While the porportion to my overall caseload has diminished in recent years due mainly to geographical changes, a significant portion of my practice when I was in the Phoenix area was federal criminal defense (mainly illegal re-entry and drug transportation cases). I can't say that my experience is ample sample size to extrapolate data to speak to the percentage of time that federal judges deviate downward, while rare is does happen. Funny story is that on my very first case I actually convinced the judge to downwardly depart from the guideline range and word got around the CADC so I had an influx of new cases. Of course, in those future cases the judges generally stayed within the guidelines so, alas, there is further proof that I cannot perform miracles.
To provide some background on federal sentencing, a person's sentencing ranges and the recommendations to the judge thereupon are not pulled out of thin air. Compared to most state sentencing guidelines, the federal sentencing code is very very complex. I remember I had to spend hours learning the code and the matrix and, at time, it felt like I was looking at a complex math problem. Nonetheless, the federal sentencing code is designed in a way to promote consistency. In practical terms, this basically means that (and it's more complicated than this) when a person is facing federal charges you look at the actual charge and then account for any prior convictions. Generally, this then gets you on a point on the chart which dictates the particular ranges a person faces at trial. In plea negotiations, the federal prosecutor will issue offer to reduce the level of charge (or eliminate multiple charges) or will maybe not allege a prior conviction. This serves to move the defendant on the chart into a lesser range. Federal prosecutors also, in my experience, like to bargain to basically promise not to recommend something higher on the range.
When a defendant reaches a plea agreement in federal court, the defendant is then referred to the federal probation office for a pre-sentence report/interview/evaluation. Unlike most state courts, the federal probation officer's evaluation and recommendation is usually very in depth, consistent with the law, and taken very seriously by the judge. At sentencing a job will rely heavily on this report. Defense counsel and the prosecutor will usually provide their own recommendations (as defense counsel usually my main goal was trying to argue against parts of the pre-sentence report and, frankly, I was far more worried about what is in that report than what the prosecutor). As one can expect, usually the prosecutor asks for the higher end of the range, the defense attorney makes arguments for the lower end of the range or, pursuant to the Booker case, asks the judge to "downwardly depart" below the range." Probation usually falls within the middle which is why, in my experience, the judge most often sentences most closely to probation's evaluation.
The above paragraph is why I found what happened in this Stone case to be so shocking. As I mentioned in the previous post, there were multiple other ways that the DOJ could have handled this if they thought the original recommendation was too harsh. Also, I'm curious now to know what pre-sentence report's recommendation is because, frankly, that's probably going to carry more weight than the prosecutor's recommendation. Nonetheless, what occurred here isn't normal and I don't see any evidence or suggestion that this was some political stunt by rogue prosecutors. Generally speaking, obtaining a position as a federal prosecutor is not easy and is a highly sought after position. In other words, most federal prosecutors are seasoned prosecutors and not some fresh out of law school prosecutors who may not have yet developed a sense of what sort of time a case merits. I would imagine in a sensitive case such as this quality federal prosecutors with experience would have been assigned and I imagine their sentencing recommendation wasn't based on whim or political bias.
But, at this point, wtf do I know?