We must decide whether extraordinary eyesight may be considered a “special skill” supporting an enhanced sentence in a casino card cheating scheme.

I

Jing Bing Liang had long gambled at cards when he met a group of players sharing a mutual dissatisfaction with the odds imposed upon them by casinos. They decided to get together and remedy the situation, so they began to cheat. It started with the game of baccarat, and the revelation that by turning the shoe in a particular way, they could see the next card to be dealt. Although this could not guarantee a win, it significantly increased the chances of a payout and, when employed over the course of an evening, apparently proved quite remunerative. For example, in November 1994 at a casino in Las Vegas, Liang and four others won $1,500,000 by peeking at the shoe.

This proved enticing, so the scale and sophistication of the group steadily increased. Members began to differentiate their roles: aside from the overall ringleaders, there were those who organized particular cheats at varying casinos and those that were lookouts; some would distract the dealers, while still others did the actual cheating. Liang’s expertise was as a cheater. He perfected his ability not only to peek, but to mark cards by “crimping” or “daubing” them, and then utilized these methods in baccarat, its mini- and midi-variations, and blackjack. All told, Liang participated in at least six cheats from 1994 through 1999 at casinos in Las Vegas, Lake Tahoe, and Atlantic City.

Federal authorities eventually caught wind of the scheme, and on October 17, 2000, Liang and his co-conspirators were indicted on charges of conspiracy to participate in an enterprise through a pattern of racketeering by cheating. See 18 U.S.C. § 1962 (d). Liang pleaded guilty on May 9, 2002.

After some wrangling at his sentencing hearing, all parties eventually agreed that at least a guideline offense level of 16 applied which, in conjunction with criminal history category I, specified a 21 to 27-month term of incarceration. The government then orally moved for a two-level sentence enhancement for the use of “special skills,” pursuant to United States Sentencing Guidelines Manual § 3B1.3 (2001) (“U.S.S.G.”). It argued that Liang had “extraordinary eyesight” allowing him to peek at the cards in the shoe, and that he had become specially trained in the art of cheating at cards. Liang objected, but the district court agreed with the government:

As a result, Liang’s offense level increased to 18 (27 to 33 months), and the district court entered a judgment sentencing him to 27 months imprisonment on October 24, 2002. Liang filed a timely notice of appeal on October 21, 2002, objecting solely to the “special skills” enhancement.

II

Section 3B1.3 provides for a two-level sentence enhancement “if the defendant abused a position of public or private trust, or used a special skill, in a manner that significantly facilitated the commission or concealment of the offense.” The application note defines a “special skill” as “a skill not possessed by members of the general public and usually requiring substantial education, training or licensing. Examples would include pilots, lawyers, doctors, accountants, chemists, and demolition experts.” § 3B1.3, cmt. n. 3.

The government argues that the district court based Liang’s “special skills” enhancement on two grounds: (1) his ability to cheat at cards; and (2) his extraordinary eyesight. We examine each in turn, de novo. See Koon v. United States, 518 U.S. 81, 100 , 116 S.Ct. 2035 , 135 L.Ed.2d 392 (1996) (“[W]hether a factor is a permissible basis for departure under any circumstances is a question of law, and the court of appeals need not defer to the district court’s resolution of the point.”).

A

The district court imposed a § 3B1.3 enhancement on Liang because of his ability to cheat at cards. However, when a “special skill” has little apparent use outside the criminal context, a § 3B1.3 enhancement seldom will be appropriate. See United States v. Mainard, 5 F.3d 404, 406 (9th Cir.1993) (“[T]he purpose of the [§ 3B1.3 enhancement] is to add to the punishment of those who turn legitimate special skills to the perpetration of evil deeds.” (emphasis added)); United States v. Green, 962 F.2d 938, 944 (9th Cir.1992) (“[T]he ‘special skill’ enhancement provision of section 3B1.3 applies only if the defendant employed a ‘special skill’ in the form of a pre-existing, legitimate skill not possessed by the general public to facilitate the commission or concealment of a crime.” (internal quotation omitted) (emphasis added)). When read in conjunction with the abuse of a “position of public or private trust” provision, it is clear that a § 3B1.3 “special skill” involves legitimate, socially valuable expertise that a criminal has perverted to uncivilized and illegal ends.

Indeed, each of the guidelines’s own illustrating examples — -“pilots, lawyers, doctors, accountants, chemists, and demolitions experts” — involve legitimate and important skills that only have the potential to be abused. § 3B1.3, cmt. n.3. Conversely, the ability to kidnap, to rob, or to smuggle is inherently ¿¿legitimate, creating little or no potential for gainful employment, and thus would not belong in this class even if arguably classified as a “skill.” See, e.g., Green, 962 F.2d at 944 (“Courts have generally rejected ap plication of the guideline merely because the offense was difficult to commit or required a special skill to complete.”).

Liang’s ability to cheat at cards cannot be understood as legitimate: It was basically useless outside the criminal context, no matter how good he got. And unlike a doctor or lawyer gone bad, Liang abused no societal trust by appropriating an otherwise positive educational investment for illegal personal gain. See Mainard, 5 F.3d at 406 (asserting that the abuse of a legitimate skill “is a breach of the trust that society reposes in a person when it enables him to acquire and have a skill that other members of society do not possess”). Rather, he only developed his criminal expertise, a course of conduct we do not condone, but which cannot form the legal basis for a § 3B1.3 enhancement. See Mainard, 5 F.3d at 406 .

Moreover, a skill is only “special” for purposes of § 3B1.3 if a district court determines “not just whether the skill is ‘not possessed by members of the general public,’ but also, as a sine qua non, whether it is a skill ‘usually requiring substantial education, training, or licensing.’ ” United States v. Lee, 296 F.3d 792, 798 (9th Cir.2002) (quoting § 3B1.3, cmt. n. 3). Perhaps most of us are unable to mark or to peek at cards. Nevertheless, the district court made no findings with respect to the extent of Liang’s training or sophistication at card cheating, and whether it required “substantial” effort to obtain. See, e.g., United States v. Petersen, 98 F.3d 502 , 507 n. 5 (9th Cir.1996) (“Only where a defendant’s ... skills are particularly sophisticated do they correspond to the Sentencing Commission’s examples of ‘special skills’.... ” (emphasis added)); United States v. Harper, 33 F.3d 1143, 1151 (9th Cir.1994) (noting that a special skill “usually require[s] substantial education, training or licensing” (emphasis added)).

Liang suggests that peeking, crimping, or daubing is relatively simple, and that anyone could learn to do it in a short amount of time. In any event, without additional findings by the district court, it certainly seems no more sophisticated than many skills that have been insufficient to trigger § 3B1.3. See, e.g., Lee, 296 F.3d at 799 (vacating § 3B1.3 enhancement for knowledge of web-based computer systems used in an Internet scam); Harper, 33 F.3d at 1152 (same for inside knowledge of an ATM system used in a robbery); Green, 962 F.2d at 945 (same for printing and photographic abilities in a counterfeiting case). In light of the district court’s limited findings on the matter, the enhancement cannot stand. See Lee, 296 F.3d at 796 (ruling that the district court abused its discretion “in view of the limited findings” regarding the defendant’s computer skills).

For these reasons, the district court erred by applying a “special skills” enhancement based on Liang’s ability to cheat at cards.

B

The government also argues that the district court correctly based the application of § 3B1.3 on Liang’s “extraordi nary eyesight.” Whether an underlying physical characteristic may form the basis for a “special skills” enhancement appears to be a question of first impression.

It is not unusual to describe certain well-developed physical capabilities as “skills.” See, e.g., Cleghorn v. Herrington, 813 F.2d 992, 996 (9th Cir.1987) (arguing that the military can “require its personnel to have the physical skills necessary to engage in [armed] combat should the situation arise”). Indeed, in some instances, formalized physical training might result in “special skills” that could be abused for illegal gain. For example, an ex-soldier theoretically might apply armed combat skills, acquired in the military, in order to facilitate a murder or a bank robbery.

Still, the fact that a skill can involve physical ability does not mean that a physical characteristic, standing alone, can also be described as a skill. While the ability to use one’s hands in a certain way might be deemed a skill, the power to use them in the first place cannot. Intrinsic physical attributes are not skills — no matter how impressive they may be — because skills involve proficiency with respect to a discrete task or set of tasks. See Webster’s New International Dictionary Unabridged 2133 (3d ed.1986) (defining “skill” as “knowledge of the means or methods of accomplishing a task”). Physical characteristics, on the other hand, are not knowledge nor acquired know-how: one cannot be skilled at “stamina” no matter how much he or she jogs. Likewise, it makes little sense to describe Liang as possessing the “skill of good eyesight.”

No matter how much it contributed to his ability to peek at cards, Liang’s extraordinarily acute vision cannot be described as a skill. And because substantial training or education is what makes a skill “special” for purposes of § 3B1.3, see Lee, 296 F.3d at 798 , Liang’s visual acuity is simply irrelevant to the question of whether he should be subject to the enhancement. Accordingly, the district court also erred by considering this factor.

Ill

For the foregoing reasons, we VACATE Liang’s sentence and REMAND to the district court for resentencing consistent with this opinion.

. A "shoe” is a device casinos use to dispense playing cards from a reservoir containing a large number of shuffled decks. . Certain cards can be marked for future detection either by slightly folding them ("crimping”) or by smearing on a small bit of petroleum jelly or similar substance that subtly can be identified by touch ("daubing”). These cards later can be identified without seeing their faces, which can thereby substantially increase the odds of winning a particular hand. . The indictment additionally included two counts of money laundering, but as specified by the terms of Liang's plea agreement, those charges were dismissed upon the entry of his guilty plea. . The plea agreement indicated that the government could file such a motion and reserved Liang’s right to object if it did. . Perhaps, in certain circumstances, a casino might employ someone because he or she is skilled at card cheating. See Noah Goldman, A Textbook Case: MIT Students Break the Bank in Las Vegas, ABC News, Sept. 15, 2003, at http://abcnews.go.com/sections/prime-time/US/MITgamblers03 0915 .html (reporting that "casinos employ specialized security agencies to spot potential cheats and card counters”). However, this specialized use of Liang’s skill only serves the purpose of catching criminals who engage in precisely such conduct. Of course, one might also cheat in order to win a (previously) friendly card game. But in any event, neither of these isolated applications legitimizes card cheating more generally for purposes of § 3B1.3.