SENTENCING MEMORANDUM
The defendant, Reyson Jose Pena (“Pena”), pled guilty to a one-count indictment charging him with Unlawful Re-entry of a Deported Alien, in violation of 8 U.S.C. § 1326. The defendant urged a sentence at the low end of the applicable range of imprisonment under the United States Sentencing Guidelines. The government filed a motion for an upward departure, on the grounds that Pena’s criminal history understated his culpability and the likelihood that he would re-offend (under U.S.S.G. § 4A.1.3). I granted the government’s motion. No other issue is contested in this sentencing.
I write this memorandum because it is my practice to write opinions whenever there is an unusual issue in a sentencing, or whenever I depart from the Guidelines, either upward or downward. I do so for the parties, the Court of Appeals, and more significantly, for the public.1
*67The government here sought an upward departure because, it argued, Pena’s “criminal history” score under the federal Sentencing Guidelines understated his criminal record. Specifically, the government noted that Pena had been convicted of a particularly heinous crime — mayhem and assault and battery — prior to his deportation. The government also cited one conviction for drug distribution and an arrest for drug trafficking that has not yet been adjudicated, both after Pena’s unlawful re-entry into the United States.
As I describe below, I concluded that although the facts underlying Pena’s conviction for mayhem and assault and battery were, indeed, heinous, it was not appropriate for me to punish Pena again— through a “criminal history” adjustment on an immigration charge — for an eight-year old conviction, which had been fully adjudicated in the state court and for which another judge had imposed an eight- to nine-year sentence in the state’s maximum security prison. Pena tried to explain what had happened eight years before; the government countered with reference to police reports and newspaper clippings. I was in no position to retry the case or to evaluate whether the state court punishment was appropriate. Nor would I consider a drug arrest, the facts of which had not been adjudicated and were not before me.
However, I did consider Pena’s earlier drug conviction because the state court judge hearing that case obviously had been misled. Pena had lied to the court about his criminal record in that case; he did not disclose his true identity. When the state court judge sentenced him, the court gave him the kind of sentence it would give a first offender, a suspended sentence rather than a term of imprisonment. As a result, the defendant received only one criminal history point for that state conviction under the federal Sentencing Guidelines calculation in this case. If the state court had given Pena a term of imprisonment, it would have affected his criminal history score under the Guidelines in this case by at least one level. Therefore, I departed upward one level.
A. Framework for Considering Criminal Record
The United States Sentencing Commission, pursuant to its authority under the Sentencing Reform Act, determined that a defendant’s “past criminal conduct”2 is directly relevant to the purposes of sentencing. First, an individual “with a record of criminal behavior” is more culpable, it found, than a first offender. Second, general deterrence requires that a message be sent that “repeated criminal behavior” will aggravate the punishment. Finally, such conduct is correlated with recidivism on the one hand, and the limited likelihood of rehabilitation on the other. U.S.S.G. § 4A1.1 attempts to identify factors that correlate with these purposes.
The starting point for the analysis consists of the defendant’s record of convictions. The Guidelines “score” a conviction depending upon the length of the sentence, and, to a degree, the circumstances of those convictions. Section 4A1.1 of the Guidelines outlines how those points are assigned: More points for sentences exceeding one year and one month, less for lesser sentences, and more points if the defendant committed the offense while under any criminal justice sentence, or less than two years after release from imprisonment.
*68At the same time, the Guidelines recognize that the formal scoring system cannot be entirely adequate to the task of characterizing criminal history. See e.g., United States v. Leviner, 31 F.Supp.2d 23, 31-34 (D.Mass.1998) (Gertner, J.). If “reliable information indicates that the criminal history category does not adequately reflect the seriousness of the defendant’s past criminal conduct or the likelihood that the defendant will commit other crimes, the court may consider imposing a sentence departing from the otherwise applicable guideline range.” U.S.S.G. § 4A1.3, p.s.
The Guidelines’ policy statement describes the kinds of circumstances in which a court would conclude that the criminal history score is not adequate to reflect the seriousness of a defendant’s past criminal conduct or the likelihood that he will be a recidivist. In effect, the drafters urge the court to ask the question — what is it about the score that created the under (or over) valuation of the record?: (i) prior sentences not counted in computing criminal history (e.g., foreign sentences), (ii) prior sentences of “substantially more than one year” that derive from independent crimes on different occasions, again not counted in the Guidelines score, see U.S.S.G. § 4A1.3(a) and (b); (iii) the timing of convictions- — did they follow closely one after another, or while another charge was pending, see U.S.S.G. § 4A1.3 (d); and (iv) the nature of the convictions — e.g., assaul-tive behavior or non-violent conduct. See Leviner, 31 F.Supp.2d at 32-33.
Significantly, in some instances, the Guidelines also permit the court to go beyond the formalities of the conviction and sentence, as the Government here urges, to reexamine sentences that the defendant already has received. Commentary to U.S.S.G. § 4A1.3 suggests that a court may second-guess the sentences imposed by other courts, i.e., “a defendant with an extensive record of serious, assaultive conduct who had received what might now be considered extremely lenient treatment in the past might have the same criminal history category as a defendant who had a record of less serious conduct.” This, as the drafters note, may be especially true of younger defendants who “may have received repeated lenient treatment, yet who may actually pose a greater risk of serious recidivism than older defendants.” U.S.S.G. § 4A1.3, comment, (backg’d).
Indeed, the language of these sections suggests that the court is permitted to go still further beyond the formalities of convictions and sentences altogether. Section § 4A1.3(e) permits the court to consider “prior similar adult criminal conduct not resulting in a criminal conviction,” so long as the information is “reliable.” Id.
Nevertheless as I noted in United States v. Footman, 66 F.Supp.2d 83 (D.Mass.1999) (Gertner, J.), “just because the Guidelines seem to give a judge the discretion to look to criminal conduct rather than convictions, it does not mean that a judge should do so. Once one ventures past formal convictions, and formal sentences, one is in troubling territory. How can the court reconstruct prosecutions long past? Should the court essentially retry the past conduct, or rely on police reports which may well be skewed? And if a matter has been dismissed, or worse, resulted in an acquittal, how can the court, many years hence, substitute its judgment for that of the original decision-maker?” Id. at 100.
In Footman, I departed upward three levels because of admissions that the defendant had made about his prior criminal conduct, admissions overheard in taped conversations introduced in the case which I had tried. I also heard the testimony of a witness who described what the defendant had done to her, and why she had not *69pressed charges. See id. at 100. I do not have such information here.
B. The 1994 Conviction for Mayhem and Assault and Battery
The 1994 conviction which the government pointed to was for mayhem, assault and battery with a dangerous weapon, and assault and battery. It derived from the egregious treatment of a child, the son of defendant’s girlfriend. I was horrified at the account.
Defendant’s counsel reported that his girlfriend received a more substantial punishment, which somehow reflected the state court’s conclusion she was the more culpable party. The defendant described his extraordinary drug habit at the time, implying that he had no idea what was happening.
After reading the police report, as reproduced in the presentence report, and the press clippings, it would be easy to say that the sentence of eight to nine years’ imprisonment at MCI Cedar Junction that the defendant received at that time was far too lenient, and that I should make up for that leniency by “throwing the book” at him in this case. It would be easy to reject the defendant’s after-the-fact characterization of those events. And if I did, no one would criticize the outcome.
But I cannot sentence someone based on press clippings and police reports alone. Another judge, presumably hearing evidence and not hearsay accounts, concluded that eight to nine years was sufficient. And that substantial sentence was counted in Pena’s criminal history score. It led to three more criminal history points. Can I say from this vantage point that the state court judge should have been more punitive? I cannot.
Can I say that the mayhem conviction predicts that Pena is more likely to reenter the country — the only charge before me? I cannot. One charge has nothing to do with the other.
Or can I say that the mayhem conviction is likely to predict he will commit similar crimes? I cannot. Apart from that one heinous crime, Pena’s record is about drugs and motor vehicle violations. There are no other convictions for violent behavior, no weapons offenses, and no other suggestions of child abuse or domestic violence.
C. The 2000 Arrest for Traffickiny in Cocaine
Section 4A1.3(e) of the Guidelines permits the court to consider “prior similar adult criminal conduct not resulting in a criminal conviction,” so long as the information is “reliable.” U.S.S.G. § 4A1.3, p.s. This section mirrors the language of the “Introductory Commentary” to Part A— Criminal History. It addresses the relationship between “past criminal conduct” or “prior criminal behavior,” on the one hand, and culpability, deterrence, and recidivism, on the other. At the same time, the policy statement notes that “a prior arrest record itself shall not be considered under § 4A1.3.” Id. Putting together the references to “past criminal conduct” or “behavior,” with the reference to a “prior arrest record itself,” suggests, as I noted in Footman, supra at 99 n. 27, that a court cannot rely only on a prior arrest without any additional reliable information. Plainly, there are serious concerns about fairness when a judge extrapolates merely from an un-adjudicated, and unchallenged arrest record. In any event, the commentary focuses on “similar” conduct; the reference suggests that the crime of prior arrest (here, drug trafficking) must be similar to the charge at bar (illegal reentry into the United States), and it is not.
*70D. The 2000 Conviction for Distribution of Heroin
Defendant’s 2000 conviction for distribution of heroin is a different matter. Pena lied to the state court. He identified himself as Jose M. Diaz-Valentin and later Jose Febus. He used a fake date of birth, and a fake social security number. He received a suspended sentence, with the proviso that he get random drug screens and remain employed. It is clear that Pena’s sentence would have been higher had he told the state court who he was, and that he had been convicted of an earlier drug possession charge and of mayhem in the 1994 case.
In this instance, it makes no sense to count this conviction as simply adding one point to Pena’s criminal history, rather than two or three, which would be the case had the state court been fully informed. Assuming three points, this raises Pena’s criminal history score to 8, and puts him in Category IV.
Accordingly, I sentenced Pena on that basis, and departed upward (albeit not as far upward as the government had urged). Rather than the 78 months the Government sought, or the 46 months the defendant sought, I sentenced Pena to 67 months in prison, which is likely to be followed by his immediate deportation.
SO ORDERED.