In the
United States Court of Appeals
For the Seventh Circuit
____________________
No. 15‐2862
UNITED STATES OF AMERICA,
Plaintiff‐Appellee,
v.
CRANE MARKS,
Defendant‐Appellant.
____________________
Appeal from the United States District Court for the
Northern District of Illinois, Eastern Division.
No. 14 CR 330‐9 — Virginia M. Kendall, Judge.
____________________
ARGUED JUNE 7, 2017 — DECIDED JULY 24, 2017
____________________
Before RIPPLE, ROVNER, and HAMILTON, Circuit Judges.
HAMILTON, Circuit Judge. This appeal illustrates why, when
an arcane and arbitrary issue arises under the Sentencing
Guidelines, the sentencing judge should ask, “Why should I
care?” That question was not asked here, so we must remand
for resentencing.
Defendant‐appellant Crane Marks pled guilty to conspir‐
ing to distribute heroin. He was sentenced to nine years in
2 No. 15‐2862
prison (108 months). That sentence is either well above or well
below the advisory range under the Sentencing Guidelines,
depending on one issue in calculating Marks’ criminal history.
The district court decided the issue against Marks, and he has
appealed his sentence.
In all candor, that one issue seems astonishingly technical
and trivial. It has nothing to do with Marks’ culpability or the
larger goals of sentencing. As we explain below, the issue is
whether, when Marks was imprisoned on his fourth state
drug conviction in 2000, he also had his state parole revoked
on any of his earlier state drug convictions and was re‐impris‐
oned on that revocation as well. From this description of the
issue, we hope readers will agree that this is one of those
guideline issues that should prompt the sentencing judge to
ask why the judge or anyone else should care about the an‐
swer.
Because the issue seems so technical and trivial, we have
examined the record in this case for any signs that the judge
would have given Marks the same sentence regardless of how
the technical criminal history issue was resolved. We found
no such signs, however, so we have considered the technical
guideline issue on the merits. We conclude that the court
made both a legal error and a factual error. The legal error was
that the court did not make the finding needed to treat Marks
as a career offender under the Guidelines. The factual prob‐
lem is that the court was not presented with reliable evidence
from which it could have found that Marks was imprisoned
on a revocation of parole on any earlier conviction. That
means that Marks does not qualify, technically, as a career of‐
fender. His advisory guideline sentencing range is lower than
No. 15‐2862 3
the range found by the district court. We therefore vacate
Marks’ sentence and remand for resentencing.
To explain, in October 2013, Marks joined a conspiracy to
distribute heroin in violation of 21 U.S.C. §§ 846 & 841(a)(1).
In 2015 he pled guilty to that crime. At the time of plea nego‐
tiations, the government knew that Marks had been convicted
of controlled‐substance offenses in state court in 1994, 1995,
1996, and 2000. The government did not then believe Marks
was a career offender under U.S.S.G. § 4B1.1 because of the
age of the offenses from the 1990s. For the career‐offender
Guideline to apply, at least two of those four convictions
would need to count for criminal history points. See U.S.S.G.
§ 4B1.1(a)(3). For a conviction to be counted, Marks had to
have been still serving a sentence of imprisonment on that
conviction within 15 years of his commencement of the of‐
fense of conviction in this case. See U.S.S.G. § 4A1.2(e)(1).
Since Marks commenced his involvement in the conspir‐
acy in October 2013, a prior conviction counts under the
Guidelines only if he was incarcerated on the conviction after
October 1998. The 2000 conviction definitely counts. The
question is whether any of the 1990s convictions count. The
parties agree that Marks was paroled on all of the 1990s of‐
fenses in June 1998. But if his parole was revoked for any of
the 1990s offenses when he committed the new drug crime in
2000, the term of incarceration on the revocation would fall
within the 15‐year window, the conviction would count, see
§ 4A1.2(k)(2)(A), and Marks would be treated properly as a
career offender under § 4B1.1(a).
In their plea agreement, the parties acknowledged that
Marks would receive criminal history points only for the 2000
conviction, not the earlier offenses. On that understanding
4 No. 15‐2862
they projected a guideline prison range of 51 to 63 months.
The probation officer who wrote the presentence report came
to a different conclusion. Based on handwritten records from
the Illinois Department of Corrections, the officer concluded
that Marks’ imprisonment in 2000 was based not only on his
2000 conviction but also on revocation of his parole for one or
more of the earlier convictions.
There is no formal court order revoking parole. The IDOC
records, which are appended to the presentence report, do not
include any narrative explaining whether Marks’ parole was
revoked or what type of custody he was in at any particular
time. The three records read in their entirety as follows:
Case #94CR7913 Case #95CR21777
Offense Date: 2‐21‐94 Offense Date: 6‐22‐95
Date Sentenced: 8‐15‐95 Date Sentenced: 8‐15‐95
Admitted to IDOC: 8‐18‐95 Admitted to IDOC: 8‐18‐95
Escape Out: 2‐21‐96 Escape Out: 2‐21‐96
Escape Return: 2‐28‐96 Escape Return: 2‐28‐96
Escape Out: 4‐4‐96 Escape Out: 4‐4‐96
Escape Return: 7‐26‐96 Escape Return: 7‐26‐96
Discharge of Sentence: 5‐22‐00 Discharge of Sentence: 5‐22‐00
Case #96CR1549301
Offense Date: 5‐7‐96
Date Sentenced: 7‐23‐96
Admitted to IDOC: 7‐26‐96
Parole Out: 6‐26‐98
Violator Return: 2‐8‐00
Discharge of Sentence: 6‐26‐00
These records say that Marks’ sentences for the 1994 and 1995
convictions were “discharged” on May 22, 2000 and that the
sentence for the 1996 conviction was “discharged” on June 26,
2000. The records for the 1994 and 1995 conviction do not list
No. 15‐2862 5
a parole date. In fact, they do not show that Marks was ever
in prison on those two convictions. Other information in the
presentence report shows that he was on electronic monitor‐
ing—not in prison—when he escaped and returned to cus‐
tody. The record for the 1996 conviction, however, lists a “Pa‐
role Date” in 1998 and a “Violator Return” date of February 8,
2000, the same day Marks pled guilty to the 2000 offense.
Based solely on these records, the probation officer inferred
that all four controlled‐substance convictions counted under
the federal Guidelines so that Marks qualified as a career of‐
fender.
At sentencing Marks disputed the career‐offender issue.
He asserted that he did not recall his parole ever being re‐
voked on any of the 1990s convictions. Without proof of a pa‐
role revocation that returned him to prison on one of these
convictions, Marks argued, they do not fall within the 15‐year
window under § 4A1.2(k)(2)(A).
The district court did not make an explicit finding about
whether Marks had been imprisoned again on a parole revo‐
cation of any of the 1990s convictions. The court concluded,
though, that each conviction counted because the sentences
were not “discharged” until 2000. On that basis, the court
found that Marks was a career offender and that his impris‐
onment range was 151 months to 188 months. In pronouncing
sentence, however, the court said that Marks did not fit the
profile of the typical career offender. The court sentenced him
to 108 months—43 months below the low end of the calcu‐
lated range, but 45 months above the top of the guideline
range without the career‐offender finding.
On appeal, Marks renews his argument that the district
court erred by classifying him as a career offender because his
6 No. 15‐2862
1994, 1995, and 1996 convictions are too old to count. He con‐
tends that the government did not meet its burden of showing
that he was sent back to prison on a revocation of his parole
on any of those offenses when he committed a new drug
crime in 2000. If anything, Marks asserts, the IDOC records
prove only that he was in state custody at the time his parole
expired on the 1990s convictions. The records do not prove,
he argues, that he was actually imprisoned again on any of
those convictions.
The government argues that the district court reasonably
relied on the statement in the presentence report that Marks’
parole had been revoked. The government contends that the
probation officer’s statement in the report is proven by the
three IDOC records, which, in the government’s view, “clearly
state” that Marks was “in custody some time during the Oc‐
tober 1998 through October 2013 window.” That argument
misses the point. Marks was in custody during that window,
but under the text of the Guidelines, for any one of the 1990s
convictions to count, Marks must have been incarcerated on
that conviction, not just on his 2000 conviction. After June 1998
he could not have been serving prison time on any of the
1990s convictions unless his parole on that conviction was re‐
voked.
For two reasons, one legal and one factual, we cannot sus‐
tain the district court’s application of the career‐offender
Guideline, § 4B1.1. First, the transcript of the sentencing hear‐
ing indicates substantial legal confusion about which guide‐
line provision was being applied and what factual findings
were necessary. Second, the government failed to offer suffi‐
cient evidence to show as a matter of fact that Marks’ parole
was revoked on at least one of his 1990s convictions.
No. 15‐2862 7
To the first point, the district court and the parties were
not on the same page about what is required to count an older
conviction under § 4A1.2(k)(2), which explains how and
when to count a prior offense when parole has been revoked
on it. The government had to show that Marks’ parole on one
of the three older convictions was revoked and that he was re‐
imprisoned based on that revocation. See United States v.
Moreno‐Padilla, 602 F.3d 802, 808 (7th Cir. 2010). The probation
officer testified at sentencing that an IDOC administrator had
told him Marks was in IDOC custody “all the way up until
some time in 2000” when his parole terms expired.
The broad concept of “custody” is not enough under
§ 4A1.2(k)(2). The focus is “incarceration.” Proving that
Marks’ parole terms did not expire until 2000 was not
enough—the government had to show that Marks was incar‐
cerated on at least one of those convictions. See § 4A1.2(b)(1)
(“sentence of imprisonment” means “a sentence of incarcera‐
tion”). And under Illinois law, a person remains in IDOC cus‐
tody even while serving a parole term outside of prison. See
730 ILCS 5/3‐14‐2(a); Cordrey v. Prisoner Review Bd., 21 N.E.3d
423, 429 (Ill. 2014). So it matters what type of custody Marks
was in for his 1990s convictions between October 1998 and
2000—in a prison or in some other, less restrictive form of cus‐
tody. The IDOC records show that Marks remained under
IDOC control—in its “custody”—on the 1990s convictions un‐
til May or June 2000.
The problem is that those records do not show incarcera‐
tion, which is what matters under the federal Guidelines. In
short, to apply the career‐offender Guideline properly, the
district court had to decide first whether Marks’ parole on any
of the 1990s convictions was revoked, and second whether he
8 No. 15‐2862
was sent back to prison because of such a revocation. It was a
legal error to find that Marks was a career offender without
making these findings. See United States v. Titus, 821 F.3d 930,
934 (7th Cir. 2016).
The second problem, the factual one, is a distinct problem
with the district court’s guideline calculation. As we view the
record, the district court lacked reliable evidence to answer
these questions so as to support application of the career‐of‐
fender Guideline.
As a general rule, a sentencing judge may rely on a presen‐
tence report if it “is well‐supported and appears reliable.”
United States v. Black, 636 F.3d 893, 897 (7th Cir. 2011); see
United States v. Heckel, 570 F.3d 791, 795 (7th Cir. 2009); United
States v. Mustread, 42 F.3d 1097, 1101–02 (7th Cir. 1994). But if
a presentence report contains “nothing but a naked or unsup‐
ported charge,” the defendant’s denial will suffice to call the
report’s accuracy into doubt. Mustread, 42 F.3d at 1102. Simi‐
larly, if the presentence report “omits crucial information,
leaving ambiguity on the face of that document,” the govern‐
ment has the burden of independently demonstrating the ac‐
curacy of the report. Moreno‐Padilla, 602 F.3d at 809.
The government emphasizes that Marks did not produce
any evidence to prove that his parole had not been revoked.
The government cites Moreno‐Padilla, 603 F.3d at 807–09, a fac‐
tually similar case in which the defendant challenged an as‐
sertion in the presentence report that his parole on an earlier
conviction had been revoked within 15 years of his commis‐
sion of the offense of conviction. In Moreno‐Padilla we re‐
peated our familiar admonition that, when faced with a
presentence report that is well‐supported and facially reliable,
No. 15‐2862 9
the defendant cannot stand by and do nothing. He must in‐
stead show the judge that disputed factual information in the
report is inaccurate. 602 F.3d at 808–09.
Moreno‐Padilla does not say, however, that any factual as‐
sertion that finds its way into a presentence report may be ac‐
cepted uncritically by the sentencing judge. Rather, we
pointed out that statements about the defendant’s parole hav‐
ing been revoked were corroborated by other details. The de‐
fendant had been in a state prison, not a local jail, for two
months before pleading guilty to a new charge. That sug‐
gested that the period of incarceration was for a parole revo‐
cation rather than pretrial detention on the new charge. Also,
the defendant’s parole term on the earlier conviction extended
beyond the new term of incarceration. Id. at 805, 809.
Marks, on the other hand, was not sent back to prison until
after he pled guilty to the new offense in 2000. There is no
indication that any of his 1990s sentences were extended be‐
yond what they would have been if the 2000 offense had not
occurred. Rather, his sentences on the 1990s convictions all
terminated in 2000 even though he was not paroled on the
2000 offense until 2002.
In short, Marks did not have the burden of producing con‐
trary evidence because this assertion in the presentence report
was not reliable on its face. When read as a whole—the text as
well as the underlying IDOC records incorporated as exhib‐
its—the presentence report “omits crucial information, leav‐
ing ambiguity on the face” of the report concerning whether
Marks’ parole on any of the 1990s convictions had been re‐
voked and he had been imprisoned on any of them. See
Moreno‐Padilla 602 F.3d at 809 (“If a defendant’s parole was
10 No. 15‐2862
not revoked, he could not have served time for a parole revo‐
cation.”); United States v. McNeil, 573 F.3d 479, 483–84 & n.4
(7th Cir. 2009) (district court plainly erred by not ordering
probation office to supplement presentence report where
omission of crucial information created uncertainty about
how federal sentence should be imposed). Indeed, the proba‐
tion officer’s testimony showed that he asked the IDOC ad‐
ministrator the wrong question: whether Marks was “in cus‐
tody” rather than whether he was incarcerated on parole rev‐
ocation for the 1996 conviction. That discrepancy undermines
the government’s reliance on the records to prove that Marks
was imprisoned on a parole revocation.
As for the 1994 and 1995 convictions, the IDOC records
demonstrate only that Marks’ sentences were discharged on
May 22, 2000. The records say nothing about a parole revoca‐
tion—ever—or a return to prison on any date after 1996, let
alone on February 8, 2000, the date he was imprisoned on the
2000 offense. In fact, the records for these two convictions
seem incomplete. Each report lists the date Marks was first
admitted to IDOC custody, the dates he escaped from and
was returned to custody, and the date his sentence was dis‐
charged. Neither report explains what type of custody Marks
was in or gives a parole or release date. Marks obviously was
released from prison at some point between the date he was
admitted into IDOC’s custody in 1995 (on the same date for
both convictions) and the date his sentences were discharged,
because he was arrested twice in 1999 while outside of prison.
And the presentence report explains that Marks was on elec‐
tronic monitoring (not imprisoned) when he “escaped” IDOC
custody twice in 1996.
No. 15‐2862 11
Taken at face value, these records fail even to prove that
Marks’ sentences on these convictions ever met the guideline
definition of “imprisonment,” let alone that he was reincar‐
cerated in 2000 due to a parole revocation. This court and
most other circuits hold that sentences served in community
treatment centers, halfway houses, home detention, or other
forms of probation are not imprisonment for guideline pur‐
poses. See, e.g., United States v. Timbrook, 290 F.3d 957, 959–60
(7th Cir. 2002) (noting that Guidelines do not elaborate on def‐
inition of incarceration but explaining that residence in non‐
secure community treatment center or halfway house is not
imprisonment for guideline purposes); United States v. Phipps,
68 F.3d 159, 162 (7th Cir. 1995) (home detention not imprison‐
ment for guideline purposes); United States v. Gordon, 346 F.3d
135, 138–39 (5th Cir. 2003) (same); United States v. Pielago, 135
F.3d 703, 711–14 (11th Cir. 1998) (community treatment center
not imprisonment for guideline purposes); United States v.
Jones, 107 F.3d 1147, 1161–65 (6th Cir. 1997) (home detention
not imprisonment for guideline purposes); United States v.
Latimer, 991 F.2d 1509, 1514 (9th Cir. 1993) (community treat‐
ment center not imprisonment for guideline purposes); but
see United States v. Rasco, 963 F.2d 132 (6th Cir. 1992) (holding
that confinement in community treatment center after parole
revocation was part of original term of imprisonment).
The IDOC record for the 1996 conviction also does not
support the needed finding. Unlike the other two records, it
at least includes the word “parole.” But it indicates only that
Marks was returned to custody in February 2000, not why he
was returned to custody. The document lists Marks’ “Pa‐
role Out” date as June 26, 1998, his “Violator Return” date as
February 8, 2000, and his “Discharge of Sentence” date as June
12 No. 15‐2862
26, 2000. While the term “Violator Return” could refer to a vi‐
olation of parole, that meaning is neither explicit nor obvious.
As Marks points out, the records do not contain any nar‐
rative showing that he was given a new term of imprisonment
for violating parole, or whether he was merely noted as being
in custody on a potential parole violation. The fact that Marks’
sentence on this 1996 conviction (as well as the 1994 and 1995
convictions) was discharged only a few months after he pled
guilty to the 2000 charge suggests that no revocation oc‐
curred. And it is difficult to understand why, if Marks’ parole
was actually revoked, the government could not have sup‐
ported the presentence report with a copy of the order of rev‐
ocation. See McNeil, 573 F.3d at 483 (explaining that “gather‐
ing such information is one of the principal tasks of a proba‐
tion officer in preparing a PSR”). In its brief and oral argu‐
ment, the government has not suggested that it obtained any
other documents supporting its contention that a revocation
occurred.
So, coming up for air from the minutiae of the IDOC rec‐
ords, we must conclude that the district court’s decision to
treat Marks as a career offender under the Guidelines was an
error. And despite the sometimes arbitrary limits in the calcu‐
lations, the guideline range must be properly calculated. Gall
v. United States, 552 U.S. 38, 51 (2007); United States v. Robinson,
435 F.3d 699, 701 (7th Cir. 2006). An incorrect calculation of
the Sentencing Guidelines is a procedural error that we pre‐
sume influenced the judge’s choice of sentence, unless the
judge said otherwise. See United States v. Adams, 746 F.3d 734,
743 (7th Cir. 2014); see generally, Molina‐Martinez v. United
States, 578 U.S. —, —, 136 S. Ct. 1338, 1347–48 (2016) (even
unpreserved error in calculating guideline range is presumed
No. 15‐2862 13
to affect defendant’s substantial rights, at least if sentencing
court did not suggest it would have imposed same sentence
absent the error).
With all that said, we return to the question we raised at
the outset. Why should anyone care? Marks came before the
district court with four prior convictions for drug felonies.
Whether or not state officials took the formal step, when
Marks was imprisoned on the fourth drug conviction in 2000,
of revoking his parole on any of his three prior drug convic‐
tions has no apparent connection to the purposes of sentenc‐
ing: reflecting the seriousness of the offense, promoting re‐
spect for the law, providing just punishment for the offense,
affording adequate deterrence of criminal conduct, protecting
the public from further crimes by Marks, and providing him
with needed correctional treatment. See 18 U.S.C. § 3553(a)(2).
If the sentencing court had asked why anyone should care,
it might have explained at the time of sentencing that the an‐
swer to the technical guideline question simply did not matter
to the court’s final sentencing decision. That did not happen
here, though. The court did not say that it would have im‐
posed the same sentence whether or not Marks technically
qualified as a career offender under the Guidelines.
The fact that the Guidelines present some technical and ar‐
bitrary questions, as in this case, does not show a flaw in the
Guidelines as a whole. They are a comprehensive and de‐
tailed system. They try to capture a broad range of factors rel‐
evant to sentencing. Such a comprehensive and detailed sys‐
tem of advice that takes into account the seriousness of crim‐
inal conduct and the defendant’s criminal record must neces‐
sarily include some arbitrary rules and cutoffs. One example
is the 15‐year requirement in § 4A1.2(e), which can be even
14 No. 15‐2862
more arbitrary as applied to parole revocations under
§ 4A1.2(k)(2).
The sometimes arbitrary aspects of criminal history calcu‐
lations have long been recognized by the Sentencing Commis‐
sion itself. Even the original 1987 version of the Guidelines
encouraged departures based on the often‐arbitrary features
of the criminal history calculations. See U.S.S.G. § 4A1.3
(1987). Both when the Guidelines were treated as binding and
now when they are advisory, the presence of such a technical
and arbitrary issue should prompt a sentencing judge to ask
him‐ or herself why anyone should care.
Even if the 1990s drug convictions are too old to count for
guideline purposes, the court may still consider them in eval‐
uating Marks’ criminal history under § 3553(a). See
§ 4A1.3(a); 18 U.S.C. § 3661. In the end, Marks’ extensive crim‐
inal history may justify a sentence above the range that ap‐
plies without the career‐offender Guideline (51 to 63 months),
but that is a question for the district court.
We VACATE Marks’ sentence and REMAND for re‐sen‐
tencing consistent with this opinion.