In 1989 Romeo Soto pled guilty to one count of conspiring to possess with intent to distribute over one thousand kilograms of marijuana and was sentenced to a ten-year mandatory term. Soto did not pursue a direct appeal of his conviction or sentence. Instead, over a year later, he returned pro se to the district court and moved, pursuant to 28 U.S.C. Sec. 2255, that his conviction and sentence be vacated, alleging (1) that his court-appointed counsel was ineffective in not advising him of the nonviability of the government's theory of the case and misleading him about the sentence he could face, (2) that his plea was involuntary because he thought he would not be sentenced under the Sentencing Guidelines and mandatory minimum laws, (3) that his sentence was based on an erroneous calculation of relevant conduct and (4) that the district court did not make findings resolving contested issues of fact in the presentence report as required by Federal Rule of Criminal Procedure 32(c)(3)(D). The district court summarily denied his motion without an evidentiary hearing pursuant to Rule 4(b) of the Rules Governing Section 2255 Proceedings, authorizing such "[i]f it plainly appears from the face of the motion and any annexed exhibits and the prior proceedings in the case that the movant is not entitled to relief." Soto now appeals this latest order.

We can quickly reject Soto's challenges to the district court's technical compliance with Rule 32(c)(3)(D) and its calculation of relevant conduct under the Sentencing Guidelines because, except under the rarest of circumstances (which are not presented here), these are not recognized grounds for a Sec. 2255 collateral attack on a sentence. See Kelly v. United States, 29 F.3d 1107, 1111-14 (7th Cir. 1994); Basile v. United States, 999 F.2d 274 (7th Cir.1993); Scott v. United States, 997 F.2d 340 (7th Cir.1993).

The opposite is true with respect to Soto's challenges to his counsel's performance. Despite eschewing direct appeal of his conviction and sentence, Soto may collaterally attack the performance of his original counsel because the claims of substandard assistance he presents depend for their success on evidence outside the existing record which must wait for post-conviction proceedings to be adduced. See Guinan v. United States, 6 F.3d 468, 472 (7th Cir.1993). However, while a bare trial record is typically inadequate to conclusively demonstrate a right to relief on such a claim, it may at the same time indicate that no right to relief exists or will emerge upon further evidentiary development. See id. at 474 (Easterbrook, J., concurring). This is not an incongruous position; it merely restates this circuit's view that ineffective assistance of counsel claims, even ones eventually adjudged inevitably meritless, are except under narrow circumstances to be taken up collaterally. Soto brought his claims of ineffective assistance in the right forum; while we can conclude that one certainly must fail on the merits, we cannot be so confident, on this wisp of a record, as to the merits of the other.

Soto first argues that his counsel's comments at sentencing demonstrate a misunderstanding of the mandatory minimum penalty to which Soto was subject by virtue of his plea. This misunderstanding, he seems to assert, led to earlier erroneous advice and hence a tainted guilty plea. Whether counsel's comments did in fact demonstrate the misapprehension that Soto alleges we seriously doubt: counsel acknowledged his cognizance of the mandatory minimum but asked for a lower sentence anyway. Counsel's plea was "ineffective" in the sense that the court did not have the authority to grant such mercy but no constitutional injury did or could result from such a vain request for lenient treatment. Soto's direct allegation that counsel misinformed him (before his plea we will assume) about the applicable sentencing range, once stating that 97 months was a low end possibility, is more serious. Actual prejudice, however, must also be shown to successfully attack counsel's actions, see Strickland v. Washington, 466 U.S. 668, 104 S. Ct. 2052, 80 L. Ed. 2d 674 (1984). Here, Soto's awareness of the mandatory ten-year term he was facing is incontrovertible. First, the plea agreement itself spelled it out: "Defendant understands that count one carries a minimum penalty of ten (10) years in prison...." Second, Soto acknowledged at his change of plea hearing that he had read and understood the agreement prior to signing it. Third, before accepting Soto's plea, the district court, in accordance with Rule 11, insured that Soto understood the applicable minimum and maximum penalties. The district court would not have accepted the plea if it were not satisfied that "the plea [was] voluntary and not the result of force or threats or of promises apart from [the] plea agreement." Fed.R.Crim.P. 11(d). Even if Soto's counsel had stated earlier, as Soto alleges, that his sentence would fall between 97 and 121 months (which it in fact did), the properly and carefully conducted plea hearing ensured that Soto actually knew that his sentence would not be less than ten years. See Barker v. United States, 7 F.3d 629, 633-34 & n. 5 (7th Cir.1993); cf. United States v. Trussel, 961 F.2d 685, 689 (7th Cir.1992). There was no prejudice to his decision to plead guilty and hence no constitutionally ineffective assistance. (For the same reason, Soto's claim that he was unaware of the sentencing consequences of his guilty plea, even if not defaulted because of his failure to appeal (about which we know nothing from this record), is without merit.)

Soto's second claim of ineffective assistance, however, is much less penetrable on this record. Soto argues that his counsel should have perceived that the government's theory that he engaged in a single long-term conspiracy was mistaken and thus his counsel should have discouraged him from pleading guilty. That theory exposed Soto to Guideline and mandatory minimum sentencing because his drug trafficking activities spanned the effective dates of both of these schemes. Count one of the indictment to which Soto pled guilty which, incidentally, is not contained in the record on appeal but is excerpted in the plea agreement which, like the transcript of the change of plea hearing, is appended to filings that are in the record--charged Soto with participating in a drug conspiracy involving over 1000 kilograms of marijuana that lasted from early 1986 until the end of March 1989. All we know about the substance of this supposed three-year long conspiracy is what the government mentioned in its offer of proof (whose specifics have never been challenged) at Soto's plea hearing and what Soto himself conceded at sentencing. This is the entire offer of proof:

If this had gone to trial, we would, first of all, call four witnesses, Mario Stein, Kathleen Jaeger, Fausto Zaffino, and Teresa Wielocowski. The first three of those witnesses are witnesses who have been convicted of drug trafficking offenses in the Eastern District. They would testify that during 1986 and 1987 they were involved in a partnership arrangement with this defendant, Mr. Soto.

Specifically they would testify that from April of 1986 to early of 1987 Mr. Soto drove in excess of 3,000 pounds of marijuana from Brownsville, Texas, to Milwaukee, Wisconsin, where he delivered it to Mr. Stein, Ms. Jaeger, and Mr. Zaffino, and they subsequently distributed. A search was conducted at Mr. Stein's residence in March of 1987, and several hundred pounds of marijuana were seized. Drug records were also seized at that time bearing a coded reference to Mr. Soto which is consistent with the statements made by Mr. Stein, Ms. Jaeger, and Mr. Zaffino.

In addition to their testimony, the government would introduce evidence that in November of 1988 Mr. Soto met with Special Agent Randy Gibas of the State of Wisconsin Division of Criminal Investigation who was acting in an undercover capacity. At that time Mr. Soto delivered approximately 6 pounds of a mixture containing marijuana and had discussions with Agent Gibas talking about future deliveries of marijuana and also talking about his prior arrangements with the Stein drug organization.

We would also introduce testimony from various undercover agents from the Waukesha County Sheriff's Department. They would testify that on four different occasions between January and March of 1989 they met with Mr. Soto, and on various occasions Mr. Soto either delivered or caused to have delivered through an intermediary various quantities of marijuana ranging from 1 pound to 34 pounds.

During these conversations as well Mr. Soto stated the willingness and the ability to provide larger quantities of marijuana. And that is essentially what we would be prepared to prove, Your Honor.

Soto's allocution only provided details of his earlier activities with the Stein gang.

Soto claims that if his lawyer were competent, he would have recognized that Soto's involvement with the extremely large quantities of marijuana that expose him to a ten-year mandatory term (1000 kilograms of marijuana activate the minimum) and a severe Guideline sentence took place before the effective dates of the mandatory minimum provisions and the Guidelines generally, and only by casting all of Soto's drug activities as part of a single conspiracy, straddling the effective dates, could the government have subjected him to the current regime's harsh mandated penalties for his prior conduct. Moreover, Soto asserts, the Stein group was broken up well before the Guidelines and the mandatory minimums came into effect and there is nothing to suggest that his more recent, much smaller scale, marijuana sales to undercover agents were in any way connected to his old arrangement with Stein and company. Especially in light of the severe sentencing consequences in Soto's case, a reasonably competent counsel would have questioned and strongly advised his client not to assent to the tenuous, and perhaps nonexistent, single conspiracy hook by which the government managed to have 3000 kilograms of marijuana relate forward.

This, in our opinion, is not an inherently deficient theory of ineffective assistance or a theory that can be rejected out of hand on this record. Appointed counsel would not be of much use at all if they are not expected to point out and urge their clients to avoid plea arrangements that subject them to penalties much more severe than what likely is legally appropriate and would result after a properly conducted trial. And in this case, without an enlarged record, it is impossible to conclusively determine that Soto's counsel's assumed failure to so advise was objectively reasonable. All we know is that Soto had been involved with a large-scale marijuana operation, that operation appeared to have ended, and after a considerable gap in time Soto sold marijuana to undercover agents. Without any discernible connection between the earlier and later incidents, we cannot begin to definitively evaluate Soto's counsel's acquiescence to the plea agreement as structured. We must remand for an evidentiary hearing in which an adequate record on this matter can be developed.

What will emerge may very well indicate that Soto's counsel performed competently. Many things inform the oft-contested issue of single versus multiple conspiracies. See, e.g., United States v. Flood, 965 F.2d 505, 508-10 (7th Cir.1992); United States v. Robinson, 956 F.2d 1388, 1393 (7th Cir.1992); United States v. Maceo, 947 F.2d 1191, 1196-97 (5th Cir.1991); United States v. Paiz, 905 F.2d 1014, 1020 (7th Cir.1990); United States v. Napue, 834 F.2d 1311, 1333 (7th Cir.1987). And just because Soto's later activities occurred after some of his earlier cohorts were arrested does not preclude the existence of a single conspiracy. See Robinson, 956 F.2d at 2394. Nor does the fact that only government agents may have been involved in his later sales; our opinion in United States v. Mahkimetas, 991 F.2d 379 (7th Cir.1993), only indicates that a single defendant cannot "conspire" with undercover officers alone and does not stand for the novel proposition that actions in furtherance of an existing actual conspiracy cannot involve investigating agents.

Moreover, the question of competent performance does not turn simply on the direction in which most of the factors may point in this case. The inquiry must center around whether the nonexistence of a single conspiracy was obvious or likely enough from what Soto's counsel knew (or should have known) at the time of the plea hearing--even when taking into account countervailing considerations like the risk of sacrificing other favorable treatment from the prosecution, the likelihood of a similar or worse outcome in an alternative scenario, the possibly low likelihood of success on such a theory, and Soto's repose--that it was unreasonable for Soto's counsel not to have raised it either to the government in negotiating the terms of the plea or to Soto in advising him whether to accept the plea. Because these matters cannot be resolved on the current state of the record, we reverse that part of the district court's order denying an evidentiary hearing in which Soto's second claim of ineffective assistance of counsel could be explored and remand for such proceedings consistent with this opinion.