1) the nature and circumstances of the offense and the
   history and characteristics of the defendant;
   (2) 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;
4                                              No. 07-1266

    (B) to afford adequate deterrence to criminal conduct;
    (C) to protect the public from further crimes of the
    defendant; and
                         *   *   *
    (6) the need to avoid unwarranted sentence disparities
    among defendants with similar records who have
    been found guilty of similar conduct.
The judge balanced the history and characteristics of the
defendant, which argued for leniency, against the retribu-
tive and deterrent ends to be served by the 46-month
sentence. There was no need for him to consider ex-
plicitly the need to avoid unwarranted sentence disparities,
as there is no suggestion that the guidelines sentence
was out of line with what other defendants with similar
records who engage in the same criminal conduct receive.
The result of the judge’s balancing of competing con-
siderations was reasonable and so binds us.
  His statement that preying on minors with the aid of
the Internet is a serious problem is supported by a study
that found that approximately one in five minors per
year receives an unwanted sexual solicitation online,
while one in 33 receives a more aggressive solicitation
in which the offender asks to meet the minor or calls or
sends him or her something. David Finkelhor, Kimberly J.
Mitchell & Janis Wolak, “Online Victimization: A Report
on the Nation’s Youth” 1 (National Center for Missing and
Exploited Children June 2000), www.missingkids.com/
en_US/publications/NC62.pdf (visited Dec. 4, 2007). Only
25 percent of those solicited told a parent about it,
and only a fraction of the solicitations were reported to
authorities. 

No. 07-1266                                              5

   The defendant’s history and characteristics were rele-
vant in possibly suggesting both that imprisonment
would be a more severe punishment for him than for the
average Internet sexual predator, which would argue for
a lower prison sentence, and that he is less likely to re-
peat his crime than the average such offender, which
would also argue for a lower prison sentence. The first
point goes to the retributive (“just punishment”) and
general-deterrence factors in section 3553(a), the second
to specific deterrence (to protect the public from further
crimes by this defendant). But neither was very persuasive.
As far as we know or the defendant’s lawyer or psycho-
logist attempted to show, the average man who trolls for
young girls in Internet chat rooms is no better adjusted
than the defendant. It is true that a sexual interest in
teenage girls is not abnormal in the sense that pedophilia
is; any female of reproductive age is a natural object of
desire on the part of a male, whether in our species or any
other. But the 12-year-old girl whom the defendant
propositioned may well have been prepubescent, and he
seemed not to care; and given that sex with 12- and 15-
year-olds is deemed a very serious crime in our society,
especially when the man is much older, a 31-year-old man
who tries to seduce these young teenagers in Internet
chat rooms is unlikely to be well adjusted.
  The guidelines sentencing ranges are designed with
reference to the average offender in each crime category
to which a given range applies. United States v. Goff,

, 260-61 (3d Cir. 2007). So if a particular
defendant is average, his case for a sentence below the
range is weak. See 

As far as the record (or
our independent research) discloses, the psychological
characteristics of our defendant are average for Internet
sexual predators.
6                                                No. 07-1266

  With regard to the probability of his repeating his crime,
the absence of a comparative dimension again under-
mines the defendant’s case. Is a 9-to-13 percent probability
higher or lower than the average probability of recidivism
by persons who commit the defendant’s crime? Only if it
is lower would it be a reason for a below-guidelines
sentence. Cf. United States v. 

, 501 F.3d at 261
n. 17. For all sex offenses, “the observed sexual recidivism
rate is typically 10% to 15% after 5 years.” R. Karl Hanson
& Kelly E. Morton-Bourgon, “The Characteristics of
Persistent Sexual Offenders: A Meta-Analysis of Recidi-
vism Studies,” 73 J. Counseling & Clinical Psych. 1154 (2005).
We have not found an estimate for Internet preying on
minors, but the Sentencing Commission may not have
found one either, and so may have assumed, in choosing
the guidelines range for our defendant’s crime, a probabil-
ity of recidivism equal to that of sex offenders in gen-
eral—which is approximately the probability that
Dr. Ostrov predicted for the defendant.
  And how valid was Dr. Ostrov’s prediction, a prediction
of a specific person’s likely rate of recidivism rather
than an average rate across heterogeneous sex crimes?
Has the algorithm that Ostrov used to derive the 9-to-13
percent estimate—Static 99—been validated by meth-
ods accepted in the psychological community? Would it
satisfy the standard in Daubert or in Rule 702 of the Fed-
eral Rules of Evidence? We are not told.
   We are not even told what “Static 99” is. Of course one
can look it up, and when one does one finds that what
it does is try to match a sex offender’s characteristics
to characteristics found in studies of convicted sex offend-
ers to be correlated with recidivism, such as age (though
only whether the offender is over 25), prior sexual offenses,
No. 07-1266                                                   7

and whether he has lived with someone for at least two
years with whom he had a sexual relationship. See, e.g.,
R. Karl Hanson & David Thornton, “Static 99: Improving
Actuarial Risk Assessments for Sex Offenders” (1999),
http://ww2.ps-sp.gc.ca/publications/corrections/
199902_e.pdf (visited Dec. 4, 2007); Andrew Harris et al.,
“STATIC-99 Coding Rules (revised 2003), http://ww2.ps-
sp.gc.ca/publications/corrections/pdf/Static-99-coding-
rules_e.pdf (visited Dec. 4, 2007). Our defendant’s charac-
teristics matched those of offenders 9-to-13 percent of
whom were found to have repeated their offense.
  The methodology employed by Static 99 to predict the
probability of recidivism has been accepted in a number
of cases. See, e.g., In re Commitment of Simons, 

, 1192 (Ill. 2004); In re Detention of Thorell, 

,
726 (Wash. 2003); In re Commitment of Tainter, 

(Wis. App. 2002); In re Commitment of R.S., 

,
75 (N.J. Super. Ct. App. Div.), aff’d, 

(N.J. 2002)
(per curiam); see Eric S. Janus & Robert A. Prentky,
“Forensic Use of Actuarial Risk Assessment with Sex
Offenders: Accuracy, Admissibility and Accountability,”
40 Am. Crim. L. Rev. 1443, 1472-75 (2003). Not that it is
perfect (what is?); even its advocates claim only “moderate
predictive accuracy.” Harris et 

. It may be
more accurate than clinical assessments, Janus & 

, but that may not be saying much. Esti-
mates of recidivism are bound to be too low when one is
dealing with underreported crimes such as sex offenses.
Static 99 treats as a recidivist only someone who is
convicted of a further sex offense, but the recidivism
concern is with someone who commits a further offense,
whether or not he is caught—yet if he is not caught, his
subsequent crime does not affect the data on which the
Static 99 calibrations are based.
8                                                 No. 07-1266

  These and other problems with efforts to predict recidi-
vism from offenders’ characteristics (such as the limited
number of potentially relevant characteristics considered
by the Static 99 algorithm—education and occupation,
for example, or finer age gradations than just 25 or older,
are excluded) have engendered skepticism. Center for Sex
Offender Management, “Recidivism of Sex Offenders”
(May 2001), www.csom.org/pubs/recidsexof.html (visited
Nov. 29, 2007); Richard Wollert, “Low Base Rates Limit
Expert Certainty When Current Actuarials Are Used To
Identify Sexually Violent Predators: An Application of
Bayes’s Theorem,” 12 Psychological Pub. Policy & L. 56
(2006).
  Evidence does not have to be admissible at a trial in order
to be considered in a sentencing hearing, which is not
governed by the rules of evidence. E.g., United States v.
Goodwin, 

, 642 (7th Cir. 2007); United States v.
Birk, 

, 899 (7th Cir. 2006); United States v.
Leekins, 

, 149-50 (3d Cir. 2007); U.S.S.G.
§ 6A1.3(a). Static 99 may be a good place to start an analysis
of the likelihood of a specific defendant’s repeating his
crime—but not to end it. Robert A. Prentky et al., “Sexually
Violent Predators in the Courtroom: Science on Trial,” 12
Psychological Pub. Policy & L. 357, 384-85 (2006). In any
event, without any effort by the defendant’s lawyer to
establish the reliability of Dr. Ostrov’s methodology—or
even to explain it—the judge was entitled to discount his
prediction, United States v. Beier, 

, 574 (7th Cir.
2007). He was entitled to discount not only the estimate of
a 9-to-13 percent probability of recidivism but also the
unsubstantiated claim that the range could be reduced
by counseling; the efficacy of treatment programs for
sex offenders has not been demonstrated. Hanson &

.
No. 07-1266                                                9

  Nor did Dr. Ostrov comply with proper Static 99 proto-
col, which requires the evaluator to include with the
offender’s Static 99 score a statement as to whether the
evaluator considers the score an accurate representation of
the offender’s risk of recidivism, given characteristics of
the offender excluded from the Static 99 assessment.
“Appendix Seven: Suggested Report Paragraphs for
Communicating STATIC-99-based Risk Information,” in
Harris et 

.
  But we should consider the possible bearing on our
analysis of the Supreme Court’s decision last month in Gall
v. United States, 

(U.S. Dec. 10, 2007). The
Court held that a sentence outside the guidelines range
must not be presumed unreasonable by the appellate
court, which also may not hogtie sentencing judges with
a rigid formula for determining whether the justifica-
tion for an out-of-range sentence is “proportional” to the
extent of the sentence’s deviation from the range. Neither
approach had been followed by this court. Even before
Rita v. United States, 

(2007), our court,
anticipating Gall in this respect, rejected the notion that a
sentencing judge can presume the reasonableness of a
sentence within the guidelines range—rejected it emphati-
cally. “The judge is not required—or indeed permitted—to
‘presume’ that a sentence within the guidelines range is the
correct sentence and if he wants to depart give a reason
why it’s not correct. All he has to do is consider the guide-
lines and make sure that the sentence he gives is within
the statutory range and consistent with the sentencing
factors listed in 18 U.S.C. § 3553(a). His choice of sen-
tence, whether inside or outside the guideline range, is
discretionary and subject therefore to only light appellate
review. The applicable guideline nudges him toward the
10                                                No. 07-1266

sentencing range, but his freedom to impose a reasonable
sentence outside the range is unfettered.” United States v.
Demaree, 

, 794-95 (7th Cir. 2006) (citations
omitted, emphasis added); see also United States v. Griffin,

, 868 (7th Cir. 2007).
  With specific reference to out-of-range sentences, we
had said, again before the Supreme Court’s decision in Gall,
that “when the guidelines, drafted by a respected public
body with access to the best knowledge and practices of
penology, recommend that a defendant be sentenced to a
number of years in prison, a sentence involving no (or, as
in this case, nominal) imprisonment can be justified only by
a careful, impartial weighing of the statutory sentencing
factors.” United States v. Goldberg, 

, 673 (7th Cir.
2007). We reversed in Goldberg because we found the
district judge’s weighing of the statutory sentencing factors
that guide the judge in deciding whether to give a sentence
outside the guidelines range to have been “unreasonable,”
id.—the standard later announced in Gall. While disapprov-
ing sentencing presumptions and rigid formulas, the
Supreme Court made “clear” in Gall, as we had tried to do
in Goldberg, “that a district judge must give serious consid-
eration to the extent of any departure from the Guidelines
and must explain his conclusion that an unusually lenient
or an unusually harsh sentence is appropriate in a particu-
lar case with sufficient justifications.” Gall v. United 

, at *6. “[A] major departure should be supported
by a more significant justification than a minor one.” 

  The judge in this case gave a guidelines sentence, and
since, as we have explained, his decision not to give a
sentence beneath the guidelines range was reasonable,
there is no basis in Gall for invalidating the sentence.
No. 07-1266                                                   11

Indeed, the emphasis that the Supreme Court in Gall placed
on sentencing discretion (see, e.g., 

in the course of
rejecting the tight girdle (amounting in the Supreme
Court’s view to de novo review, see 

that the
Eighth Circuit had placed on that discretion, reinforces
our conclusion that the sentence meted out in this case,
because it is reasonable, must be
                                                     AFFIRMED.
A true Copy:
        Teste:

                           _____________________________
                           Clerk of the United States Court of
                             Appeals for the Seventh Circuit




                     USCA-02-C-0072—1-10-08