THE RELEASE last week of a report commissioned by the Glendening administration on the effect of race and geography on Maryland's death penalty is likely to start as many arguments as it settles. Most of the initial reaction to the report has focused on its conclusions that race and geography play a disproportionate role in determining who will face execution.

These factors are disturbing and a cause for thoughtful concern, whatever one's view of the death penalty. And they forcefully rebut Gov. -elect Robert L. Ehrlich Jr.'s charge that the death-penalty moratorium imposed by Gov. Parris N. Glendening while waiting for the report was "politically motivated." To the contrary, those who support the death penalty have an even greater stake in ensuring that the "machinery of death" is perceived to be fair and equitable. Yet the report's conclusions, disturbing as they are, are flawed. They distract attention from the fact that there is an unacceptable risk that people - white or black - will be executed for crimes they did not commit. This risk transcends race and politics, but the report does not address it. The new governor and the General Assembly must do that.

The report's most surprising finding, drawn from an exhaustive statistical analysis of nearly 6,000 Maryland homicide cases dating back more than 20 years, is that, taken alone, a defendant's race is irrelevant in determining whether the death penalty is imposed. "There is no evidence that the race of the defendant matters at any stage once case characteristics are [accounted for]." In other words, an African-American charged with capital murder in Maryland is no more likely to face execution than a white charged with the same crime.

The victim's race, on the other hand, appears to matter very much. "Offenders who kill white victims, especially if the offender is black, are significantly and substantially more likely to be charged with a capital crime."

That the victim's race appears to be more important than the defendant's in deciding who will die is closely related to another finding that is no less striking. That's the critical role played by individual prosecutors across the state in deciding whether to seek the death penalty.

As the arcane statistical jargon of the report puts it, "When the prosecuting jurisdiction is added to the model, the effect for the victim's race diminishes substantially, and is no longer statistically significant."

Demographically, jurisdictions with high rates of death-penalty prosecutions - Baltimore and Harford counties are the top two - also have many white victims. The jurisdiction with the lowest rate of death-penalty prosecutions - Baltimore City - has more black victims. A "death-eligible" murder committed in Baltimore County is 13 times more likely to result in prosecutors' seeking the death penalty than a virtually identical murder committed in Baltimore City. Similarly, Prince George's County has a relatively high proportion of black victims and low rate of death-penalty prosecutions.

If the jurisdictions with the highest rates of death-penalty prosecutions are predominantly white and the jurisdictions with the lowest rates predominantly black, it's not surprising that death-penalty prosecutions disproportionately involve white victims. It does not make the victim's race a cause of the decision to seek the death penalty (in any particular case).

Indeed, Baltimore County State's Attorney Sandra O'Connor argues that her policy of limiting prosecutorial discretion by seeking the death penalty in virtually every case permitted by statute is the best way to minimize the risk of discriminatory prosecutions. There's a certain primitive logic to her argument, even if it amounts to a practice of avoiding the appearance of discrimination by trying to kill as many people as possible.

The report's focus on race and geography leaves open the question of whether we are at risk of executing an innocent person. The answer, clearly, is yes.

Each year, I begin my class in courtroom evidence and trial procedure by asking my law students, "Do we have time machines?" The question is fanciful, but the point is deadly serious. We don't have time machines, so it isn't possible to travel back to the scene of a crime to see what happened. When we need to know something about the past that has serious legal consequences, such as who committed a murder, our only choice is to hold a trial, consider the evidence and reach a "verdict," which serves as a necessary substitute for our knowing the answer, and as a society we have agreed to accept the verdict as truth. As columnist Charles Krauthammer observed, "A civilized society needs such conventions to resolve the unresolvable. Even though we know that the ultimate truth can never be known, we ask 12 ordinary citizens for their opinion and make that our truth."

So long as we use procedures calculated to maximize accuracy, we can be reasonably confident we get the answer right most of the time. But there are risks. One is that occasionally, a jury gets it wrong. Either a guilty person will be acquitted or an innocent person will be convicted. This is inevitable.

Yet, we must have criminal trials, do our best and, when someone is convicted of a serious crime, send him to prison. We must do so even though we realize that some people will go to prison for crimes they didn't commit. There is simply no alternative to having a functioning criminal justice system in a civilized society.

Practical matter

But it is possible to have a functioning criminal justice system without a death penalty. The rest of the world's leading democracies (excluding Japan, which still executes convicts) are doing fine without the death penalty - better, some might argue, than the United States.

For those who are unpersuaded by the morality of the issue, there is the matter of practicality, which is where I come down.

When I'm not teaching law students, I try cases for a living.

Unfortunately, I've lost cases I should have won and - truth be told - won cases I should have lost. The system is not perfect, but it's not designed to be. The nature of the litigation process compels judgments based on probabilities, not certainties. Trials are untidy ways of determining facts.

In trials, we use the best evidence available at the time - evidence that is often flawed and more incomplete than many people imagine. But we recognize that society doesn't have unlimited time and resources to devote to each case - what Oliver Wendell Holmes famously referred to as a "concession to the shortness of life." We can tinker with the fact-finding process and try to make it better, but no matter what we do, it will never be possible to get the correct result 100 percent of the time.

When advocates of capital punishment say they're for the death penalty, but only when we're "sure" of the defendant's guilt, what they're really saying is that they are against the death penalty. No matter how certain we think we are in any particular case, there will be times we're wrong.

Too great a price

In more than 100 documented cases during the past two decades, death row inmates have been exonerated, usually because of DNA evidence. Virtually every prosecutor, judge and juror in those trials was undoubtedly "sure" of each defendant's guilt. But as we now know, they got it wrong.

There are many convictions in which no DNA evidence exists. Not infrequently, convictions are based largely on the testimony of one very believable eyewitness who identifies the defendant to the satisfaction of the jury. Sometimes, very believable witnesses are mistaken or lying.

That's just the way it is: We don't have time machines and we never will. No matter how hard we try, no system of evidence and trial procedure can get it right all the time, and when the system inevitably gets it wrong, we can't ensure that we will discover the error in time.

We are left with the unavoidable question of whether society's need for the death penalty is worth the price of sometimes executing an innocent person. To me, the answer to that question is the only aspect of the criminal justice system absolutely free from doubt: Of course not.

Andrew D. Levy is a trial lawyer in Baltimore and an adjunct professor at the University of Maryland School of Law.