OPINION

I. INTRODUCTION

Michael Jeffries caused a traffic accident that fatally injured his passenger. A jury convicted him of second-degree murder under AS 11.41.110. The court of appeals affirmed. We consider here whether a reasonable jury could find that Jeffries displayed "extreme indifference to the value of human life," as required for a second-degree murder conviction under AS 11.41.110. There was evidence that Jeffries, after drinking aleoholic beverages at home, drove from his home to a social club, where he drank at least six more beers before he decided to drive himself and a passenger home. Jeffries was extremely intoxicated, with a blood alcohol level of about 0.27 percent, when he drove his car directly and slowly in front of an oncoming car on a well-lit, icy, five-lane street. Evidence of Jeffries's prior convictions for drunk driving, court orders to attend substance abuse programs, and a condition of probation that he abstain from drinking alcohol permitted a finding that he had a heightened awareness that driving while grossly intoxicated was highly dangerous. We hold that the evidence was sufficient to support the verdict, and affirm.

II. FACTS AND PROCEEDINGS

was "well-lit" with streetlights. While grossly intoxicated, Michael Jeffries drove an automobile and caused a February 7, 2000 traffic accident that fatally injured his front seat passenger, Beulah Dean. The two were eastbound on DeBarr Road in Anchorage, driving home at 8 P.M. from a social club where alcohol had been served, when Jeffries made an abrupt left turn-but at slow speed-directly in front of Mark Ber-geron's oncoming westbound car. The posted speed on the five-lane street where the collision occurred was forty-five miles per hour and Bergeron was driving at about thirty-five miles per hour. There was a thin layer of packed snow in the center lane from which Jeffries turned. There was a sheen of ice in the traffic lanes and the road was icy and slippery. It was dark, but the street Bergeron and the investigating detective believed that headlights of Bergeron's car were on at the time of the collision. The right front corner of Bergeron's car struck the passenger door beside Beulah Dean, penetrating more than twelve inches into the passenger compartment and fatally injuring Dean. She was taken to a hospital but died soon after.

*915Jeffries's blood alcohol content was 0.27 percent when it was tested about seventy minutes after the accident. There was evidence that Jeffries had been drinking alcoholic beverages before noon on the day of his accident, that after drinking in the morning, he drove to a social elub, where he consumed at least six more beers before he drove home and caused the fatal accident. There was also evidence that Jeffries may have been drinking while he was driving home. A po-Hice officer found an empty beer can on the floor on the driver's side of the car. As the court of appeals later observed, "[vliewing the evidence presented at trial in the light most favorable to the State, Jeffries downed approximately twenty beers over the course of several hours.1

Jeffries was indicted on a charge of see-ond-degree murder under AS 11.41.110(2)@) for engaging in conduct that resulted in death under cireumstances manifesting extreme indifference to the value of human life. He was also charged with manslaughter, negligent homicide, driving while intoxicated (DWT), and driving with a suspended license.

At trial, the prosecution introduced evidence of Jeffries's long 'history of driving while intoxicated. The jury heard evidence that Jeffries had six prior DWI convictions, that his license had been suspended since 1989, that he had four times failed to participate in court-ordered substance abuse programs, and that as a condition of probation he had been ordered to abstain from drinking alcohol, Jeffries objected to the admission of this evidence as irrelevant, unfairly prejudicial, and improper character evidence.

. After the close of evidence Jeffries moved for a judgment of acquittal on the second-degree murder count. He argued that the state had failed to present sufficient evidence that he was driving in a manner that exhibit, ed extreme indifference to the value of human life. The superior court denied Jef-fries's motion; in doing so, it relied in part on Jeffries's history of driving while intoxicated. The jury found Jeffries guilty of see-ond-degree murder, driving while intoxicated, and driving with a suspended license.

Jeffries appealed, arguing that extreme-indifference murder should be reserved for cases in which an intoxicated driver operates his vehicle in a particularly dangerous or heedless manner. After a thorough review of Alaska and nationwide case law, the court of appeals agreed with the superior court that Jeffries's past convictions for driving while intoxicated, his repeated refusal to participate in court-ordered treatment for alcohol abuse, his decision to drive despite his Heense suspension or revocation for prior DWI convictions, as well as his "extreme intoxication" on the night of the accident were sufficient to allow the murder charge to go to the jury, even if the defendant did not engage in "egregiously dangerous driving.2 Given this holding, the court also rejected Jeffries's contention that the evidence of the DWI convictions and Jeffries's failure to comply with court-ordered substance abuse treatment and abstinence from alcohol was irrelevant and unduly prejudicials.3

We granted Jeffries's petition for hearing.

IIL. DISCUSSION

A. The Evidence of Jeffries's Extreme Indifference to the Value of Human Life Was Sufficient To Allow a Reasonable Jury To Convict Him of Second-Degree Murder.

Jeffries challenges the superior court's denial of his motion to acquit. In reviewing the denial of a motion to acquit, we must determine whether there is "such relevant evidence which is adequate to support a conclusion by a reasonable mind that there was no reasonable doubt as to [the defendant's] guilt."4 In making this determination, we "will consider only those facts in the record most favorable to the prosecution and *916such reasonable inferences as a jury may have drawn from them." 5

. Jeffries was convicted of second-degree murder under AS 11.41.110. Subsection (a)@) of this statute provides that a person commits murder in the second degree if "the person knowingly engages in conduct that results in the death of another person under cireumstances manifesting an extreme indifference to the value of human life."6 This provision is adapted from the Model Penal Code, which was "adopted in large measure" by the Alaska legislature in 1978.7 In interpreting this provision, we follow the approach of the court of appeals 8 and look not only to Alaska case law, but also to authorities interpreting the Model Penal Code. ©

When determining what culpable mental states must be proved for each element of an offense, the court of appeals in Neifzel v. State explained that both the Model Penal Code and Alaska's revised criminal code divide the elements of a eriminal offense into three categories: conduct, surrounding circumstances, and result.9 Applying this framework to AS 11.41.110(a)(2), the court of appeals found that the conduct is "performing an act," that the surrounding cireum-stances must manifest "an extreme indifference to the value of human life," and that the result is "the death of another person.10 The court of appeals concluded that knowledge is the mens rea for conduct,11 and that recklessness is the mens rea for the surrounding cireumstances and the result.12 The court of appeals also identified four factors the jury must consider in determining whether a defendant has displayed extreme indifference to the value of human life:

(1) the social utility of the actor's conduct;

(2) the magnitude of the risk his conduct creates including both the nature of foreseeable harm and the likelihood that the conduct will result in that harm;

(3) the actor's knowledge of the risk; and

(4) any precautions the actor takes to minimize the risk, [13]

These factors have been in use in Alaska since 1982 and provide a proper framework to distinguish extreme-indifference murder from manslaughter.

The commentaries to the Model Penal Code suggest that extreme-indifference murder is intended to allow actors to be convicted of murder if their actions, while not purposeful or knowing with regard to the resulting death, demonstrate equivalent indifference to the value of human life.14 According to the commentaries, "there is a kind of reckless homicide that cannot fairly be distinguished in grading terms from homicides *917committed purposely or knowingly." 15 Recklessness is defined as "an awareness of the creation of substantial homicidal risk, a risk too great to be deemed justifiable by any valid purpose that the actor's conduct served." 16 For a reckless homicide to be classified as murder instead of manslaughter, the factfinder must find that "the actor's conscious disregard of the risk, under the cireumstances, manifests extreme indifference to the value of human life.17 The commentaries advise that the factfinder must determine whether "extreme indifference to the value of human life" exists:

Whether recklessness is so extreme that it demonstrates similar indifference [as to purposeful or knowing homicide] is not a question, it is submitted, that can be further clarified. It must be left directly to the trier of fact under instructions which make it clear that recklessness that can fairly be assimilated to purpose or knowledge should be treated as murder and that less extreme recklessness should be punished as manslaughter. [18]

Because the question whether an actor's conduct demonstrates extreme indifference to the value of human life is primarily one for the factfinder, only rarely will evidence favorable to the defendant as to a single factor in the Neitgel analysis prevent the case from going to a jury. Neitgel's four factors provide a test in which particularly convincing evidence as to one factor may compensate for lack of evidence as to another. Thus, although attempting to drive normally while intoxicated usually renders the driver who causes a death eulpable of only manslaughter,19 such conduct might be found to demonstrate the requisite extreme indifference if the other Neitgel factors all point strongly towards greater culpability. The - court may only intervene if the evidence, viewed as a whole, cannot be reasonably interpreted as demonstrating the type of heightened recklessness that is equivalent to purposeful or knowing homicide.

Jeffries contends that the only way to ensure a clear distinction between manslaughter and extreme-indifference murder is to reserve murder for cases in which the objective risk of death or serious physical injury posed by the defendant's actions is "very high." This is a correct statement of the law, but we do not agree with his implicit contention that the objective risk posed by his conduct was not "very high."

Jeffries correctly notes that many intoxicated drivers whose convictions of extreme-indifference murder were affirmed on appeal in Alaska operated their vehicles in an exceptionally dangerous manner over an extended period of time. In Ratliff v. State, for example, the intoxicated defendant swerved across the road, causing an accident and a near-miss, and then drove at an excessive speed on the wrong side of a divided highway for two miles until he collided head-on with another car.20 In S#iegele v. State, the intoxicated defendant drove on the left side of the road at eighty-five miles per hour with passengers in the back of his truck who were screaming for him to stop, and finally crashed when he could not negotiate a corner.21 And in Pears v. State, the intoxicated defendant ran stop signs and red lights and eventually collided with another car when he ran a red light without even slowing.22

Although the defendants in those cases engaged in more egregious driving conduct than Jeffries, this does not mean that his driving was not in fact egregious. -

*918When viewed in the light most favorable to the state, the evidence at trial-including the expert testimony concerning the impairing effects of a .27 blood alcohol level and the testimony describing the accident itself-would have enabled a reasonable jury to find not just that Jeffries was extremely intoxicated, but also that his intoxication extremely impaired his ability to drive, so that he lacked the ability to identify and react to common and easily avoidable hazards of everyday driving. In other words, the evidence tended to show that he was literally "blind" drunk to oncoming cars, not merely distracted or somewhat slowed down. Severe impairment of his kind would pose a grave danger at every intersection Jeffries crossed, not just at the place where his (and Dean's) luck happened to run out; and the danger of driving while blind to surrounding hazards is no less egregious merely because it poses a covert rather than an overt risk.

Nor is prolonged driving misconduct over an extended period of time inherently necessary for an extreme-indifference murder conviction. Jeffries has not identified any case in which this court or the court of appeals has overturned a jury verdict of extreme-indifference murder because the evidence of objective risk was insufficient.

Furthermore, Alaska defendants who have driven while severely intoxicated and who have engaged in driving conduct comparable to Jeffries's have been convicted of extreme-indifference murder.23 In two such reported cases, Richardson v. State24 and Pusewicz v. State,25" the defendants did not challenge their convictions on appeal, and indeed both pleaded no contest to charges of extreme-indifference murder.26 Both defendants unintentionally crossed the center line and collided with oncoming vehicles.27 Both were convicted of extreme-indifference murder for driving conduct that essentially involved fatal lapses of attention or control by very intoxicated drivers who, like Jeffries, knew or should have known they should not be driving. Neither engaged in inherently reckless or intentional gravely dangerous driving conduct, such as swerving in and out of traffic or driving at high speed, that might have justified extreme-indifference murder charges even against sober drivers. Both collisions seem to have occurred relatively soon after the defendants began or resumed driving. Thus, neither case involved prolonged .or overtly "egregious" driving misconduct apart from erratic driving resulting from each defendant's severe intoxication. Although both appellate decisions were only sentence appeals, it is significant that no party or court in either case appears to have detected any obvious legal or evidentiary flaw in basing an extreme-indifference murder conviction on a death attributable to this sort of driving conduct. No one seems to have thought that prolonged and overtly "egregious" driving conduct was necessary to support either conviction under all the cireumstances in each case. Indeed, the absence of such driving conduct did not generate much mitigating force with respect to sentencing in either case. These cases illustrate that Jeffries's proposed restrictions on extreme-indifference murder would be a sharp break from the long-accepted view of the offense in Alaska. And as a practical matter, that only a few such appellate cases have arisen during the past decades refutes Jeffries's claim that his proposed restrictions are needed to avoid some sort of an endless slippery slope that threatens to swallow all repeat DWI offenders.

Jeffries cites cases from outside Alaska that purportedly demonstrate that extreme-*919indifference murder is not an appropriate charge for intoxicated drivers attempting to drive normally. Several of these cases, such as Park v. State,28 State v. Jensen,29 and Blackwell v. State,30 were decided some years ago, when public awareness of the dangers of driving while intoxicated was far less than it is today.31 Other cases cited by Jeffries hold that a typical drunk driving accident should not be grounds for extreme-indifference murder, but appear to leave open the possibility that aggravating factors that could justify a murder conviction are not limited to especially egregious driving over a long period of time. For example, in Allen v. State, the Alabama Court of Criminal Appeals held that "the 'situation' that will support a conviction for reckless murder must involve something more than simply driving after having consumed alcohol and becoming involved in a collision.32 "[Slome shocking, outrageous, or special heinousness must be shown.33 * But nothing in that opinion suggests that operation at a very high level of intoxication and driving directly in front of oncoming traffic at such a slow speed that a passenger-side collision is sure to happen could not prove "special heinousness." Furthermore, the Alabama Supreme Court upheld Allen's murder conviction although his driving was no more egregious than Jeffries's34 Allen was not speeding at the time of the crash and had a much lower blood alcohol content than Jeffries.35 The accident resulted from his inability to keep his car in the proper lane of travel.36

Similarly, in United States v. Fleming, the Fourth Cireuit held that a conviction for reckless murder was appropriate because "the facts show a deviation from established standards of regard for life and the safety of others that is markedly different in degree from that found in most vehicular homicides." 37" Although Fleming engaged in a series of dangerous maneuvers, this does not mean that Jeffries's conduct did not meet the legal standard set forth in Fleming.38 Like Allen, Fleming leaves open the possibility *920that a "deviation from established standards" may be found on the basis of factors other than a prolonged period of egregious driving.

Jeffries also relies on scholarly authority as theoretical support for his argument. He cites a student note in the American Criminal Law Review that argues "the average drunk driver who drives poorly simply because of alcohol-induced sense distortion" cannot be found guilty of extreme-indifference murder.39 The note reasons that extreme indifference "can only be proven from inferences drawn from the defendant's conduct." 40 In Neitzel, the court of appeals took a different view of "sense distortion," when it concluded that "recklessness may be found despite unawareness of a risk where intoxication accounts for the failure to perceive the risk.41 4 We agree with the court of appeals' holding that drunk drivers are responsible for their actions when their intoxication prevents them from perceiving dangers that a sober driver would notice. Indeed, this holding is dictated by the plain language of the Alaska Criminal Code, which defines the culpable mental state "knowingly" to require a finding of knowing conduct when the defendant's failure to perceive surrounding cireumstances results from voluntary intoxication: "[A] person who is unaware of conduct or a cireum-stance of which the person would have been aware had that person not been intoxicated acts knowingly with respect to the conduct or cireumstance." 42 In our view, the circumstances present here, including the defendant's extreme intoxication, his knowledge based on prior convictions that such intoxication was unjustifiably dangerous, and his conduct in driving directly in front of an oncoming car that had no opportunity to stop allow a jury to infer a defendant's disregard for the lives of others. The student note concedes the point that previous drunk driving convictions should be "placed in the scales" in determining whether a murder charge is appropriate.43

Jeffries also argues that Professor LaFave has stated that extreme-indifference murder requires conduct creating a "very high degree of risk." 44 We agree. But Jef-fries neglects to mention that Professor La-Fave concedes that the precise degree of objective risk necessary to support a charge of extreme-indifference murder varies depending on the circumstances. "[IJt is what the defendant should realize to be the degree of risk, in the light of the surrounding circumstance which he knows, which is important, rather than the amount of risk as an abstract proposition of the mathematics of chanee." 45 Furthermore, Professor LaFave states that "the social utility of [the defendant's] conduct is a fact to be considered." 46 Our approach is consistent with Professor LaFave's reasoning. As did Professor La-Fave, we consider the concrete facts at play in each case, not merely the abstract risk of driving while intoxicated. We hold that, in a case such as Jeffries's, in which the Neitzel factors weigh heavily against the defendant when taken together, the actual degree of risk required for murder has been met.

A review of the evidence in this case in light of the four Nettzel factors demonstrates that the jury acted reasonably in finding that Jeffries acted with extreme indifference to the value of human life. We consider each factor in turn.

1. Social utility

The state argues that driving with a blood alcohol content of two-and-a-half times the legal limit has "no social utility." Jeffries concedes that the utility of driving a vehicle home from a bar while intoxicated is "limited," but apparently does not agree it is nonexistent. In the past, the court of appeals *921has held that the utility of driving while intoxicated is "marginal, at best47 " and "substantially reduced" from the utility of driving sober.48

We disagree with those decisions insofar as they suggest that driving home while intoxicated necessarily has some social utility. As the state points out, public awareness of the dangers of drunk driving has increased in recent years, as have penalties. While there is certainly utility in driving, that utility is, except in rare circumstances, completely negated by the grave danger posed to society by an extremely intoxicated driver. In this case, there was no evidence of extenuating cireumstances, such as the need to take a critically ill friend or family member to the hospital or the lack of any alternative means of getting home (e.g., taxis, buses, or friends) that might require a conclusion that Jef-fries's driving had some limited social utility. In addition, there was evidence in the record that Jeffries had been drinking on the morning of the accident and that after drinking he drove to a social club where he consumed at least six more beers before attempting to drive home. Also, Jeffries may have continued to drink while he was driving. A witness to the accident testified that "the entire vehicle smelled like alcohol," and one of the responding police officers discovered a beer can on the floor on the driver's side of the vehicle. A reasonable jury is not obliged to give an extremely drunk driver any credit for the social utility of "merely attempting to drive home."" This is especially so after he has chosen to consume aleohol in the morning at home, gets behind the wheel of a car, and drives to a social elub to continue his drinking, before rolling the dice by trying to drive home, perhaps drinking in the car while driving despite already being grossly intoxicated.49

2, Magnitude of the risk; nature and - likelihood of foreseeable harm

Jeffries argues that his actual driving was not particularly egregious and did not create "a very high risk of death." He minimizes the riskiness of his behavior by characterizing it as a "poorly executed left turn." We are unpersuaded by this characterization. Jeffries's conduct was much more risky than the conduct in a typical drunk-driving accident for two reasons.

First, the evidence suggests that Jeffries's error in judgment was severe. Jeffrics was attempting to make a left-hand turn across DeBarr Road, a five-lane street with a speed limit of forty-five miles per hour, against oncoming traffic. Jeffries was traveling as slowly as ten miles per hour when he pulled directly in front of Bergeron's oncoming car. Bergeron's headlights were on, and although it was dark, streetlights lit the street well. Bergeron was traveling at or below the posted speed limit, probably at about thirty-five miles per hour. The street was icy and slippery. Bergeron "had about enough warning to take [his] foot off the gas" before the collision, but not enough time to stop or swerve to avoid the accident. Bergeron's car hit Jeffries's passenger door-almost the center of the car. The point of impact demonstrates that Jeffries either badly misjudged the speed of the oncoming car or altogether failed to see it. His speed of ten miles per hour was too slow to permit him to cross safely in front of Bergeron's oncoming car and left Jeffries's passenger gravely and . predictably vulnerable to a side impact.

Second, Jeffries was highly intoxicated on the night of the accident. Jeffries's apartment maintenance supervisor testified that Jeffries smelled of beer during an encounter with Jeffries between 10:30 A.M. and noon on the day of the crash. In response to the supervisor's concerned inquiry about Jef-*922fries's ability to drive, Dean commented that "he's been worse than this." At roughly 8:80 P.M. Jeffries and Dean arrived at the Veterans of Foreign Wars (VFW) club in Moun-tainView. The bartender testified that Jef-fries drank "only" six beers there and left with Dean at 8:00 P.M. After the crash, an empty beer can was found on the passenger floorboard of the car. Police investigating the crash testified that Jeffries smelled strongly of alcohol and failed a field sobriety test. A blood test performed at 9:25 P.M., an hour and ten minutes after the accident, measured Jeffries's blood alcohol content at 0.27 percent. The state's expert testified that had Jeffries begun drinking at noon, he would have had to consume 28.6 drinks to reach a blood alcohol content of 0.27 percent by 9:25 P.M. ©

The evidence established that Jeffries's blood alcohol content made it highly dangerous for him to drive. An expert witness for the state testified about a study that demonstrated that the probability of causing an accident increases "exponentially" as blood alcohol content increases. While a driver with a .08 percent blood alcohol content is three times more likely to cause an accident than a sober driver,50 a driver with a 0.15 percent blood aleohol content is twelve times more likely to cause an accident than a sober person. Jeffries's blood aleohol content was nearly twice the highest level discussed by the expert. Thus, there was evidence that the probability Jeffries would cause an accident was at least twelve-and probably many more-times that for a sober driver. The fact that the roads were "icy" and "slick" on the night of the accident probably increased the risk even more because the- condition of the road made it more difficult for oncoming drivers to altogether avoid a collision by stopping or swerving or to minimize the consequences by slowing down.

At least one court, the Kentucky Supreme Court, has upheld murder convictions of intoxicated drivers based primarily on their extreme intoxication at the time of the accident.51 Although we do not decide here whether a murder conviction might be warranted on the basis of extreme intoxication alone, we do conclude that Jeffries's intoxication, at over two-and-a-half times the legal limit,52 was extreme. The jury could properly find the objective risk posed by a driver with Jeffries's level of intoxication to be significantly higher than that of a typical drunk driver.

Likewise, the nature of the harm-the risk of death or serious bodily injury-inherent in abruptly turning and driving slowly across the path of oncoming traffic on slippery streets is both great and readily foreseeable. And it is very likely, and foreseeable, under such cireumstances that the conduct will cause that harm.

3. Awareness of the risk

Jeffries's heightened awareness of the risks of drinking and driving differentiates this case from other deaths involving drunk drivers. The evidence relevant to his awareness was strong. The parties stipulated that Jeffries had six prior DWI convictions between 1981 and 1996, four of them occurring in the 1990s. The parties also stipulated that Jeffries's license had been continuously revoked since 1989, that it would remain revoked until 2018, and that Jeffries was aware that his license had been revoked.53 They also stipulated that a DWI conviction was the basis for the 2000 revocation. The state presented evidence that four times between 1989 and 1994 Alaska courts had ordered Jeffries as a condition of his probation to *923report to a probation program that sereens offenders and assigns them to alcohol treatment programs and that he failed to comply with each order despite the possibility that he could be sent to prison for noncompliance. Finally, the state presented evidence that as a condition of his probation Jeffries was forbidden from drinking alcohol at the time of the most recent accident. There is no claim Jeffries did not have actual knowledge of his past drunk driving convictions and of the court orders requiring him to get treatment. In short, there was significant evidence that Jeffries had a heightened awareness of the dangerousness of his conduct, the need to refrain completely from any driving and to refrain completely from any drinking, and of the danger of driving intoxicated.

As Superior Court Judge Dan A. Hensley explained in determining that evidence of Jeffries's past problems with alcohol was relevant to this inquiry:

[A] person who drinks, drives, causes an accident, gets arrested, goes to jail, is ordered to aleohol treatment, ordered not to drink and then drinks and drives again, and then drinks and drives again, and then drinks and drives again, not only has the intellectual understanding of the risks associated with drinking and driving but also has the very real understanding. Which in my view is relevant to show the heightened awareness of those risks. Experience is the best teacher. ©

The superior court was correct in its assessment. An intoxicated driver with a record as long as Jeffries's cannot possibly be unaware of the significant threat that his actions pose in the eyes of society. A reasonable jury could have inferred from this evidence that Jeffries had a heightened awareness of the risk of drinking and driving and could have given this factor substantial weight in its analysis under Neitzel.54

4, Precautions to minimize the risk

The state argues correctly that there is no evidence that Jeffries took any precautions to minimize the risk of his conduct. In fact, Jeffries's past failures to follow orders to participate in substance abuse programs and to refrain from either driving or drinking demonstrate a willful refusal to take precautions to minimize the risk. A reasonable jury could have properly taken Jeffries's refusal into account in evaluating whether he exhibited extreme indifference to the value of human life. |

5. A reasonable jury weighing these factors could convict Jeffries of murder.

.We agree with the admonition in Pears that "a charge of second-degree murder should only rarely be appropriate in a motor vehicle homicide.55 " But Jeffries is distinguishable from the typical intoxicated driver by his heightened awareness of the risk resulting from his past history of drunk driving offenses and the revocation of his license, his extreme level of intoxication, and his inherently dangerous conduct in driving his car directly in front of an oncoming car that had no opportunity to react."56The evidence allowed a reasonable jury to conclude that Jeffries's conduct was not only reckless, but *924also demonstrated extreme indifference to the value of human life.

B. The Superior Court Did Not Abuse Its Discretion by Admitting Evidence of Jeffries's Failure To Complete Alcohol Treatment and His Probation Condition that He Not Consume Alcohol.

Jeffries argues that evidence of his failures to complete court-ordered alcohol treatment and his probation condition which prohibited him from drinking alcohol on the night in question was irrelevant and prejudicial. The court of appeals reasoned that this evidence demonstrated Jeffries's heightened awareness of the risks of his conduct.57" We review the superior court's admission of evidence under an abuse of discretion standard.58 Erroncously admitted evidence is deemed harmless if "it did not appreciably affect the jury's verdict." 59

With regard to the evidence of his failure to obtain substance abuse treatment, Jeffries argues that while "a rational inference of knowledge [of the risk of drinking] may be based upon completed treatment, a similar inference [cannot] be based upon a failure to complete treatment." ©

Jeffries's argument is unpersuasive. A jury could infer that Jeffries's awareness of his dangerousness to society was heightened by the court orders to participate in the substance abuse programs as a condition of probation.

Jeffries argues that the evidence that he was prohibited from drinking on the night of the accident by the terms of his probation is irrelevant because it "does not tend to establish a knowledge of the dangers of driving while intoxicated." We reject this argument as well. As a result of his persistent drunk driving a probation condition ordered Jeffries not to drink. The jury could infer that this probation condition reinforced Jeffries's awareness that his drinking and driving was a serious matter that posed such a great risk to the community that it was deemed necessary to prohibit him from drinking altogether.

Finally, Jeffries argues that the probative value of both the probation condition and his repeated failure to participate in the treatment program was "outweighed by the danger of unfair prejudice."60 As noted above, this evidence was probative of Jef-fries's heightened awareness of the risk of his actions. We acknowledge that evidence that Jeffries violated court orders could tempt the jury to form negative opinions about Jeffries's character. But we hold that the probative value of this evidence outweighs any unfair prejudice Jeffries might have suffered. Because it is usually impossible to present direct evidence of a defendant's mental state, indirect evidence that establishes that the defendant was informed of the dangerousness of his actions is often essential to establish the defendant's awareness of the risk. In this case, the challenged evidence is highly probative.

IV. CONCLUSION

We hold that an intoxicated driver may be guilty of extreme-indifference murder if all four Neitzel factors, taken together, permit a reasonable jury to find extreme indifference to the value of human life. Evidence of extreme intoxication, inherently dangerous conduct while driving, and heightened awareness of the dangers of driving while intoxicated is highly relevant to this determination. Because a reasonable jury could find Jeffries guilty of extreme-indifference murder on the evidence presented, we AFFIRM the decision of the court of appeals that affirmed his judgment and commitment.

MATTHEWS, Justice, with whom FABE, Justice, joins, dissenting.

with whom FABE, Justice, joins, dissenting.

Today's opinion holds that a drunken driver who causes a fatal accident can be guilty *925of reckless murder if he has a high blood aleohol content and a prior record of driving while intoxicated. I do not join in this conclusion.

The legislative history of our reckless murder statute makes it clear that the statute was meant to be confined to cases in which reckless conduct closely resembles an intentional or knowing murder. Conduct was contemplated that is similar in degree of risk to that which is encompassed in the phrase that governs knowing or intentional second-degree murder: "substantially certain to cause death or serious physical injury."1 The examples of such conduct offered by the senate committee which recommended the adoption of the statute were shooting into a tent or persuading a person to play Russian roulette.2

In my view, drinking too much-even way too much-and then attempting to drive safely is not substantially certain to cause death, nor is it comparable in terms of risk, or in terms of utility or anti-social mind-set, with these examples. As the court of appeals observed in Nettzel v. State:

For both murder and manslaughter, the harm to be foreseen is a death. Therefore, the significant distinction is in the likelihood that a death will result from the defendant's act. Where the defendant's act has limited social utility, a very slight though significant and avoidable risk of death may make him guilty of manslaughter if his act causes death. Driving an automobile has some social utility although substantially reduced when the driver is intoxicated. The odds that a legally intoxicated person driving home after the bars close will hit and kill or seriously injure someone may be as low as one chance in a thousand and still qualify for manslaughter. Where murder is charged, however, an act must create a much greater risk that death or serious physical injury will result, [3]

I agree with those cases and authorities that suggest that if a fatality caused by a drunken driver is to be murder rather than manslaughter, the driver must have engaged in egregiously unsafe maneuvers such as extreme speeding, wrong-way driving, or running stop lights.4 Merely being drunk and *926attempting without success to drive normally can be manslaughter, but not murder.5

Today's opinion may generate an unintended result. A prosecutor now will be able to charge murder whenever a repeat offender with a high blood aleohol reading causes a fatal accident. The murder charge will go to the jury and the jury will be instructed that it may consider the defendant's prior DWI record. But the defendant's record might not be admissible if the charge were merely manslaughter rather than murder.6Given the likely impact on the jury's deliberations of a prior history of DWI convictions, it may turn out to be easier in such cases to obtain a conviction of murder than a conviction of manslaughter.

For the reasons stated, I would revers‘e Jeffries's second-degree murder conviction.