Bad Reporting on Matthew Keys' Possible Sentence Conceals Prosecutorial Power

For years I've been grumbling ineffectually about how the media gets criminal justice stories wrong. In particular, I've repeatedly complained that the media distorts stories about federal criminal prosecutions by reporting the statutory maximum sentence that defendants face without noting that the actual remotely plausible sentence they face is usually much lower.

Here's a case in point, and here's why you should care.

Yesterday a federal jury in Sacramento convicted former Reuters journalist Matthew Keys under the extraordinarily flexible, antiquated, and confusing Computer Fraud and Abuse Act. The feds have long relied upon the ambiguity of that statute to pursue whomever they want to pursue. Keys was convicted on the theory that he provided LA Times login information to hackers who made nonsensical edits to an article on the Times' web site.

Although some media outlets reported the result accurately, many emphasized the statutory maximum sentence Keys could face. CNN: "Keys, 28, faces up to 25 years in prison when he is sentenced on January 20, 2016." "The Washington Post — "the end result may mean 25 years in prison." Huffington Post: "Keys Faces Up To 25 Years In Prison." Newsweek: "he faces up to 25 years in prison." Wired: "Keys faces a possible sentence of up to 25 years."

But there's no plausible chance that Keys will get anything like 25 years. The court will use the United States Sentencing Guidelines as a starting point — a recommended sentence — and absent any bizarre factors not present here, will impose a sentence close to that recommendation or below it. Based on the government's outlandish claim that the LA Times hack caused almost a million dollars in damages — more on that in a moment — Keys will likely face a recommended sentencing range of (by my calculation) 51-63 months as a starting point for sentencing arguments,1 and while the court may go considerably lower, it's very unlikely the court will go much higher. Keys' attorneys will get a chance to dispute these calculations, and may convince the court to reject the government's loss calculation (driving the recommended sentence lower) or otherwise adjust it, so that the judge ultimately sentences him based on a different recommended range. But there is effectively no chance that the court will go years higher — a sentence substantially above the guideline range is vulnerable to attack on appeal, and doesn't happen in cases like this.2 That calculation of 51 to 63 months recommended range is consistent with the government's comment that it will seek less than five years.

I can see why Keys would think that this is annoying pedantry, but it's not. Leave aside, for the moment, that the media is misinforming the public about the criminal justice system. Forget the fact-distorting discussions this always creates: "how can you get 25 years for hacking when murderers get out in 10?" The inaccurate reporting is a problem because it conceals a grave problem: the vast power of federal prosecutors to drive the sentence federal defendants get.

Keys gave hackers the ability to deface an LA Times online article for 40 minutes with silly gibberish. The government is arguing that caused $929,977.00 in damages. The jurisdictional cut-off for the charged crime is only $5,000. What part of a response to a hack gets counted as damages, and what doesn't? The line is obscure and flexible. Evidence here suggested that the Times initially calculated damages at $3,800, and that its agents tried to inflate the numbers to make it more likely that the perpetrators would be prosecuted federally. The government can take a very low-key approach and argue that only a low, easily provable number represented damages — or it can aggressively pursue an extravagant theory. If the judge accepts the government's stance — and judges usually do — it makes a very substantial difference in the sentence that the defendant faces. In this case, if the government were only arguing (let's say) $6,000 in damages, Keys would be facing a recommended sentencing range of 15-21 months instead of 51-63 months.3 Those years make quite a difference in somebody's life, and in the decision of whether to go to trial or not. The government can — and routinely does — say to defendants "if you accept this plea we'll stipulate that the damages are only $6,000, but if you go to trial, we're going to argue that they are almost a million dollars."

When journalists report the statutory maximum sentence, they help promote a misleading narrative that prosecutors merely charge cases and judges determine the sentence. In fact, prosecutors have extraordinary power to determine a defendant's likely sentence based on what they choose to charge and what facts they choose to argue. Bad reporting conceals that. Stop it.

Edited to add: Dan Nguyen helpfully points out that in its opposition to Keys' motion to suppress, the government asserted that the labor cost of investigating and responding to the hack was 333 hours worth less than $18,000, but implied that figure didn't include other costs that no doubt inform its nearly-a-million figure. Again, it's about the government's discretion to argue vastly disparate positions that drive a sentence.

  1. The applicable Guideline is 2B1.1. The base offense level is 6, the government's ludicrous damages allegation yields 14 more points under (b)(1)(H), and 4 more levels for the type of computer crime under (b)(17)(A)(ii), for a total offense level of 24. He gets no credit for acceptance of responsibility because he took the case to trial. Assuming he has no criminal record he's in Criminal History Category I. Offense level 24, Criminal History Category I yields a recommended sentence range of 51-63 months. The offense level could go up a few points if the government convinces the judge that he obstructed justice in the course of the case or that he led or organized other people, but that seems unlikely.  
  2. Judges occasionally go way over the guidelines in cases involving child pornography or violence to vulnerable people.  
  3. The calculation is the same as in footnote 1, above, except that between $5,000 and $10,000 yields an increase of 4 levels instead of 14 levels, so the final offense level is 14 instead of 24.  

Last 5 posts by Ken White


  1. shg says

    There will be a PSR. There will be a battle over the guidelines calculation. There will be a dispute as to whether it's $3800 or $929,977. There will come a point in time when an intelligent discussion of Keys' sentencing exposure can be had. That time is not now.

    The compulsion of media to leap over these details to sensationalize the significance of the conviction adds to public ignorance. And it wasn't just non-lawyer journalists doing this, but lawyers who should know better. You are right that writing of statutory exposure serves only to make people stupider, but writing about sentence at all when the battle has yet to be joined in earnest is pure speculation that serves to perpetuate ignorance of the system. While journalists have no concern about this, lawyers should.

  2. Nick42 says

    The inflation of damages has been going on since the 80's. The Hacker Crackdown documented a case where Craig Neidorf stole an 12 page AT&T technical document whose value was calculated to be $80,000. Apparently (per wikipedia) this valuation included the entire cost of the VAX workstation it was typed up on. His lawyer pointed out at trial that AT&T sold the document to the general public for $13.

    I don't have a citation, but I believe there was a relatively recent case that stated that fixing the original vulnerability can now be included when calculating damages.

  3. jdgalt says

    If it takes this much time and effort for even the defendant or his lawyer to figure out (or find out) what the likely sentence will be, then it seems to me the full statutory maximum was at least threatened, and therefore plausible, and those other media should not be blamed for citing it in their stories.

    I believe we need to make it one of the demands of any reform movement that sentences be fixed and predictable much earlier in the process, both to make trials fairer (by letting the jury know the sentence that will likely result from conviction before they convict) and to take away prosecutors' unjustifiable power to punish non-cooperation by increasing the sentence.

  4. zzz says

    @jdgalt But Ken just posted what is the general sentencing guidelines which offer a much more static count of what the sentence will be.

    The variable part in this case comes mainly from the supposed damage calculations.
    If you have a good idea to streamline that/ make it more predictable: A) please let everyone know as it can be applied to pretty much all lawsuits which leads to B) unless it is flexible/limited/abusible enough to not matter, trial lawyers will fight against it tooth and nail making it a real battle to get anything passed.

    The other flexible part is how some statues are vague and flexible enough to apply to anyone the feds want to prosecute, but I don't think you'll talking about that.

  5. MisterFister says

    @jdgalt: Prosecutorial threats to argue for a twenty five year sentence against the realities of how a sentence is treated on appeal (realities that cannot be taught even in three semesters of law school) does not make something "feasible." It is "feasible" that you and I can both win the same PowerBall Jackpot, for each of two different Jackpots, within the next six months. Does that mean that it's responsible for you to go into a car dealership and ask for a car loan by saying, "Well, I COULD come into five hundred million dollars within the next six months, and this is a reasonable figure because we've seen several jackpots in recent years in excess of three hundred million each."

    To your second point — absolutely not. No no no no. It is patently unfair to ask a jury to consider the sentence when deciding on factual guilt. There are few things you could do to make a sentencing system LESS fair than by asking jurors to decide guilt based on likely consequence.

  6. En Passant says

    Ken wrote:

    When journalists report the statutory maximum sentence, they help promote a misleading narrative that prosecutors merely charge cases and judges determine the sentence. In fact, prosecutors have extraordinary power to determine a defendant's likely sentence based on what they choose to charge and what facts they choose to argue. Bad reporting conceals that. Stop it.

    The article you cited above now includes this quotation from a prosecutor:

    The statutory maximum for Keys's crimes is 25 years, but in a statement given after the trial, a spokesperson for the US Attorneys Office said Keys would likely face less than five years.

    "While it has not been determined what the government will be asking the court for, it will likely be less than 5 years," the spokesperson said.

    And this "update" from the Vice reporter or editor:

    Update: This story was updated to include a statement from Special Agent in Charge Miller and further comment from case prosecutors. The original headline of the story has been changed to reflect the US Attorney's statement that the government would likely seek a sentencing of less than five years.

    Whether Vice updated its reportage in response to your critique, or sua sponte, it appears that your objection to Vice's reporting in this particular example has been somewhat mooted by that update.

    Now, may a hundred flowers bloom, to set right the other hundred cases of bad reporting that just happened since you posted your critique.

  7. PonyAdvocate says


    There are few things you could do to make a sentencing system LESS fair than by asking jurors to decide guilt based on likely consequence.

    Yes and no. I agree that as an epistemological matter, factual guilt has nothing to do with what sentence can be imposed. But many think that justice ought to be one of the actual goals of a "justice" system, instead of being an accidental by-product. Consider, for example, the notorious three-strikes laws, where a third criminal conviction, no matter how minor (shoplifting a few candy bars, for example), can result in an extremely harsh sentence being imposed on a defendant. A jury aware of such a consequence might find that such a result "shocks the conscience", and decide to return a verdict of "not guilty", notwithstanding the facts ("jury nullification"). The Fully Informed Jury movement works towards achieving such results, if I'm not mistaken.

  8. Erik says

    As somebody who works in IT security, this is how I see the damages:

    * Sharing the password – $25. This is what your average low-end employee will sell their username / password for. If your systems aren't designed accordingly, then they're out of step with reality.

    * Their IT staff not disabling the account when he was terminated – $Whatever's Left.

    This may not be the IT staff's fault. There are many instances where an employee leaves and their management is wishy-washy about when to remove access just in case they come back. The correct solution in those cases is to disable access but not delete the account, but again, IT reports to management and is subject to stupid decisions just like everyone else.

    Heck sometimes employees leave and nobody tells IT. But they should be auditing their accounts to see who hasn't logged in in the last 3-4 weeks, and raising red flags accordingly. This literally takes just a few minutes, and is rife with opportunity to get back at managers that have pissed you off lately ("this dolt is threatening our security by not telling us when people leave", etc).

    Or their IT department could be some combination of lazy, disorganized, and stupid. I've seen all of the above and more.

  9. Dan Weber says

    Just because my old boss didn't have the locks changed when I quit doesn't mean I'm off the hook for giving out the copy I made to people while encouraging them to do harm.

    The fact that the IT department didn't disable his account is bad form on their part. It doesn't excuse his behavior.

    I have colleagues who do cleanup after break-ins are discovered. For large companies, a six-figure bills is typical.

  10. En Passant says

    Ken White says October 8, 2015 at 10:42 am:

    @En Passant: That's why I linked to that article for a discussion of the case and didn't include it in the paragraph of bad examples.

    D'oh! Thanks for the clarification. I plead caffeine deprivation.

    Now (as of my typing this), the WAPO story (which I now understand was one of the bad examples), contains a brief statement (a standalone sentence displayed as a single 'graf) from the prosecutor:

    Prosecutors said they are likely to ask for less than a five-year sentence but stressed that Keys’s conduct was damaging.

    While that does not entirely undo the bad reporting, it at least vatguely suggests your point that "In fact, prosecutors have extraordinary power to determine a defendant's likely sentence based on what they choose to charge and what facts they choose to argue."

    I agree that bad reporting increases the readership's stupidity quotient. But bad online reportage can be changed quickly to include or more clearly state salient facts. This leaves some readers ignorant, and some more fully informed.

    Heraclitus was certainly right (as reported by Plato) that you can't step in the same river twice. But it appears that on the intarwebz you can't even step into the same river once.

  11. Dan Weber says

    If it takes this much time and effort for even the defendant or his lawyer to figure out (or find out) what the likely sentence will

    It doesn't take much effort. Ken showed in a previous post the charts that are used. It's really rather simple. It's about as much work as calculating fireball damage in D&D.

    But journalists have a groove and they sticking to it, dammit.

  12. Tim! says

    @Erik: what you describe aren't damages. I suppose if you really stretch credibility, the Times could choose to get into the business of selling their own passwords, such that the $25 could be counted as lost revenue. Or I suppose if you stretch in a different way the login details could be considered intellectual property belonging to the Times, and in this case property worth $25 was stolen. It breaks my brain to think too hard about either of those.

    The other cost you bring up is normal cost of doing business during termination of employment for any employee. The Times would (should) have taken on that cost regardless of whether a crime occurred.

    As I understand it, damages are the costs that were paid because the crime happened, that would not have been paid if the crime had not happened. In this case these should include the costs of investigating the mode and extent of the breach and cleaning up the altered data, plus some amount for loss of revenue or reputation.

  13. AH says

    @Tim! I think you missed the point. The point was that most people will sell their password for 25 USD. because of that it's arguably negligent to rely on passwords for security.

  14. Tim! says

    @AH: I understand that Erik was making good points about security best practices. I was addressing specifically his use of the word "damages." It was misused here, and confusing since the main post is about damages from a crime and their effect on sentencing, not about IT security and its implementation costs.

    Nick42 did mention in passing a rumor that the cost of fixing security vulnerabilities could be wrapped into damages when someone exploits them. But nobody has substantiated that rumor, Erik made no mention of it, and even if that rumor were true, I highly doubt that a court would interpret a widespread (if poor) industry practice as a vulnerability for which the defendant should pay the cost of replacing.

    Password-based security is not negligence. It's not negligent if I put a lock on the door to my house, and give everybody who lives there a key. I might be negligent if I forget to ask for the key back when I kick someone out, but that doesn't mean he gets to make copies, pass them out to his friends, and invite them to come pour mayonnaise on my couch.

  15. M. Alan Thomas II says

    Do the districts still vary wildly in percentage of non-plea sentences with upward/downward departures and the ratio thereof? I know they used to.

    Actually, my favorite upward departure was in SDNY—notorious for regular downward departures and no upward departures—on a plea deal(!), much to the consternation of the defendant. It established the precedent, later effectively written into law, that the loss of cultural artifacts could not be measured by dollar value and the guidelines authorized departure to account for losses not measured.

    Mind you, among the plethora of bad moves made in the case was arguing to a federal judge that the letter written by George Washington to John Jay on the occasion of the opening of the first session of the Supreme Court—one of the stolen documents—was not really important or worth anything. I suspect that the federal judiciary thinks that the history of the federal judiciary is kinda important.

    United States v. Spiegelman, 4 F. Supp. 2d 275 (SDNY 1998). The book about it is The Book Thief: The True Crimes of Daniel Spiegelman by Travis McDade.

  16. Ryan says

    Although I'm a fan of her work, Sarah Jeong's Motherboard article regarding the case should be taken to task alongside and above all the other outlets mentioned, representing a distinct flaw in some otherwise great reporting. The URL and the headline shared across Twitter (though not the headline given above the body of the article itself) reads, "Former Reuters Journalist Matthew Keys Found Guilty of Hacking, Faces 25 Years." I understand that the headline was altered, but that's still the message they ran with initially and- through Twitter- the headline being shared to attract readers; visitors to the article don't see the emendation until they've already been exposed to the error.

    This not only quotes the sensationalist "25 years" figure, but, distinctively, doesn't even waive responsibility for that sensationalism with the "up to" crumple zone built into similar headlines. For know-nothing jerks like me, "Faces Up To 25 Years" is a tell that I've learned to spot, but "Faces 25 Years" is a crucial step too far; it makes it sound like sentencing has already occurred, which it hasn't.

    I'm not necessarily coming down on Jeong. I understand that the byline doesn't necessarily write the headline, and that even if Jeong had issue with how Vice/Motherboard presented her work, that bitching about it on Twitter probably wouldn't be prudent, so I don't blame her for the apparent breach. But it seemed an oversight yesterday, as it does today in this article, to see (I assume) Ken on the Popehat Twitter rightly pushing back against the ubiquitous and nefarious quotation of the maximum sentence by press outlets reporting on the case, while linking to the (even worse) Motherboard article a couple hours later, and letting it slide in the roundup of transgressors today.

  17. barry says

    The government is arguing that caused $929,977.00 in damages. The jurisdictional cut-off for the charged crime is only $5,000. What part of a response to a hack gets counted as damages, and what doesn't? The line is obscure and flexible.

    Well that's just nuts! Damages should be the amount it takes to put it back to the way it was before it was damaged, and any punitive on top of that.
    If someone breaks my gate lock and steals a bucket, that's $20 for the lock and $5 for the bucket. If I then pay Blackwater a million dollars to investigate, report, and upgrade to a better security system to protect my property in the future, that's on me, not the guy who stole my bucket.

    What @Tim said:

    The Times would (should) have taken on that cost regardless of whether a crime occurred.
    As I understand it, damages are the costs that were paid because the crime happened, that would not have been paid if the crime had not happened.

    Where security is needed, leaving it weak until someone breaks in should not be encouraged by the courts.

  18. Dan Weber says

    With a physical intrusion, you can easily sweep the place and determine what had happened. But Keys was bragging (in chat logs) about having superuser access and making other users. "Rogue sysadmin" is so expensive to remedy because you can't just lock out his account. You have to sweep every single corner to see if he left any backdoors.

    It's a shame that he was giving away superuser credentials instead of just a regular shmoe's password. But you take your victims as you find them.

  19. LordSega says

    Question… can a defendant use the government's deal offer in it's arguments about the related crime cost? I.e. in your example, that the government's statement of $6000 during deal negotiations should be used. Also can it be argued that the threat of a $1,000,000 cost is duress or blackmail?

  20. barry says

    @Dan Weber

    But Keys was bragging (in chat logs) about having superuser access

    I can see my lock & bucket analogy falls down a bit there. But still, whatever part of the total cost of fixing it went to upgrading the system shouldn't be included (if it was).

  21. Cromwell Descendant says

    Ken, the thing that you totally miss is that even if the media did a better job reporting the facts, the public wouldn't care. Prosecutorial misconduct that only affects people already convicted of some sort of naughtiness, blah blah blah they already stopped listening. They could be torturing this guy to death with some weird partial execution protocol and most people wouldn't care. "He was convicted, right?" is the response you'll get after explaining this shit. People that are willing to consider understanding the problem are not using the "news" as a primary source of explanation or analysis; it only is competent to report the subjects and general locations of events, and quote biased analysis from public figures with no technical qualifications in the subject they're discussing.

    If you want people to care, you need to create some sort of popular TV drama that explores your point of view.

  22. Ssssssss says

    @Cromwell Descendant

    I would watch the **** out of "Law & Order: Prosecutorial Misconduct."

  23. Dan Weber says

    As someone who thinks this guy is a inveterate liar, is a remorseless vandal, is someone with an incredible level of privilege who abused a position of trust that most people will never see, and deserves to spend some time staring at the inside walls of a cell:

    The four years he is actually facing is too entirely too much.

  24. Burnside says


    I can see the show now, and it's beautiful…

    "In the criminal justice system, the people are represented by two separate yet equally important groups: the police who investigate crime and the district attorneys who prosecute the offenders. They are assholes. CHUNG CHUNG"

  25. Lokiwi says

    Is the amount of damages a jury issue? Or is it up to the judge during sentencing once the prosecution proves past the jurisdictional minimum?

  26. says

    Don't they use those maximum sentences to scare people into pleading guilty? That's what they did to Aaron Swartz. As a non-lawyer, I can't help but think that pointing out the maximum sentence is relevant because he was probably originally threatened with that number. Or is that wrong?

    Either way, the "harm" seems to be mostly made up in these type of cases. Like Homer Simpson telling the insurance adjuster what was lost in the fire: "Let's see, there was the Monet and the Picasso…" and the adjuster replies "I'm sorry mister Simpson, but we only cover actual losses."

  27. says

    I laughed out loud when I got to the last page of the Tribune's Damage Expense report. Seriously? Newspaper staffers who make $50.00 an hour? I need to change companies.

  28. Papillon says

    @Kat: It's likely cost to the company, not take home pay. This would include not only benefits but also cost of their cube space, computer, IT support, etc.

  29. Dan Weber says

    This is getting stupid. Just because someone works for a newspaper doesn't mean they are low paid. When a newspaper wants to hire a doctor or a lawyer or a security specialist, they'll need to pay market wages.

    And if you need to bring in outside people to investigate your computers after a rogue sysadmin, you are going to be paying a fucking shitload more than $50 for an hour of their time. In the LA market, they're gonna be bill you for triple or quadruple that.

    $50/hour for security people is an extremely frugal rate. Most of my colleagues would laugh — and I mean literally laugh — at an offer like that in a market like LA.

  30. Robert says

    Ssssssss says

    October 9, 2015 at 4:56 am

    @Cromwell Descendant

    I would watch the **** out of "Law & Order: Prosecutorial Misconduct."

    Err…isn't that just regular "Law & Order"?

  31. Thomas Ptacek says

    Worth knowing, for those of you considering the provenance of damage figures for minor-seeming breakins:

    When large organizations have breaches of any sort, they’re often required (for instance, by their insurers) to engage external “forensics” firms to do an assessment. These assessments gauge the extent of the breach, determine if regulated information was taken (which might necessitate a formal breach disclosure to customers) and attempt to rule out backdoors that might have been installed by attackers.

    A forensics investigation on a newspaper content management system isn’t going to cost $900,000, but it isn’t going to cost $3,000 either. $50,000 is a decent ballpark figure for a reputable firm.

    Costs like these aren’t for closing the vulnerability or improving security or making the next compromise less likely; they’re the direct cost of that particular intrusion.

  32. Cromwell Descendant says

    As to cost of just examining the damages, I do software and system administration consulting. I am a nobody, with no fancy clients. I live in a small market with a low cost of living. The lowest rate I would ever bill is $65/hr, and that rate makes me the lowest priced available consultant in North America on almost every project I bid on. I also have to pay overhead, so that certainly isn't my take-home.

    I would absolutely expect a low end in-house sysadmin with basic competence to be costing $50/hr in LA. Probably about $15 of that is overhead, and $35 is wages+benefits. Assuming they're bottom level, the tech equivalent of a "wrench turner" in a mechanics shop.

    A lower wage IT person isn't actually an IT person, they're probably a telephone support specialist with a flow chart that represents their implied technical knowledge.

  33. MoratNZ says

    A naïve question: how is "if you accept this plea we'll stipulate that the damages are only $6,000, but if you go to trial, we're going to argue that they are almost a million dollars" compatible with telling the truth, the whole truth and nothing but the truth.

    Again, I know it's naïve, but could one not present that in one's defence, saying "look, they said that they'd be happy to call the damages $6k if I pled guilty; now they're saying it was a million dollars. How did my pleading not guilty increase the damage I allegedly did by a couple of orders of magnitude?".

  34. Dwight says


    It goes thusly (and please someone jump in if I'm off here):

    "We accuse you as guilty of this crime. Our opinion is that the evidence is you did $1 million damage perpetrating it. However, because that still needs to be proven in court and that takes time and money, and there is uncertainty that the judge will make that finding, our office is willing to set aside contention over the amount of damage if we can reach a joint agreement that you are guilty of the crime and did at least $6,000 damage, and get sentenced based on that $6,000."

    That is how, I believe, it is presented and that may indeed even be the truthful opinion of the prosecutor. Maybe it isn't their actual opinion and it is a shitpiling exercise to use fear to leverage the accused into capitulation, but it is at least plausible that it is their opinion and the law is loose enough because judges willing enough to give slack that it is plausibly arguable in court. "Truth" is a fuzzy thing out here in the real world.

  35. David says

    Isn't the base level 7 due to the statutory maximum of 25 years?

    Base Offense Level:

    (1) 7, if (A) the defendant was convicted of an offense referenced to this guideline; and (B) that offense of conviction has a statutory maximum term of imprisonment of 20 years or more; or

    (2) 6, otherwise.

  36. Hasdrubal says

    Non lawyer question on how things work: Does the prosecutor need to prove the amount of damages that he is claiming? Does he have to lay out the books, and say "They paid $900,000 in consulting fees to identify any intrusions and rectify them." Can the defense then say "Woah there! Yes, they were billed $900,000, but a.) that's 3x the normal rate for this type of service in this area: These types of services generally run between $200 and $300/hour but they were billed $750/hr. And b.) about 2/3 of the services provided were audits and redesigns of features unrelated to the actual intrusion, and thus shouldn't be considered as damages."

    Second, isn't a jury supposed to determine the facts of a case? That seems to be a common thread in appellate commentary from Volokh Conspiracy posts, at least. If so, shouldn't the damages have been identified by the jury during the trial rather than just put in front of the judge afterwards? I vaguely remember either a post here or over at the Volokh Conspiracy about a judge adjusting a sentence based upon facts not presented to the jury, but I can't remember if the end result was a fair or foul call on that.

  37. FJ says

    @MoratNZ and Dwight:
    Dwight's summary is, indeed, a pretty good picture of how plea negotiations go. But as a general matter, the rules of evidence forbid parties from introducing anything anyone says in settlement negotiations. Imagine if that rule didn't exist. A defendant who says, "I'm willing to plead guilty, admitting that I committed this crime, if…" and suddenly the prosecutor stands up and says, "Thanks for the confession, see you at trial."

    @Hasdrubal: This is a fantastic point, and it is a very relevant area of recent SCOTUS caselaw. Yes, the judge may only consider "facts" that have actually been proven, but when those facts had to be proved and to whom (the judge or the jury) depends. In particular, Alleyne v. United States (2013) says that any fact that affects the statutorily permissible sentencing range (either the statutory maximum or statutory minimum) must be found by the jury beyond a reasonable doubt. An earlier case, US v. Booker (2005), said that any factors that affect the permissible sentence under the guidelines must be found by the jury beyond a reasonable doubt. The federal sentencing guidelines are now merely advisory: judges are permitted to depart from them, although (as Ken points out) abuses of that discretion are reviewable on appeal. So while *technically* this defendant faces up to 25 years of imprisonment, *as a practical matter* such a sentence is very unlikely to be imposed and would very likely be overturned on appeal if it were. SCOTUS hasn't said that the Constitution requires sentencing to be a complete crapshoot, although I suppose there's always next term.


  1. […] 3rd Section: “Just life in prison.” Thank you for joining journalists who promote ignorance by focusing on the statutory maximum sentence. Really? Does anyone who has practiced law for a day think that Bergdahl is going to receive the statutory maximum? Do you realize that his possible punishment in a General Court-Martial ranges from “no punishment” to the statutory maximum? That’s a big range, capable of accounting for the total weight of any aggravation, extenuation, and mitigation. For more information about poor journalism in the reporting of maximum sentences in order to achieve sensationalism, refer to Popehat…. […]