Defendants Carl Stewart, Kevin Ferguson, and Darryl Wayne Hall were charged in a fourteen count indictment with drug-related crimes involving crack cocaine. A jury convicted Ferguson and Stewart of conspiracy to distribute and possess with the intent to distribute cocaine base, in violation of 21 U.S.C. § 846, and distribution of cocaine base, in violation of 21 U.S.C. § 841. The jury convicted Hall of distribution of cocaine base, in violation of 21 U.S.C. § 841, but acquitted him on the conspiracy charge. All three now appeal a multitude of issues. We affirm.
I. Background
Law enforcement agents conducted an extensive investigation of the Marion, Illinois area cocaine industry, in which they recruited a large number of informants who made controlled buys from the defendants. The evidence at defendants’ joint trial consisted mostly of testimony from these informants and from co-defendants who were not tried with the defendants because they pled guilty. Because of the number of issues raised, we only briefly summarize the facts here. We enter into more detail as necessary when addressing defendants’ arguments below.
The evidence at trial revealed that Stewart1 and Ferguson were joint venturers in the business of crack cocaine. Stewart was Ferguson’s “right-hand man.” The two distributed drugs from three primary locations: Ferguson’s residence on Griggs Street, Stewart’s residence on Monroe Street, and a neighborhood park. Ferguson was one of the main suppliers in the Marion, Illinois area. Ferguson funneled the majority of his drugs through Stewart and Stewart relied on Ferguson for his supplies. Ferguson also had others distribute for him, including juveniles. Witnesses testified to purchasing directly from both Stewart and Ferguson, as well as from others who obtained the drugs from one of these two. Based on this evidence, Ferguson and Stewart were convicted by a jury of conspiracy to distribute and possess with the intent to distribute cocaine base, as well as several counts of distribution of cocaine base. Ferguson and Stewart were sentenced to 360 and 216 months in prison, respectively.
In contrast, defendant Hall was acquitted of the conspiracy charge and convicted only *1231of one count of distribution. This conviction was based on the following controlled buy. On March 27, 1995, Hall was at Stewart’s residence on Monroe Street. There he met Cynthia Baldwin, a government informant who was looking for Stewart so that she could purchase some drugs. Hall told her that Stewart was busy. Instead of retrieving Stewart, Hall took Baldwin’s money, obtained the drugs from Stewart himself, and delivered them to Baldwin. In exchange for this service, Hall asked for a bit of the crack cocaine, which Baldwin gave him. Baldwin did not know Hall at the time, but later identified him from a spread of five photographs. After his conviction, Hall was sentenced to 210 months in prison.
All three defendants appeal their convictions and sentences. Because Ferguson and Stewart present each issue jointly, we will address the issues in the same manner. After addressing the joint arguments of Ferguson and Stewart, we address those raised by Hall.
II. Ferguson & Stewart
1. Evidence of Gang Affiliation
In response to a question on cross examination by Stewart’s counsel, a government witness testified that Ferguson was making gang signs in a photograph. After this statement, Stewart’s counsel continued on the subject of gangs in an apparent attempt to show that the witness was not qualified to identify gang signs.2 The photo was later admitted into evidence. As an initial matter, Stewart has no ground for objection to this evidence, since it was his counsel that brought out the gang affiliation testimony. See United States v. Fulford, 980 F.2d 1110, 1116 (7th Cir.1992) (“It is well-settled that where error is invited, not even plain error permits reversal.”). However, for the same reason that this evidence does not merit a new trial for Ferguson, it would not merit a new trial for Stewart even if he could object despite the circumstances.
Defendants first argue that this evidence of gang affiliation violated their First Amendment rights to free association. See Dawson v. Delaware, 503 U.S. 159, 112 S.Ct. 1093, 117 L.Ed.2d 309 (1992). Second, they argue that the judge abused his discretion under Federal Rule of Evidence 403 when he admitted the evidence because it was overly prejudicial. We need not address whether either error actually exists, because we find the admission of this evidence to be harmless. We note that Justice Blackmun, in his concurrence in Dawson, 503 U.S. at 167-71, 112 S.Ct. at 1099-1100, questioned whether harmless error analysis was appropriate when gang affiliation evidence violated a defendant’s First Amendment rights, given the *1232potential chilling effect use of such evidence could have. However, because any such chilling effect is minimal when the gang affiliation evidence was introduced by the defense, as it was here, we find harmless error analysis to be appropriate. Accord People v. Shatner, 174 Ill.2d 133, 220 Ill.Dec. 346, 356, 673 N.E.2d 258, 268 (1996) (finding that Dawson court held harmless error analysis to be applicable to erroneous introduction of gang affiliation evidence); Dawson v. State, 608 A.2d 1201, 1203 (Del.1992) (same).
We should allow the conviction to stand despite error “if we are convinced that the error did not have ‘a substantial and injurious effect or influence on the jury’s verdict.’ ” United States v. Irvin, 87 F.3d 860, 866 (7th Cir.) (quoting United States v. Hanson, 994 F.2d 403, 407 (7th Cir.1993)), cert. denied, — U.S. -, 117 S.Ct. 259, 136 L.Ed.2d 184 (1996); see also Fulford, 980 F.2d at 1117 (applying harmless error analysis in this context). The references to gang membership were minimal, occupying no more than two pages of an over 800 page trial transcript. Further, they were made only at the instigation of defense counsel. Defendants do not contest that the government made no argument based on gang affiliation. See Irvin, 87 F.3d at 867 (holding extensive gang evidence and prosecutor’s closing argument based on gang membership harmless when defendant himself testified to gang membership and evidence was overwhelming). Given that the government did not use gang affiliation as any part of its case — either as direct or indirect evidence— we think it exceedingly unlikely that the jury found Ferguson guilty purely by virtue of association. Thus, even if the admission of this testimony was error, it was harmless.
2. Sufficiency of the Evidence of Conspiracy
Sufficiency of the evidence challenges will be rejected if any rational jury could have found defendants guilty beyond a reasonable doubt. United States v. Durrive, 902 F.2d 1221, 1225 (7th Cir.1990). When defendants challenge their convictions on sufficiency of the evidence grounds, we view the evidence and all reasonable inferences in the light most favorable to the government. Id.; United States v. Monroe, 73 F.3d 129, 131 (7th Cir.1995); United States v. Rodriguez, 925 F.2d 1049, 1051-52 (7th Cir.1991). There must be substantial evidence supporting the conclusion that Stewart and Ferguson were involved in the conspiracy, Durrive, 902 F.2d at 1229, however, that evidence may be wholly circumstantial. United States v. Yusufu, 63 F.3d 505, 508 (7th Cir.), cert. denied, — U.S. -, 116 S.Ct. 578, 133 L.Ed.2d 501 (1995); United States v. Kellum, 42 F.3d 1087, 1091 (7th Cir.1994).
Defendants argue that the government failed to prove conspiracy. According to defendants, it proved no more than a buyer-seller relationship. To prove conspiracy, the government must prove an agreement to commit a crime separate from the mere purchase or sale of the drug. United States v. Lechuga, 994 F.2d 346, 347 (7th Cir.1993). That agreement can be an implicit understanding between the parties regarding the subsequent resale of drugs. See United States v. Clay, 37 F.3d 338, 341 (7th Cir. 1994). The jury may infer such an understanding, and thus the existence of a conspiracy, from the course of dealings between the alleged conspirators. See id.; see also Lechuga, 994 F.2d at 350 (“Prolonged cooperation is neither the meaning of conspiracy nor an essential element, but it is one type of evidence of an agreement' that goes beyond what is implicit in any consensual undertaking, such as a spot sale.”). Routine sales for resale can create a mutual interest between buyer and seller in maintaining their continuing relationship that extends beyond the moment of each individual sale. Clay, 37 F.3d at 341. Thus, “a prolonged and actively pursued course of sales coupled with the seller’s knowledge of and a shared stake in the buyer’s illegal venture is sufficient to sustain a finding of conspiracy.” Id.
The government bases its argument that a conspiracy existed on the continuing relationship between Stewart and Ferguson. The evidence indicates that there existed more than just a series of disinterested sales between the two. Instead, Stewart and Ferguson were engaged in a cooperative effort. In addition to the continuing sales between Stewart and Ferguson, there was evidence *1233that Stewart relied on Ferguson to supply him for his sales to others. For example, when Stewart was out of drugs he would indicate he was waiting for Ferguson to supply him with more. On one occasion when a witness wanted to buy crack from Stewart, Stewart sent for Ferguson who then brought him some crack to sell to the witness. There was also testimony that Ferguson would refer buyers to Stewart. Most importantly, a witness testified that on one occasion Ferguson and Stewart travelled together to Tennessee where they purchased and split a significant amount of crack. In the words of one witness, Stewart was Ferguson’s “right-hand man.” The evidence fully supports the inference of an interdependent relationship between Stewart and Ferguson. Reading the evidence and drawing all inferences in the light most favorable to the government, Yusufu, 63 F.3d at 508, the evidence is sufficient to support the conclusion that there existed not just a mere buyer-seller relationship, but a conspiracy between Stewart and Ferguson to distribute crack cocaine.
3. Evidence of Drug Quantity
Defendants Stewart and Ferguson next challenge their sentences, claiming there was insufficient proof of the drug quantities attributed to them. Unfortunately, their brief on this issue repeatedly refers to “defendant” without specifying which defendant, thus failing to make clear which quantities they challenge as to whom. Because the brief only develops this argument as to Ferguson, the undeveloped argument challenging the drug quantities attributed to Stewart is waived. United States v. Berkowitz, 927 F.2d 1376, 1384 (7th Cir.1991) (“We have repeatedly made clear that perfunctory and undeveloped arguments ... are waived----”). The best argument we can glean is that Ferguson challenges the drug quantity as based on unreliable information. He argues that three witnesses in particular were not credible: Harper, Dismore and Thompson. He specifically challenges the calculation of 438 grams attributed to him based on a calculation in the presentence report (“PSR”).
The drug quantity for sentencing purposes must be proven by a preponderance of the evidence. United States v. Howard, 80 F.3d 1194, 1202 (7th Cir.1996). The district court’s calculation of the quantity of drugs involved in an offense is a finding of fact to be reversed only for clear error. Id.; United States v. Acosta, 85 F.3d 275, 279 (7th Cir.1996). We must affirm unless we are “left with the definite and firm conviction that a mistake has been committed.” United States v. Beler, 20 F.3d 1428, 1431 (7th Cir. 1994) (internal quotations omitted). Ferguson’s arguments are based primarily on the fact that the challenged witnesses were drug addicts and informants and thus the district court should have scrutinized their reliability with particular care. Beler, 20 F.3d at 1435. However, Ferguson points to nothing that would indicate to us that the district court failed to adequately assess these witnesses’ reliability. The district court had the benefit of viewing Harper and Dismore at trial, including their cross examinations. It also made further inquiry at sentencing into the statements made by Thompson to the agent who interviewed him. We are satisfied that the district court adequately assessed the reliability of this evidence and “[i]t is well established that a sentencing court’s credibility determinations are accorded exceptional deference.” Acosta, 85 F.3d at 280.
Ferguson also challenges the PSR’s extrapolation of 438 grams attributed by Thompson based on an equation of $20 rocks of crack to 0.1 grams. As long as the information is reliable, the district court is entitled to rely on the PSR, as it did here. When the district court adopts the PSR’s findings, the defendant must offer more than a bare denial of its factual allegations to mount a successful challenge. United States v. Taylor, 72 F.3d 533, 547 (7th Cir.1995); United States v. Mumford, 25 F.3d 461, 467 (7th Cir.1994). In this ease, Ferguson offered no more than a bare denial of the drug quantity information in the PSR and he did not submit any contrary evidence at the sentencing hearing. Moreover, this calculation was supported by the testimony of Deputy *1234Sheriff Roland at the sentencing hearing. Compare Howard, 80 F.3d at 1204 (vacating sentence based on similar calculation when there was no evidence as to how many ounces a $20 rock of crack cocaine constituted). We find that the government proved by a preponderance of the evidence that Ferguson is accountable for the 438 grams attributed by Thompson. Thus, having reviewed the record, we do not think the district court committed clear error in calculating the drug quantity involved in Ferguson’s offense.
4. Position of Leadership
The district court increased Ferguson’s offense level by four for his leadership role under § 3Bl.l(a) of the Sentencing Guidelines. Ferguson first argues that the district court did not state with sufficient specificity the basis for its finding that he took on a leadership role.3 We review a district court’s factual findings regarding a defendant’s role in the offense for clear error. United States v. Lewis, 79 F.3d 688, 690 (7th Cir.1996). Ferguson cites to United States v. Jewel, 947 F.2d 224, 235-36 (7th Cir.1991), for the proposition that the court must state with specificity its basis for applying the § 3Bl.l(a) role enhancement. Jewel is distinguishable, however, because the court in that case rejected the PSR’s recommendation as to what role enhancement was appropriate. Id. at 235. In contrast, after stating its conclusion that Ferguson qualified for the aggravating role enhancement, the court in the instant case stated it was adopting the findings in the PSR with only one exception not relevant to this issue. Although Federal Rule of Criminal Procedure 32(c) generally requires the district court to make explicit findings with regard to factual matters in the PSR disputed by the defendant, no “magic words” are required, as long as the court actually resolves the disputed issues. United States v. McKinney, 98 F.3d 974, 982 (7th Cir.1996), cert. denied, — U.S. -, 117 S.Ct. 1119, 137 L.Ed.2d 319 (1997). Where the defendant objects but does not offer any evidence of the PSR’s inaccuracy, the rule that the court must make findings as to disputed issues can be satisfied by reference to the PSR. Id. In this respect the instant case is very similar to McKinney, 98 F.3d at 982. As in McKinney, the court in the instant case heard all the testimony and other evidence at trial, it relied on the PSR, and it heard additional evidence when Deputy Sheriff Roland testified in support of the PSR’s findings at the sentencing hearing. In contrast, Ferguson offered no contradictory evidence at this hearing. Thus, even without explicit findings as to the leadership role, we may affirm if the record supports the enhancement. Id.; see also United States v. Carson, 9 F.3d 576, 585 (7th Cir.1993).
Ferguson’s second argument challenging the enhancement for his role in the offense is that there was insufficient evidence of control or supervision; rather he was no more than a seller in buyer-seller relationships. To receive a four-level enhancement under § 3Bl.l(a), the defendant must be “an organizer or leader of a criminal activity that involved five or more participants or was otherwise extensive.” This circuit has held that to merit a four-point enhancement for one’s leadership role under § 3Bl.l(a) when the criminal activity involved five or more participants, the defendant must have had control, direct or indirect, over at least four other participants in the offense. United States v. Guyton, 36 F.3d 655, 662 (7th Cir. 1994); Carson, 9 F.3d at 584; see also United States v. McGuire, 957 F.2d 310, 317 n. 4 (7th Cir.1992). When we use the word “control,” however, we use it in a broad sense to mean some kind of supervisory or organiza*1235tional role with respect to those participants, including recruitment of other participants, Guyton, 36 F.3d at 662; Carson, 9 F.3d at 585.
The evidence is sufficient to support the role enhancement applied to Ferguson. First, evidence shows that Ferguson had a leadership role with respect to Stewart. For example, one witness testified that Stewart was Ferguson’s “right-hand man.” Testimony also revealed statements by Stewart indicating that Ferguson was responsible for obtaining the drugs Stewart distributed. In addition, the PSR states that Ferguson had five to ten juveniles distributing for him. See United States v. Ferguson, 35 F.3d 327, 334 (7th Cir.1994) (citing fact that defendant had others distribute for him as evidence of leadership role), cert. denied, — U.S. -, 115 S.Ct. 1832, 131 L.Ed.2d 752 (1995); United States v. Graciani, 61 F.3d 70, 76 (1st Cir.1995) (holding numerosity requirement satisfied when appellant led group comprised mostly of juveniles). Ferguson offers no more than a bare denial of this aspect of the PSR. Because Ferguson offered no contradictory evidence, the district court was entitled to sentence based upon this information. Taylor, 72 F.3d at 547; Mumford, 25 F.3d at 467. Moreover, the evidence adduced both at the sentencing hearing and at trial supports the PSR’s findings. Deputy Sheriff Roland testified at the sentencing hearing that Ferguson had six to seven juvenile distributors working for him. He testified specifically regarding one juvenile, Thompson, who distributed for Ferguson. A witness at trial, Allen, also testified that Ferguson had juveniles between 12 and 16 years old distributing for him. Allen specifically testified that he saw Ferguson deliver crack to a juvenile named Marshall, that he bought from Marshall on occasion and that Marshall said he got his drugs from Ferguson. According to Allen, Gunn, a third juvenile, also sold for Ferguson and had told Allen he couldn’t front drugs to him because the drugs were Ferguson’s. The three named juveniles in addition to Stewart satisfy the numerosity requirement of the § 3Bl.l(a) role enhancement. Given this evidence, we cannot find that the district court’s application of the four-point enhancement for Ferguson’s leadership role was clear error.
5. Firearm Possession
Defendants object that there was insufficient evidence to enhance either Ferguson’s or Stewart’s sentence by two levels based on possession of a firearm. We review for clear error the district court’s decision to enhance a sentence for gun possession under § 2Dl.l(b)(l). United States v. Wetwattana, 94 F.3d 280, 283 (7th Cir.1996). The enhancement is warranted if the government proves by a preponderance of the evidence that the weapon was possessed during the course of a conspiracy. Id. There is no requirement that the defendant use the weapon. Id. at 285. Under the Guidelines, the enhancement should apply unless it was “clearly improbable that the weapon was connected with the offense.” U.S.S.G. § 2D1.1, comment, (n.3). Thus, the government need not prove that the weapon was connected to the offense, as long as it was possessed during the offense or during related relevant conduct as defined in U.S.S.G. § 1B1.3. Wetwattana, 94 F.3d at 283. A number of witnesses testified to seeing both Ferguson and Stewart with a weapon. For example, after Ferguson discovered that the drugs he had purchased in Tennessee with Stewart were missing, Ferguson pulled out a gun and threatened to kill the person who had taken the drugs. There was also testimony that Stewart had a gun which he had pulled out of its gun ease because he was storing “weed” in the case. We do not think that the enhancements for firearm possession are clearly erroneous.
6. Type of Cocaine Base
Both Ferguson and Stewart argue that there is insufficient evidence to support the enhanced penalties they received for distribution of “cocaine base.” “Cocaine base” is punished at a 100:1 ratio as compared to “cocaine,” such that 5 grams of “cocaine base” is equivalent to 500 grams of “cocaine” *1236for sentencing purposes. U.S.S.G. § 2Dl.l(c). Under the Sentencing Guidelines, “ ‘Cocaine base,’ ..., means ‘crack.’ ‘Crack’ is the street name for a form of cocaine base, usually prepared by processing cocaine hydrochloride and sodium bicarbonate, and usually appearing in a lumpy, rock-like form.” U.S.S.G. § 2Dl.l(c), note (D). See also United States v. Booker, 70 F.3d 488, 494 (7th Cir.1995) (defining “cocaine base” as used in 21 U.S.C. § 841(a)(1) as meaning “crack cocaine”), cert. denied, — U.S. -, 116 S.Ct. 1334, 134 L.Ed.2d 484 (1996).
Stewart and Ferguson argue that the government failed to prove that they distributed crack cocaine, as opposed to any other form of cocaine base. This argument is meritless. We have previously held that expert testimony such as that which the government presented, that numerous samples of the drugs sold by defendants contained cocaine base, is sufficient to support the use of enhanced penalties for crack cocaine. See United States v. Reddrick, 90 F.3d 1276, 1283 (7th Cir.1996). Defendants rely on United States v. James, 78 F.3d 851 (3d Cir.), cert. denied, — U.S. -, 117 S.Ct. 128, 136 L.Ed.2d 77 (1996), as precedent for holding that there is insufficient proof that the form of cocaine base distributed by defendants was crack. In James, the court held that the evidence that defendant distributed crack cocaine was insufficient when the controlled substance was always referred to as “cocaine base” in the indictment, the plea agreement, and by the court in the plea colloquy, and the only time it was identified as “crack” was by the government during the plea colloquy. Id. at 855-56. The instant case is clearly distinguishable. Witness after witness testified that the substance defendants distributed was “crack.” This testimony is particularly persuasive given that the Guidelines define “crack” with reference to how that term is used on the street. See U.S.S.G. § 2Dl.l(c), note (D); Booker, 70 F.3d at 493 n. 19 (holding that Congress intended to punish trafficking in what is - “commonly known as ‘crack’ ”). In addition, one witness testified to viewing Stewart cook powdered cocaine into crack using baking soda (i.e., sodium bicarbonate), the usual method of manufacturing crack. U.S.S.G. § 2Dl.l(c), note (D). Thus, the evidence presented was clearly sufficient to support punishing the defendants using the enhanced penalties for crack cocaine.
III. Hall
Hall’s conviction was based largely on the testimony of Cynthia Baldwin, a government informant who made a controlled sale of drugs to Hall. On March 27, 1995, Baldwin wore a wire and went in search of a crack dealer. She couldn’t find who she was looking for, so at around 3 p.m., she went to Stewart’s house. There she met Hall, although she did not know who he was at the time. She asked to see Stewart so that she could buy some crack. Hall informed her that Stewart was busy, but that he would get it for her if she gave him the money. This she did. After entering the house, Hall returned with crack which he delivered to Baldwin who was sitting in her car. After getting Baldwin to give him a bit of the drug for his services, Hall stuck his head in the car and warned Baldwin that she had better not be affiliated with the police. On March 31, Baldwin picked Hall’s picture out of a five person photographic line-up. She also identified him at trial.
1. Prior Inconsistent Statement Instruction
Baldwin testified at trial that she had never seen Hall’s co-defendant Ferguson in her life, and even at trial she was not sure if Ferguson was in the courtroom. However, prior to trial the police showed Baldwin a picture of Ferguson. Baldwin told police that she saw Ferguson, although she did not know his name, when she sold the drugs to Hall. Hall joined Ferguson’s request to give a prior inconsistent statement jury instruction based on Baldwin’s testimony. The court refused to give the instruction. Hall argues the instruction was crucial to show Baldwin was not credible given that her testimony was the basis of Hall’s conviction. Although it may not imply Baldwin was lying, the inconsistency does reflect on Baldwin’s memory and perception in ways that may *1237harm the credibility of her identification of Hall.
“[A]n instruction need be given only when it addresses an issue reasonably raised by the evidence.” United States v. Akinrinade, 61 F.3d 1279, 1287 (7th Cir.) (internal quotations omitted), cert. denied, - U.S. -, 116 S.Ct. 541, 133 L.Ed.2d 444 (1995). The government argues that there was in fact no inconsistency in Baldwin’s statements because she never testified that it was Ferguson in the yard; she just identified someone as being in the yard who happened to be Ferguson. However, when she saw Ferguson face to face, she did not recognize him as someone she had seen before. According to Hall, the statements are inconsistent because Baldwin had previously stated she saw Ferguson (even though she did not know who he was) in the yard, the same person she claimed later never to have seen in her life. Hall argues this reflects poorly on her ability to remember faces, and thus makes her identification of Hall less believable. The judge concluded that the prior inconsistent statement instruction was not appropriate because Hall did not impeach Baldwin based on that statement.
We review refusals to give jury instructions to determine if the defendant was prejudiced. United States v. Rodriguez, 53 F.3d 1439, 1449 (7th Cir.1995). If the instructions on the whole treated the issues fairly and adequately, the refusal to give a jury instruction will not merit reversal on appeal. Akinrinade, 61 F.3d at 1287. Assuming Hall properly preserved the issue, which the government contests, we find that Hall did not suffer prejudice from the court’s refusal to give this instruction. First, the inconsistent statement did not seriously discredit Baldwin’s identification of Hall. In contrast to Hall, with whom Baldwin testified she was at one point almost “nose to nose,” Baldwin testified that she barely even looked at anyone else at the scene. Moreover, the instruction does not highlight the inconsistencies in Baldwin’s statements as damaging her credibility. Instead, it focuses on the fact that the previous inconsistent statement is not to be considered for its truth, but only as impeachment. The instruction reads as follows:
Evidence that on some former occasion a witness made a statement inconsistent with his testimony in this case may be considered by you only in determining the credibility of the witness and not to establish the truth of the matters contained in that prior statement.
The jury additionally was instructed to make its own credibility determinations and that they should consider the witnesses’ memory and ability to observe when making credibility determinations. The court also gave an instruction specifically addressing the credibility of identification witnesses. The jury heard the evidence of the prior identification and thus was able to judge Baldwin’s memory and perception based on her later inability to recognize Ferguson. We are convinced that failure to give this instruction did not impair the jury’s assessment of Baldwin’s testimony and did not prejudice Hall.
2. Photo Display and In-Court Identification
After making the buy from Hall, Baldwin was shown a photographic display of five people, from which she identified Hall as the person from whom she bought the crack cocaine. Hall objects to this identification procedure as unreliable because Baldwin knew one of the men portrayed (Carl Stewart) and a second picture was too dark to be recognized; hence, Baldwin essentially was only viewing three pictures. Hall further argues that because Baldwin does not recall whether she told officers if Hall had a goatee and only one person pictured had a goatee, her choice was further limited to two possibilities. Hall also asserts that the in-court identification was prejudicial because it was tainted by the pretrial identification.
We review the decision to allow identification testimony for clear error, engaging in a two-step analysis. United States v. Funches, 84 F.3d 249, 253 (7th Cir.1996). First, the defendant must show that the identification procedure used was unduly suggestive. Id. If it was, the court next must determine whether, given the totality of the *1238circumstances, the testimony was reliable despite the suggestiveness. Id. We look to five factors to assess reliability: “(1) the opportunity of .the witness to view the event and the actor; (2) the degree of the witness’s attention; (3) the accuracy of the witness’s description; (4) the witness’s level of certainty; and finally, (5) the time elapsed between the crime and the identification.” United States v. Fryer, 974 F.2d 813, 821 (7th Cir.1992); see also Neil v. Biggers, 409 U.S. 188, 199-200, 93 S.Ct. 375, 382-83, 34 L.Ed.2d 401 (1972).
We need not address whether this line-up was unduly suggestive; we hold that Baldwin’s testimony was sufficiently reliable in spite of any suggestiveness. First, Baldwin had a very good opportunity to observe Hall. It was the middle of the afternoon, and she was able to get quite close to him, at one point almost “nose to nose.” Second, the witness testified that she had in mind the agents’ instructions that she was to pay particular attention to whom she saw because she would have to identify him or her later. The third factor is accuracy of the witness’s description. At trial, Baldwin testified to having given a fairly detailed description of Hall. Only having seen the photograph from the line-up, it is difficult for us to judge its accuracy.. However, Hall makes no argument on appeal that it is inaccurate. Fourth, Baldwin seemed certain in her identification. At trial she testified that she “automatically recognized” Hall when she saw the line-up. Finally, only four days elapsed between the controlled buy and the identification at the photographic line-up. Considering these factors, we believe Baldwin’s testimony to be sufficiently reliable, despite any suggestiveness in the line-up.
Hall also claims that Baldwin’s in-eourt identification was tainted by the suggestiveness of the line-up. Having determined the identification to be reliable despite any suggestiveness, we must reject this claim. See United States v. Duprey, 895 F.2d 303, 308 (7th Cir.1989) (affirming admission of in-court identification after unduly suggestive photographic line-up because there existed sufficient indicia of reliability). Moreover, we cannot agree with Hall that Baldwin’s in-court description of Hall as having a ruddy complexion and one eye that “went in” is an “obvious” attempt to conform her description to the photograph. We cannot see how Hall’s picture portrays a person of particularly ruddy complexion; nor can we discern from the picture that he has one eye that “goes in.” As Hall’s counsel brought out at trial, there are legitimate reasons why one might be skeptical of Baldwin’s identification of Hall. But these criticisms go only to the weight of her testimony and are not so significant as to render it inadmissible. “Juries are not so susceptible that they cannot measure intelligently the weight of identification testimony that has some questionable feature.” Duprey, 895 F.2d at 308 (quoting Manson v. Brathwaite, 432 U.S. 98, 116, 97 S.Ct. 2243, 2254, 53 L.Ed.2d 140 (1977)). Because Baldwin’s identification contained sufficient indicia of reliability, it was not clear error to allow the identification testimony.
3. Sufficiency of the Evidence
Sufficiency of the evidence challenges will be rejected if any rational jury could have found defendant guilty beyond a reasonable doubt. Durrive, 902 F.2d at 1225. Defendant argues the evidence was insufficient to convict him of distributing crack cocaine because it was based solely on Baldwin’s testimony and her identification of Hall was unreliable. We have already found Baldwin’s identification to be sufficiently reliable to be admissible. Beyond that, we will not review the credibility determinations of the jury. See Stewart v. Duckworth, 93 F.3d 262, 268 (7th Cir.1996) (rejecting sufficiency of the evidence claim based on credibility of identification testimony); United States v. Griffin, 84 F.3d 912, 927 (7th Cir.) (“It is for the jury-not the court of appeals — to judge the credibility of witnesses, and attacks on witness credibility are insufficient to sustain a challenge to the sufficiency of the evidence.”), cert. denied sub nom. Rux v. United States, — U.S. -, 117 S.Ct. 495, 136 L.Ed.2d 387 (1996), and cert. denied sub nom. Scurlock v. United States, — U.S. -, 117 S.Ct. 536, 136 L.Ed.2d 421 (1996). Viewing the evidence in the light most favorable to the government, a rational jury could have found Hall guilty beyond a reasonable doubt.
*12394. Career Offender Status
Hall was sentenced to 210 months for distribution of 0.8 grams of cocaine base, including 0.4 grams from the sale to Baldwin and 0.4 grams from another sale considered at sentencing as relevant conduct. Hall’s offense level was enhanced from 16 to 32 based on his criminal history as a career offender. Hall argues that this sentence increase over-represents the seriousness of his offense.
HalTs arguments require little discussion. First, Hall argues that increasing his sentence by such a large degree based on his status as a career offender violates the double jeopardy clause. However, we have already rejected the argument that the career offender enhancement violates double jeopardy when the sentence remains within the statutory maximum. United States v. Saunders, 973 F.2d 1354, 1365 (7th Cir.1992); United States v. Alvarez, 914 F.2d 915, 919-20 (7th Cir.1990). Hall’s 210 month sentence is within the statutory maximum to which the court could have sentenced him without his career offender status. See 21 U.S.C. § 841(b)(1)(C) (establishing 20 year maximum sentence for distribution of less than 5 grams of cocaine base). We decline Hall’s invitation to revisit the double jeopardy issue.
Hall also challenges the district court’s refusal to depart downward. We generally will not review the court’s discretionary decision to refuse to grant a downward departure unless that refusal was based on a legally erroneous belief that the court did not have the power to depart. United States v. Murray, 89 F.3d 459, 463 (7th Cir.1996); United States v. Helton, 975 F.2d 430, 434 (7th Cir.1992). Hall asserts that the court erred by basing its refusal to depart on conduct for which Hall was acquitted. Specifically, he asserts that the court assumed his relevant conduct to include distribution of 0.4 grams of crack testified to by Allen. Hall argues, however, that Allen’s testimony regarding Hall’s sale of the second 0.4 grams of crack should be discredited. Allen also testified that on one occasion when Hall did not have any crack to sell, Hall stated- he would have to get some from Stewart. Because the jury acquitted Hall of conspiring with Stewart, Hall asserts Allen’s testimony is not credible. However, as Hall should well know given the other issues on appeal, merely purchasing drugs from someone does not necessarily amount to a conspiracy. It is entirely possible that the jury credited Allen’s testimony that Hall purchased drugs from Stewart on this one occasion, while finding that the evidence was insufficient to prove that Hall conspired with Stewart. In any event, the record is entirely devoid of evidence that the second 0.4 gram sale of crack had any impact on the court’s refusal to depart. The court recognized that the drug quantity, 0.8 grams, was small, but found no basis for departure given Hall’s criminal history. Because the district court did not base its refusal to depart on either legal or factual error, we have no jurisdiction to review its discretionary decision not to depart downward. Murray, 89 F.3d at 463; Helton, 975 F.2d at 434.
IV. Conclusion
For the reasons stated above, the convictions and sentences of Ferguson, Stewart, and Hall are Affirmed.