RALPH B. GUY, Jr., Circuit Judge.

A jury awarded Doris Berry (plaintiff), six million dollars for the death of her son, Lee Berry, Jr., who was shot by a Detroit police officer. On appeal, the City of Detroit (City) raises three arguments. First, the City asserts that plaintiff failed to offer any evidence that the City’s policymakers had adopted a policy or ratified a custom of unconstitutional use of deadly force. Second, the City argues that plaintiff’s evidence of the City’s alleged failure to train and discipline its officers failed to show any custom or policy of the City that amounted to deliberate indifference to the rights of citizens or, alternatively, that there was no evidence that any alleged failure on the City’s part was a *1344 proximate cause of Lee Berry’s death. Finally, the City takes issue with the jury instructions. We conclude that plaintiffs evidence was insufficient to prove deliberate indifference on the part of the City such as to give rise to municipal liability under 42 U.S.C. § 1983. Therefore, we reverse the judgment against the City.

On June 23, 1987, Lee Berry, Jr. (Lee), who was employed in a family moving business, was driving the company van in rush-hour traffic to the family residence in Detroit. According to the testimony of Lee’s sixteen-year-old brother, Dwayne Berry, and his eight-year-old nephew, David Askew, both of whom were in the van, Lee committed several misdemeanor traffic violations during this journey. The infractions consisted of running a red light, driving on the wrong side of the road, and passing three cars that were stopped for a red light.

These traffic offenses attracted the attention of Officer Joseph Hall, a 17-year veteran of the Detroit Police Department. Exactly what happened immediately thereafter, however, is a matter of some dispute. On the witness stand, Hall testified — in contrast to the testimony of the surviving occupants of the van and in partial contradiction of his earlier deposition testimony — that the violations prompted a dangerous, high-speed chase involving the police. Dwayne and David, on the other hand, offered a markedly different account of what transpired. They testified that Lee had not been speeding, had not attempted to flee from arrest, and had not committed any traffic violations other than those described above.

Moments after the van arrived at the Berry family home, Hall arrived on the scene. Hall confronted Lee Berry and a struggle then ensued, during which Hall shot Lee in the back from a distance estimated as being between three and ten feet. 1 Although no eyewitnesses observed the shooting, David, Dwayne, and several bystanders stated that Hall, upon his arrival at the Berry home, used profanity, threats, and physical force in attempting to effect an arrest. Hall, however, claimed that Lee attacked him; that the gun fired accidently when Lee, during their struggle, attempted to wrestle the pistol away from him; and that he continued to fire at Lee while Lee was fleeing because the firearm was set on “automatic” and because his (Hall’s) vision had been impaired by blood. Plaintiffs expert witnesses opined that Lee had not touched Hall’s revolver because Lee was several feet away from the gun when it fired, that Hall had not been blinded by blood, and that Hall’s service revolver lacked any type of “automatic” firing mechanism.

Shortly after the incident, the police department conducted an investigation that exonerated Hall. No criminal charges were brought against him and no sanctions were imposed upon him. At the time of trial, Hall remained a member of the Detroit Police ■Department on disability retirement with full pension benefits.

Plaintiff initiated this suit against Hall and the City 2 under 42 U.S.C. § 1983. Plaintiffs complaint also listed several state law causes of action. Plaintiff claimed that the City pursued a deliberate policy of failing to train- or discipline adequately its police officers in the proper use of deadly force, which failures caused the violation of Lee’s constitutional rights under the Fourth and Fourteenth Amendments. At the conclusion of the trial, the jury returned verdicts against both Hall and the City in the amount of six million dollars.

Both Hall and the City appealed from the denial of their motions for j.n.o.v. or for a new trial. While this appeal was pending, the parties settled Hall’s appeal and jointly moved for an order remanding that appeal for consideration of the settlement. As a result of this settlement, the City’s obligation under the judgment was reduced to two and

*1345 one-half million dollars. The only question before this court is the issue of municipal liability for Lee Berry’s death.

II.

Pursuant to the Detroit City Charter, the Board of Police Commissioners (Board) is the agency with authority to establish City policy for the Detroit Police Department. The chief of police administers the department pursuant to the policies, rules, and regulations established by the Board. The City argues that plaintiff failed to introduce any evidence of a custom or a deliberate choice of the Board to adopt a policy of inadequate training or discipline with regard to the use of deadly force, and that evidence of such a custom or choice is necessary to sustain a judgment against the City under 42 U.S.C. § 1983. This is an incorrect statement of the law.

Any discussion of municipal liability under § 1983 begins with Monell v. Department of Social Services, 436 U.S. 658, 98 S.Ct. 2018, 56 L.Ed.2d 611 (1978). In Monell, the Supreme Court held that municipalities can be sued directly under § 1983 where the action of the municipality itself can be said to have caused the harm, as when “the action that is alleged to be unconstitutional implements or executes a policy statement, ordinance, regulation, or decision officially adopted and promulgated by that body’s officers.” Id. at 690, 98 S.Ct. at 2035-36. Municipal liability for the actions of employees may not be based on a theory of respondeat superior. Monell went on to clarify that municipal liability may be predicated upon grounds other than explicit expressions of official policy:

Moreover, although the touchstone of the § 1983 action against a government body is an allegation that official policy is responsible for a deprivation of rights protected by the Constitution, local governments, like every other § 1983 “person,” by the very terms of the statute, may be sued for constitutional deprivations visited pursuant to governmental “custom” even though such a custom has not received formal approval through the body’s official decisionmaking channels.

Id. at 690-91, 98 S.Ct. at 2036.

More recently, the Court addressed the issiie of municipal liability under § 1983 when the plaintiff has alleged that his or her rights were violated as a result of a municipality’s failure to train its police officers adequately. In City of Canton v. Harris, 489 U.S. 378, 109 S.Ct. 1197, 103 L.Ed.2d 412 (1989), the plaintiff had been arrested and brought to the police station, whereupon she slumped to the floor on two separate occasions and was left lying there. The plaintiff brought a § 1983 action against the city for failure to train its officers to recognize when a person in custody is in need of medical assistance. The Court held that

the inadequacy of police training may serve as the basis for § 1983 liability only where the failure to train amounts to deliberate indifference to the rights of persons with whom the police come into contact. This rule is most consistent with our admonition in Monell, 436 U.S., at 694 [98 S.Ct., at 2037-38], and Polk County v. Dodson, 454 U.S. 312, 326 [102 S.Ct. 445, 454, 70 L.Ed.2d 509] (1981), that a municipality can be liable under § 1983 only where its policies are the “moving force [behind] the constitutional violation.” Only where a municipality’s failure to train its employees in a relevant respect evidences a “deliberate indifference” to the rights of its inhabitants can such a shortcoming be properly thought of as a city “policy or custom” that is actionable under § 1983.

Id., 489 U.S. at 388-89, 109 S.Ct. at 1204-05 (footnote omitted).

The City observes that the only policy decision it made concerning the use of deadly force by police officers was the deadly force policy that the Board formally established on June 19, 1986. No one disputes that this policy, as written, is constitutional. The City argues that, because there was no other formal policy decision made by the Board, it cannot be held liable. The Supreme Court, however, has concluded that there need not be a formal policy for there to be an unconstitutional custom that amounts to a policy. “It would be a narrow conception of jurispru *1346 dence to confine the notion of ‘laws’ to what is found written on the statute books, and to disregard the gloss which life has written upon it. Settled state practice ... can establish what is state law.” Nashville, Chattanooga & St. Louis Ry. v. Browning, 310 U.S. 362, 369, 60 S.Ct. 968, 972, 84 L.Ed. 1254 (1940) (quoted in Monell, 436 U.S. at 691 n. 56, 98 S.Ct. at 2036 n.56). In City of Canton, the Court clarified this pronouncement when it concluded that only where a city’s failure to train “evidences a ‘deliberate indifference’ to the rights of its inhabitants can such a shortcoming be properly thought of as a city ‘policy or custom’ that is actionable under § 1983.” 489 U.S. at 389, 109 S.Ct. at 1205. Here, plaintiffs theory is that Hall’s behavior was not an isolated incident but, rather, emblematic of a consistent and pervasive pattern of misuse of fatal force by Detroit police officers. Plaintiff asserts that the City’s failure to train the officers appropriately and to discipline them when they violated policy amounted to deliberate indifference to the rights of the City’s inhabitants.

[I]t may happen that in light of the duties assigned to specific officers or employees the need for more or different training is so obvious, and the inadequacy so likely to result in the violation of constitutional rights, that the policymakers of the city can reasonably be said to have been deliberately indifferent to the need. In that event, the failure to provide proper training may fairly be said to represent a policy for which the city is responsible, and for which the city may be held liable if it actually causes injury.

Id. at 390, 109 S.Ct. at 1205 (footnotes omitted). Thus, in order to prevail, plaintiff was not required to produce evidence of any deliberate or conscious choice by the Board to adopt an unconstitutional policy or custom.

In addition to administering the police department, the chief of police is responsible for training and internal discipline. The manner in which a police force is trained “is fairly attributed to a municipality when it is designed, implemented, and its trainees supervised by municipal officials to whom the municipal governing body has effectively delegated final authority so to act.” Spell v. McDaniel, 824 F.2d 1380, 1390 (4th Cir.1987), ce rt. denied, 484 U.S. 1027, 108 S.Ct. 752, 98 L.Ed.2d 765 (1988). “[T]he mere fact that the governing body or some other official retains ultimate but unexercised authority over both police training policy and policymakers does not allow a municipality to disclaim the policy as its own.” Id.

The City contends that it is not liable because the record here shows that its written deadly force policy was constitutional and because any failure to enforce that policy through appropriate training or discipline would be a departure from, rather than an implementation of, such policy. This theory was expressly rejected in City of Canton. 489 U.S. at 387, 109 S.Ct. at 1204 (“[W]e reject petitioner’s contention that only unconstitutional policies are actionable under the statute.”).

III.

We next turn to what we consider to be the heart of this case: Was the evidence sufficient for a jury to conclude there was such a consistent and pervasive pattern of unconstitutional use of deadly force, evidencing a failure to train or discipline police officers, that it amounted to the City being deliberately indifferent to the rights of its citizens? To establish liability under a failure to train theory, “the plaintiff must prove: that the training program is inadequate to the tasks that officers must perform; that the inadequacy is the result of the city’s deliberate indifference; and that the inadequacy is ‘closely related to’ or ‘actually caused’ the plaintiffs injury.” Hill v. McIntyre, 884 F.2d 271, 275 (6th Cir.1989) (citing City of Canton).

In reviewing the City’s arguments concerning the sufficiency of the evidence, this “‘court must view the evidence in the light and with all reasonable inferences most favorable to the party who secured the jury verdict.’ ” Igo v. Coachmen Indus., 938 F.2d 650, 655 (6th Cir.1991) (quoting Portage II v. Bryant Petroleum Corp., 899 F.2d 1514, 1523 (6th Cir.1990)). A motion for a j.n.o.v. shall be granted only when the evidence is so one- *1347 sided that there can be but one conclusion drawn from the evidence offered. Id. This court “must not pass on the credibility' of witnesses, weigh the evidence, or substitute our judgment for that of the [factfinder].” Agristor Leasing v. A.O. Smith Harvesters Prods., 869 F.2d 264, 268 (6th Cir.1989).

The testimony concerning the training that Detroit police officers received upon joining the force was not disputed. As of the time of trial, officer candidates at the Detroit Police Academy underwent 600 to 700 hours of training. During their time at the Academy, they were given 60 hours of firearms training and were required to take both a certification examination and an additional written examination. Plaintiff offered no evidence regarding any specific deficiencies in the initial training. Indeed, plaintiffs own expert testified that the initial training program was adequate:

I have no problems with the training programs, I think the training programs are fine. The problem is that when you get these officers that have been on the street 10 or 12 years, they don’t follow Department policy. It has nothing to do with their basic training.

(App. 769.)

The evidence further revealed that officer candidates at the Detroit Police Academy were trained extensively, both in the field and in the classroom, as to the use of fatal force. In fact, the number of compulsory instructional hours exceeded the minimum figure mandated by the state. Every graduate of the Academy was required to score 100 percent on a written fatal force policy examination, and they each received a legal manual that set forth the deadly force policy. This legal manual was updated periodically after graduation.

Moreover, Detroit police officers continued to receive training even after they graduated from the Academy and joined the force. Each officer was compelled to attend an annual refresher course concerning the use of deadly force, and officers frequently received training bulletins and updates during roll call. Forty hours of in-service training was required annually. Officers also had to qualify annually in firearms usage.

In Hall’s case, his records indicate that he qualified for firearms each year from 1975 through 1982 and in 1984. There is no indication of testing in 1985, 1986, or 1987. Plaintiff argues that from Hall’s records the jury could have chosen to discount the City’s assurances that officers were required annually' to qualify with a firearm. However, Hall’s wrongful actions here stemmed not from a lack of knowledge as to how to use his firearm but from a lack of discretion as to when to use it.

In .this regard, the official written rules and policies promulgated by the Board concerning the use of deadly force by Detroit police officers embraced the constitutional criteria dictated in Tennessee v. Garner, 471 U.S. 1, 105 S.Ct. 1694, 85 L.Ed.2d 1 (1985). In Gamer, the Supreme Court concluded that the Fourth Amendment mandated that resort to fatal force by a police officer to effect an arrest was constitutionally permissible only if the officer had probable cause to believe the suspect posed a significant threat of death or serious injury to the officer or to others. In Pleasant v. Zamieski, 895 F.2d 272, 274 (6th Cir.), cert. denied, 498 U.S. 851, 111 S.Ct. 144, 112 L.Ed.2d 110 (1990), we concluded that Detroit’s written deadly force policies did not violate the Constitution.

In addition, the evidence at trial disclosed that these policies were well known to all members of the police department, including Hall. In sum, the evidence developed during the trial was simply insufficient to support the jury’s finding that the City pursued an official policy or custom of failing to train its officers adequately in the proper use of deadly force. 3 We now turn to plaintiff’s failure to discipline theory.

There are three primary problems with the award of a money judgment against the City under a failure to discipline theory.

*1348 1. Jury Perception

First, there is the problem of the jury’s perception of a case of this nature. The jury determined that plaintiffs damages were six million dollars. The liability of the police officer and the City was joint and several. The jury was not told that, under the collective bargaining agreement existing between the City and the police, the City was obligated to pay any judgment of this nature rendered against an officer. 4 Jurors would be less than human if they did not factor in a plaintiffs chances of ever directly collecting six million dollars from a police officer. In short, there is tremendous pressure to find a deep pocket if the jury believes the wrong committed was egregious, as they justifiably could have concluded here.

It is also significant that, pursuant to the Supreme Court’s holding in Monell, there is no respondeat superior liability as to municipalities. 436 U.S. 658, 98 S.Ct. 2018. Although a shooting looks like a garden-variety tort, a jury is, in this context, prohibited from making the master liable for the tort of the servant, unless it goes through the “custom or policy” analysis and ties in a city in that manner.

Although the Supreme Court never has suggested that we look with special care at this type of judgment rendered against a municipality, we do not decide cases in a vacuum. Particularly when we deal with sins of omission, as opposed to sins of commission, we should be certain the causal linkage is present that makes the omission the proximate cause of the wrong. To illustrate, if a city had an announced policy that deadly force could be used against fleeing felons, regardless of the circumstances, one would have little problem holding that city responsible if a person is wrongfully injured or killed by a police officer using deadly force. On the other hand, if a city has a proper policy and it is not followed, one has to speculate on the reasons why any one of that city’s officers (here, the 5,000 Detroit police officers) would have failed to follow the policy.

Stated another way, it is a familiar concept in the law to hold that a policy that instructs a subordinate to act in a certain manner implicates both the policy and its promulgators when the subordinate follows the policy. The reverse does not hold true, however. An order not to do something that is disobeyed opens up a whole host of plausible possibilities as to why the order may not have been followed.

2. Nature of Expert Testimony

The second problem with the award of a money judgment in this instance stems from the nature of the “expert” testimony offered by plaintiff. There are two facets to this problem. First, we have grave concerns about the qualifications of plaintiffs witness, Frederick Postill, to opine as he did. As we have now defined the issue on appeal, Postill’s testimony must have provided the basis for the jury to determine that the alleged failure of the Detroit Police Department to discipline properly officers other than Hall was the proximate cause of Hall shooting Lee Berry. As we view the record, Postill did not have the qualifications to testify as an expert on this question, and, if he did, no proper foundation was laid for his ultimate opinion.

In order not to be simply conclusory on this point, a detailed review of Postill’s credentials must be undertaken. Postill received a degree in sociology in 1971, and a master’s degree in education in 1976. He also took courses in criminal justice, but how many and what kind was never revealed. His law enforcement career started when he received an appointment as a deputy sheriff in 1966, a position for which, Postill admits, no qualifications were required. Equally important, however, is that he did not receive formal training. Postill simply started and worked with more experienced officers. The only information conveyed by the record as to what he did or learned as a deputy sheriff is' the following:

*1349 Q. All right. During your tenure as a deputy sheriff, could you tell us what types of duties you were involved in?

A. Basically, general patrol, investigation work, riding a black and white, responding to calls.

Q. Were you ever certified to instruct as a deputy sheriff other members of — of the Sheriffs Department?

A. I was an instructor there in the Department. I taught defensive tactics, things of that nature.

Q. What do you mean by “defensive tactics”?

A. Physical control of subjects that you would be arresting.

(App. at 679.)

Postill also admits that he was fired twice during his two-year stint as a deputy, apparently as a result of trying to organize the deputies into a labor union against the wishes of the incumbent sheriff. Postill’s disagreement with the sheriff culminated in his decision to run against the sheriff in the 1972 election. Although his election bid was successful, the electorate decided to retire him four years later. Interestingly, not one question-was asked during trial to elicit what Postill did or learned during his four years as sheriff, a job for which the only necessary qualification is the ability to get elected.

Postill’s testimony as to what he did after being deposed as sheriff illustrates the problem that courts and juries have with expert testimony, particularly of the type offered here. If an expert has a degree in electrical engineering, there are some assumptions that safely can be made relative to his general training, since all electrical engineers receive similar training. In this same vein, if an electrical engineer says his first job was designing generator motors, there is some notion of what that job entails. But when a sociologist cum sheriff is allowed to testify as to all manners of police practices and procedures, the slopes become slippery indeed.

Postill’s next job after being sheriff was to work for four years for the Justice Department where he “developed the training criteria to train sheriffs and managers of large sheriffs[’] departments_” (Trial Tr. Vol. Xlla at 7.) Since up to this point in time he had had little or no training himself, exactly why he got this assignment is difficult to understand.

From 1981 to the present, Postill appears to have conducted seminars in police management techniques; taught; formed a one-man corporation called “Criminal Justice Consultants”; testified in court; and purchased a hunting and fishing resort in Canada, where he spends five months a year. The fact that some of this would sound impressive to a jury is its vice, not its virtue. The fault lies not with Postill but, rather, with the system. The courts have had a difficult time in appropriately cabining the opinion testimony of “scientific” experts. At least with scientific testimony there is some objective criteria to look to and there is some basis for evaluating a given expert’s testimony against the scientific norm on the subject about which an opinion is offered.. The problem that “junk science” has caused in the courtrooms has not gone unnoticed. 5 The Supreme Court recently revisited the controversy that has swirled around expert testimony for the last several years. In Daubert v. Merrell Dow Pharmaceuticals, Inc., - U.S. -, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993), the Court made a number of clarifying pronouncements. These holdings, however, are in the context of scientific evidence, and thus are only of limited help in the instant case. 6 It would appear obvious, however, that evidentiary problems are exacerbated when courts must deal with the even more elusive concept of non-scientific expert testimony, as is the case here.

The distinction between scientific and nonscientific expert testimony is a critical one. By way of illustration, if one wanted to ex *1350 plain to a jury how a bumblebee is able to fly, an aeronautical engineer might be a helpful witness. Since flight principles have some universality, the expert could apply general principles to the case of the bumblebee. Conceivably, even if he had never seen a bumblebee, he still would be qualified to testify, as long as he was familiar with its component parts.

On the other hand, if one wanted to prove that bumblebees always take off into the wind, a beekeeper with no scientific training at all would be an acceptable expert witness if a proper foundation were laid for his conclusions. The foundation would not relate to his formal training, but to his firsthand observations. In other words, the beekeeper does not know any more about flight principles than the jurors, but he has seen a lot more bumblebees than they have.

Carrying this example into the instant case, it is clear that Postill’s testimony, if admissible at all, would have been from the viewpoint of the beekeeper, not the aeronautical engineer. If there is some formal training that would allow one to testify from a scientific standpoint on how failure to discipline officer “A” would impact on the conduct of his peer, officer “B,” it is clear that Postill does not have such training. Thus, for this kind of testimony to be admissible, a foundation would have to have been laid based upon the witness’s firsthand familiarity with disciplining police officers and the effect of lax discipline on the entire force. To the degree one might be tempted to argue that “everyone knows that if discipline is lax more infractions occur,” this argument proves too much. If everyone knows this, then we do not need an expert because the testimony will not “assist the trier of fact to understand the evidence or to determine a fact in is-sue_” Fed.R.Evid. 702.

If, for example, Postill had testified that when he became sheriff there were “x” number of incidents involving alleged excessive force, but two years after he instituted a training program, there were “x-y ” incidents, we would have a starting point. If he then said that after the training program was in place, he increased the regularity and severity of discipline for infractions and incidents fell to “x-y'1,” then at least there would be some basis for his opinions. Here, however, the only discipline experience referenced was the discipline imposed upon Postal.

Although, as indicated, Daubert dealt with scientific experts, its language relative to the “gatekeeper” function of federal judges is applicable to all expert testimony offered under Rule 702. “[Ujnder the Rules the trial judge must ensure that any and all ... testimony or evidence admitted is not only relevant, but reliable.” — U.S. at -, 113

S.Ct. at 2795.

In providing guidance as to how a court should carry out its gatekeeping function, the Court in Daubert indicated that one of the first inquiries should be whether the “theory or technique ... can be (and has been) tested.” Id. at -, 113 S.Ct. at 2796. As pointed out earlier, there is no indication of any testing of Postill’s discipline theory.

Second, the Court noted that “[ajnother pertinent consideration is whether the theory or technique has been subjected to peer review and publication.” Id. at-, 113 S.Ct. at 2797. There certainly is no testimony as to any peer review of Postill’s theory, and, as to publication, the record reveals the following:

Q. Mr. Postill, what articles have you written regarding police — any aspect of police procedure or policy, whatever?

A. I don’t know I’ve written any specific articles on police procedure or policy. I’ve written articles and — I co-authored a textbook on jail administration.

Q. You wrote a textbook on jail administration?

A. That’s correct.

Q. Did you bring that with you today? A. No.

Q. You were asked at the deposition — it was April 28th — about the articles and— and asked to supply us with copies. Did you bring any of your articles that you’ve written with you today in court?

A. No.

Q. Okay. Did you bring that alleged textbook that you wrote?

*1351 A. No,, I didn’t.

(Trial Tr. Vol. Xlla at 27.) If there were any relevant publications, the jury was left in the dark as to what they were.

The Court in Daubert, concludes the gatekeeper discussion by indicating that “general acceptance” still has a bearing on the inquiry. 7 Here again we know nothing about whether the discipline theory offered by Postill was other than his own.

The Daubert Court’s final observation is also pertinent here: “[T]he Rules of Evidence — especially Rule 702 — do assign to the trial judge the task of ensuring that an expert’s testimony both rests on a reliable foundation and is relevant to the task at hand.” — U.S. at-, 113 S.Ct. at 2799. Although Daubert was decided after the trial in the instant case, earlier decisions of our court have provided guidance relative to expert testimony and its admission. One admonition not followed here is that found in United States v. Kozminski 821 F.2d 1186 (6th Cir.1987), affd in part and remanded in part, 487 U.S. 931, 108 S.Ct. 2751, 101 L.Ed.2d 788 (1988), relative to the proffer of an expert. Although the practice is different in some state courts, the Federal Rules of Evidence do not call for the proffer of an expert after he has stated his general qualifications. In Kozminski this court counseled against putting some general seal of approval on an expert after he has been qualified but before any questions have been posed to him. 8 The issue with regard to expert testimony is not the qualifications of a witness in the abstract, but whether those qualifications provide a foundation for a witness to answer a specific question.

In Kozminski the following colloquy occurred after the party offering the expert had set forth his qualifications:

Counsel: Your Honor, we would offer Dr. Stock as an expert in the field of Forensic Psychiatry and offer his expert testimony to the jury.

The Court: I don’t pass on qualifications, it’s up to the jury to. Go ahead and interrogate him.

Id. at 1219.

In the case at bar, after the qualifying questions were asked of the expert, plaintiffs attorney stated:

MR. MATEO: Your Honor, at this point in time I would offer Mr. Frederick Postill as an expert so that he may render opinion testimony as to appropriate police policies and practices.

(Trial Tr. Vol. Xlla at 15.)

Commenting on this type of proffer, Koz-minski counseled:

When the attorney conducting the direct examination of an expert witness concludes that portion of the examination relating to the expert’s qualifications, it is common to so indicate to the court in language not unlike that quoted here. The person offering the expert is not seeking a ruling but, rather, signalling that the preliminaries are over and the other side may now voir dire if they desire. If the other side conducts a voir dire and then objects or just objects without a voir dire, the court must then make a ruling if it is possible to do so. Often it is not because the key question is not the expert’s general qualifications in some field, but whether the precise question on which he will be asked to opine is ■within his field of expertise or meets some other test such as that set forth in the majority opinion. To illustrate here, the question was not Dr. Stock’s qualifications — those were impeccable. The question was what he was going to be asked which would trigger an answer comprising an “expert opinion.” Obviously, no ruling can be made until the expert is actually asked the question.

821 F.2d at 1219-20.

Unfortunately, the trial judge in the case at bar replied to the proffer by stating:

*1352 THE COURT: Very well. The Court will accept the qualifications of the witness as an expert in the field of proper police policies and practices, based upon the testimony as to his background and experience in that — in that field.

(App. at 683.)

The danger in doing what the trial judge did here is that there is no such “field” as “police policies and practices.” Any witness, like our hypothetical beekeeper, may be allowed to opine based on observation and experience if a proper foundation is laid. But here, there was no foundation at all for discipline testimony, even though it would fall under the general label of “police policies and practices.” This term, however, is so broad as to be devoid of meaning. It is like declaring an attorney an expert in the “law.” A divorce lawyer is no more qualified to opine on patent law questions than anyone else, and it is a mistake for a trial judge to declare anyone to be generically an expert.

With all due respect to Mr. Postill, his credentials as set forth in the record do not qualify him to know any more about what effect claimed disciplinary shortcomings would have on the future conduct of 5,000 different police officers than does any member of the jury. Among other things, Pos-till’s testimony assumes that all 5,000 police officers would know enough about the facts of each claimed event to evaluate the sufficiency of the discipline, 9 and that the officers would formulate their future course of conduct accordingly. It also assumes, without any basis in fact or logic, that police officers will be extravagant in their use of deadly force if they know discipline will not be severe if a shooting occurs. We are not talking about cheating on overtime here, or some other minor peccadillo; we are talking about the taking of the life of another person.

Additionally, Postill’s methodology was as suspect as his conclusions. 10 Postill analyzed the “shots fired” reports from the Detroit Police Department for 1982 through 1987. These annual reports consist of statistical information of all shooting incidents for that year summarized in a one-page report. Pos-till concluded that the 636 incidents covered in the reports, involving the discharge of I,538 shots, fired at an average distance of 30 to 35 feet from the officers’ “opponents,” represented a “fairly frequent” use of fatal force by Detroit police officers.' Postill opined that the frequency of the incidents could not be justified by “self-defense,” a term listed for 502 of the incidents.

Postill also testified that 277 of the 636 incidents included shootings at burglary suspects. Postill then testified that even prior to the Supreme Court’s decision in Tennessee v. Gamer, 471 U.S. 1, 105 S.Ct. 1694, 85 L.Ed.2d 1 (1985), the Detroit Police Department had a policy requiring officers to exhaust all reasonable means before shooting at a burglar. Again, it should be emphasized that Postill’s statistics were taken from the one-page summary sheet of shooting incidents for each year. Postill did not review the underlying facts in these cases to determine, among other things, whether the suspect was armed. Pursuant to Gamer, fatal force may be used when an officer has reason to believe that a felon poses a danger to himself or to others.

Postill also testified concerning his review of the actual records of 161 of the 636 incidents. 11 In his opinion, 78 of these inci *1353 dents were “unjustifiable” on the basis of “existing law and Detroit City policy.” Pos-till concluded that out of these 161 incidents only 15 officers were disciplined and, in seven of these cases, the disciplinary action was held in abeyance. Postill then described IB incidents in which he thought police officers should have been disciplined. In one of these incidents, the officer fired his weapon, accidently hitting a bystander, only after the officer already had been shot twice. Postill still felt that the officer should have been disciplined. The fact that the officer fired only after having been shot twice was not volunteered by Postill and was only revealed upon cross-examination.

After describing these 13 specific cases, Postill concluded that “the department tolerates this kind of reckless use of firearms. They either get no discipline or a slap on the wrist.” It was his “opinion that their failure to direct and discipline and train their officers not to use improper deadly force constitute a pattern of gross negligence.” Postill equated “gross negligence” with “deliberate indifference” and defined it as “conscious knowledge of something and not doing anything about it.” 12

This latter testimony was particularly suspect. It was carefully couched in the precise language used in ease law, either indicating a keen awareness by Postill of the direction in which he had to head, or else careful coaching prior to his testimony. “Gross negligence” is not enough to ground liability according to the Supreme Court, so Postill testified that gross negligence equates to “deliberate indifference.” We apparently now have an etymology expert testifying.

We also believe this testimony was received in violation of the Federal Rules of Evidence. Although an expert’s opinion may “embrace[] an ultimate issue to be decided by the trier of fact[,]” Fed.R.Evid. 704(a), the issue embraced must be a factual one. The expert can testify, if a proper foundation is laid, that the discipline in the Detroit Police Department was lax. He also could testify regarding what he believed to be the consequences of lax discipline. He may not testify, however, that the lax discipline policies of the Detroit Police Department indicated that the City was deliberately indifferent to the welfare of its citizens.

It would have been easy enough for the drafters of the Federal Rules of Evidence to have said that a properly qualified expert may opine on the ultimate question of liability. They did not do so. When the rules speak of an expert’s testimony embracing the ultimate issue, the reference must be to stating opinions that suggest the answer to the ultimate issue or that give the jury all the information from which it can draw inferences as to the ultimate issue. We would not allow a fingerprint expert in a criminal case to opine that a defendant was guilty (a legal conclusion), even though we would allow him to opine that the defendant’s fingerprint was the only one on the murder weapon (a fact). The distinction, although subtle, is nonetheless important.

Furthermore, “deliberate indifference” is a legal term, as the questioning of Postill indicated. It is the responsibility of the court, not testifying witnesses, to define legal terms. The expert’s testimony in this regard invaded the province of the court. The courts have addressed this issue, and typical *1354 of their holdings is that reached in Hygh v. Jacobs, 961 F.2d 359 (2d Cir.1992). Jacobs was an excessive police force case. The expert was allowed to testify that “in his opinion, the use of a baton or flashlight to strike a person in the head would constitute ‘deadly physical force’ that would not be ‘justified under the circumstances.’” Id. at 361-62.

In condemning this line of testimony, the court stated: “This circuit is in accord with other circuits in requiring exclusion of expert testimony that expresses a legal conclusion.” Id. at 363. In concluding on this issue, the court held:

Even if a jury were not misled into adopting outright a legal conclusion proffered by an expert witness, the testimony would remain objectionable by communicating a legal standard — explicit or implicit — to the jury. Whereas an expert may be uniquely qualified by experience to assist the trier of fact, he is not qualified to compete with the judge in the function of instructing the jury....

Far more troubling, [the expert] testified that Jacobs’ conduct was not “justified under the circumstances,” not “warranted under the circumstances,” and “totally improper.” We have held that an expert’s testimony that a defendant was “negligent” should not have been allowed. We see no significant distinction in [the expert’s] con-elusory condemnations of Jacobs’ actions here, which, in the language of the advisory committee, “merely [told] the jury what' result to reach.”

961 F.2d at 364 (citations omitted). If the rule were other than as indicated by the Second Circuit, we would soon breed a whole new category of “liability experts” whose function would be to tell the jury what result to reach — exactly what the expert did here.

3. Failure of Testimony to Support the Jury’s Verdict

The third problem with the award of damages under these circumstances stems from the fact that the testimony offered does not support the jury’s verdict on the failure to discipline issue, since the conduct shown does not amount to “deliberate indifference” as defined by the Supreme Court. Plaintiff asserts that his expert provided sufficient “statistical, anecdotal and expert opinion” from which a jury could find that there was a failure to discipline properly. However, the issue is not whether there may have been a failure to discipline in certain instances. In order for plaintiff to prevail, there must be sufficient evidence for the jury to conclude that the City’s failure in this regard amounted to deliberate indifference to the rights of its citizens, which failure proximately caused the death of Lee Berry.

In the failure to discipline context, it is appropriate to apply the deliberate indifference standard adopted by the Supreme Court in City of Canton v. Harris, 489 U.S. 378, 109 S.Ct. 1197, 103 L.Ed.2d 412 (1989), to require a showing of a history of widespread abuse that has been ignored by the City. Id. at 397,109 S.Ct. at 1209 (“The lower courts that have applied the ‘deliberate indifference’ standard we adopt today have required a showing of a pattern of violations from which a kind of ‘tacit authorization’ by city policymakers can be inferred.”) (O’Con-nor, J., concurring in part and dissenting in part); see Wellington v. Daniels, 717 F.2d 932, 936 (4th Cir.1983).

In this regard, it should be noted that Postill included in the shootings he considered to be “unjustifiable” those that violated Detroit Police Department policy, whether or not they violated the Constitution. The same evidence upon which Postill relied showed that in 38 of the 636 incidents the officer involved was required to attend retraining sessions and an additional 52 officers involved in the 636 incidents were disciplined. Postill based his opinion concerning the 636 incidents on nothing more than raw numbers, with no information surrounding the actual circumstances of the incidents. As for Pos-till’s opinion concerning the 161 incidents about which he felt he had “sufficient” information to base an opinion, at best, it shows that discipline was not as frequent or as severe as he would have liked. It does not show a consistent pattern of ignoring constitutional violations. No other evidence, such as witnesses or even the actual police files, was offered concerning any of the incidents *1355 in which the discipline allegedly was lacking or inadequate.

Cases in which it was determined that plaintiffs produced sufficient evidence showing the existence of inadequate training or discipline, evidencing deliberate indifference on the part of the municipality, stand in stark contrast to the ease at bar. In Spell, 824 F.2d 1380, the plaintiffs evidence consisted of seven lay citizens testifying about brutality complaints they filed that resulted in no corrective action; eight present or former officers of the city police department testifying about a “code of silence” within the department and training in inappropriate methods of subduing suspects; an assistant state district attorney who had investigated and prosecuted officers for the use of excessive force; the former legal advisor to the police department who testified that he left the department because of an attitude favoring the use of excessive force; and internal records of the city police department showing frequent complaints about excessive force that were consistently dismissed or disregarded.

In Fiacco v. City of Rensselaer, 783 F.2d 319 (2d Cir.1986), cert. denied, 480 U.S. 922, 107 S.Ct. 1384, 94 L.Ed.2d 698 (1987), the plaintiff offered evidence that the city’s public safety board had adopted rules relating to police discipline, but that the police chief never had seen them. The rules required complaints to be in writing, yet did not provide a form for civilians to make a complaint. No hearings were held by the public safety board in the five years prior to the plaintiffs injury at the hands of the police. Additionally, the plaintiff called as witnesses four of the six citizens who had actually lodged complaints against officers. Each citizen testified to being physically abused by an officer and complaining to the chief of police. The chief investigated the complaints on his own by simply talking to the officer involved. No investigatory file was created, no formal statement or disciplinary action was taken, and there was not even so much as a note placed in the officers’ personnel files.

In Zuchel v. The City and County of Denver, 997 F.2d 730 (10th Cir.1993), the plaintiffs’ expert “supported his opinion that the training program was inadequate with specific deficiencies supported by the record.” Id. at 739 n. 4. Prior to the incident in which the plaintiffs’ decedent was killed, the Denver district attorney had sent a letter to the police chief discussing the recurring incidents of the use of deadly force by Denver police officers, including five incidents over the previous six weeks where citizens were injured or killed and one in which an officer was seriously injured. The letter made numerous suggestions for improving the training in this area. The plaintiffs presented evidence that even after receiving this letter the police department made no effort to improve its training practices.

Here, the City had an acceptable training and follow-up program to communicate to its officers its constitutional policy on the use of deadly force. It had procedures in place for reviewing instances when that policy may have been violated. Over a five-year period, those reviewing procedures resulted in 38 officers being retrained and an additional 52 officers being disciplined. Plaintiffs evidence was insufficient to enable a reasonable juror to conclude that the City was deliberately indifferent to the rights of its inhabitants.

Finally, the jury never found that the alleged failure to discipline properly constituted “deliberate indifference.” They never reached this conclusion because they never were asked this question. The only question on the verdict form relating to discipline reads as follows:

We, the jury, on the § 1983 federal Civil Rights claim against the City of Detroit for failing to adequately discipline its police officers for the improper use of deadly force find for:

x The Estate of Lee Floyd Berry, Jr.

_City of Detroit

Nowhere did the jury conclude in its answers to the verdict form’s written interrogatories that the failure to discipline constituted “deliberate indifference.” Nor was the question of proximate cause put to the jury in any of the verdict form questions. All of *1356 this amplifies the error in allowing plaintiffs expert to testify .that the failure to discipline constituted “deliberate indifference.”

The jury verdict against the City of Detroit is REVERSED, and the case is REMANDED for an entry of judgment as a matter of law in favor of the City of Detroit. Fed.R.Civ.P. 50(b).

Notes

. The forensic evidence supported the inference that Lee was fleeing from Hall when he was fatally shot.

. The first amended complaint also named Detroit Police Chief William L. Hart as a party defendant. Chief Hart was dismissed from the action on summary judgment, which dismissal is not at issue in this appeal.

. The case was not submitted to the jury under a failure to supervise theory.

. In the recent case of Larez v. Holcomb, 16 F.3d 1513 (9th Cir.1994), the court concluded it was an error requiring reversal for the trial judge to tell the jury that the defendant police officer would be indemnified for any compensatory damages the jury might assess against him.

. See, e.g., Peter W. Huber, Galileo’s Revenge, Junk Science in the Courtroom (1991); Kenneth R. Foster et al„ Phantom Risk, Scientific Inference and the Law (1993).

. The Court stated, relative to its holdings in Daubert: “Rule 702 also applies to 'technical, or other specialized knowledge.’ Our discussion is limited to the scientific context because that is the nature of the expertise offered here.” - U.S. at-n. 8, 113 S.Ct. at 2795 n. 8.

. This language is important since the Court in Daubert rejects, to a large degree, the holding of Frye v. United States, 54 App.D.C. 46, 293 F. 1013 (D.C.Cir.1923), the seminal “general acceptance" case.

. Although the discussion of this issue in Koz-minski is found in a dissent, it is riot the issue on which the dissent is based. The "proffer” issue was raised by the parties in that appeal but not addressed at all by the majority in Kozminski.

. Postill did testify that police departments are a "sub-culture” (no definition given) and word gets around. There is no foundation for these conclusions either.

. On Postill’s direct examination, the following exchange occurred:

Q. And what did you do — what conclusion can you draw from reviewing those 187 shootings instances for the period June 1982 through June 1987?

A. Okay. First let me explain the methodology that I used. This is not what I would consider a scientific study ....

(App. 712) (emphasis added).

. Apparently, Postill was given the records for 187 incidents selected by plaintiff’s counsel. Postill testified that 26 reports did not contain enough information for him to form an opinion. These records had been provided by the City. Plaintiff's counsel stated that he just tried to select various cases where the targets were humans as opposed to animals. The City complains that there was no random sampling and therefore Postill’s testimony could have been skewed. This was made clear to the jury. Regardless of the sampling, Postill was still examin *1353 ing the reports of 187 incidents. Any problem with the selection of those 187 incidents would have a bearing on the weight to be given the testimony, not its admissibility.

. The exact exchange that occurred during direct examination is as follows:

Q. With regard to the history of pattern and practices of the Detroit Police Department, do you believe the Detroit Police Department has had a past pattern and practice of condoning the use of improper deadly force?

A. It’s my opinion that their failure to direct and discipline and train their officers not to use improper deadly force constitute a pattern of gross negligence in my opinion, yes.

(App. 741.)

Q. The phrase "deliberate indifferen[ce]" is a legal term, is that correct?

A. That’s my understanding, yes.

Q. What does it mean?

A. Conscious knowledge of something and not doing anything about it.

Q. All right. Do you have any other way of describing it?

A. Sometimes it's described as being gross negligence when you don’t do something that you obviously need and must do.

(App. 740-41).