Per Curiam.

Defendant pleaded guilty to armed robbery, MCL 750.529; MSA 28.797, and to being a second-offense habitual offender, MCL 769.10; MSA 28.1082. He was sentenced to ten to thirty years’ imprisonment and now appeals as of right. We remand.

Defendant argues that he is entitled to resentencing because of two alleged scoring errors that occurred in determining the guidelines’ range for his armed robbery conviction. In light of our Supreme Court’s recent opinion in People v Cervantes, 448 Mich 620; 532 NW2d 831 (1995), we must first determine whether the sentencing guidelines are still pertinent in determining an habitual offender’s sentence. Six of the justices in Cervantes stated, in conformity with past precedent, that the sentencing guidelines do not strictly apply to habitual offenders. Id. at 625, 630, 634. However, while three justices (Riley, Mallett, and Weaver, JJ.) would give absolutely no consideration to the guidelines’ ranges when sentencing an habitual offender, id. at 625-626, three other justices (Cavanagh, Levin, JJ., and Brickley, C.J.) opined that the guidelines are a useful tool or starting point in determining whether an habitual offender’s sentence is proportionate, id. at 634. Using the guidelines as a starting point or as a useful tool in sentencing habitual offenders is consistent with past practice in Michigan. See, e.g., People v Martinez (After Remand), 210 Mich App 199, 201; 532 NW2d 863 (1995); People v Derbeck, 202 Mich App 443, 446; 509 NW2d 534 (1993). Justice Boyle expressed no opinion in Cervantes regarding the guidelines and their usefulness in *213 determining habitual offenders’ sentences. Thus, the Cervantes opinion does not present a majority opinion concerning this issue. We are bound to follow past precedent and will continue to use the guidelines as a starting point and useful tool in determining whether an habitual offender’s sentence is proportionate.

Defendant argues that the trial court incorrectly scored Offense Variable (ov) 9. Ov 9 is to be scored as ten points when a defendant is a leader in a multiple offender situation. The instructions state that the entire criminal episode should be taken into account in determining whether the offender is a leader. In this case, defendant testified that he was the person who grabbed the victim, stuck a knife to the victim’s throat, and took the victim’s wallet. Moreover, defendant testified that the other codefendant (defendant’s brother) did not know that defendant was going to rob the victim. On the basis of these facts, we conclude that sufficient evidence existed on the record to support the score of ten points for ov 9. People v Johnson, 202 Mich App 281, 289-290; 508 NW2d 509 (1993).

Defendant also challenges the scoring of Prior Record Variable (prv) 7. Prv 7 is scored as twenty points when a defendant has two or more subsequent or concurrent felony convictions and as ten points when a defendant has one subsequent or concurrent felony conviction. Because it appears from the record that defendant had only one subsequent or concurrent felony conviction (unarmed robbery), he should have been assessed only ten points for prv 7. 1 There was no evidence to support a score of twenty for prv 7. See Johnson, supra at *214 288. Because the scoring of prv 7 decreases the prior record level from category d to c and also decreases the guidelines’ range from five to twenty-five years to three to eight years, we remand to the trial court so that it can consider whether a different sentence should be imposed. * 2

We remand. We do not retain jurisdiction.

We note that defendant asserts that his second-offense habitual offender conviction was counted as a concurrent felony conviction and was used to score prv 7 as twenty points. To the extent that defendant’s assertion is correct, such action would be inappropriate because the habitual offender statute does not create a substantive offense *214 that is separate from and independent of the principal charge. People v Bewersdorf, 438 Mich 55, 67; 475 NW2d 231 (1991); People v Connor, 209 Mich App 419, 426; 531 NW2d 734 (1995).

We are not implying that a different sentence must be imposed or that we think defendant’s sentence of ten to thirty years’ imprisonment is disproportionate. Furthermore, as previously stated, the sentencing guidelines are a useful tool in determining an habitual offender’s sentence. Martinez, supra. Although a strict mathematical approach should not be used to determine the proportionality of an habitual offender’s sentence, Cervantes, supra at 626, 634, the guidelines’ recommended range can be compared to the degree of sentence enhancement authorized by the Legislature under the habitual offender statutes, id. at 634; People v Chandler, 211 Mich App 604, 615; 536 NW2d 799 (1995). However, most importantly, the key test is whether the sentence is proportionate to the seriousness of the offense and the circumstances of the offender. People v Milbourn, 435 Mich 630, 661; 461 NW2d 1 (1990); Chandler, supra at 616.