trict court about the length of her sentence. She did not, however, want to withdraw her guilty plea to the conspiracy and cocaine possession charges. After considering this complaint, the district court reviewed the transcript of the change of plea hearing and decided that Ms. Barnes was bound by all of the terms of her plea agreement. Therefore, it imposed the concurrent eight-year sentences.
These sentencing decisions by the district court led to the Barneses’ appeals.
II.
Michael Barnes finds two reasons why the district court should have allowed him to withdraw his guilty plea. First, he argues that his plea was invalid because neither he, nor the prosecutor nor the court knew at the time of the plea colloquy that he would be sentenced as a career offender. Barnes believes that the parties’ ignorance of this fact created a mutual mistake that voided his plea agreement. Second, he contends that his plea is invalid because his own counsel provided him ineffective assistance by offering him an inaccurate prediction of his sentence before the change of plea hearing.
A.
The first question is whether a mutual mistake undermined Mr. Barnes’ plea agreement. Interestingly, Mr. Barnes derives his argument on this point from the principles of contract law. He characterizes the sentencing consequences of his plea bargain as essential terms of his agreement with the government. If they are essential terms, the government, the defendant and the court must have a reasonably accurate knowledge of them before going forward with a guilty plea governed by Federal Rule of Criminal Procedure 11.
Plea agreements are governed by ordinary contract principles. See United States v. Ingram, 979 F.2d 1179, 1184 (7th Cir.1992), cert. denied, 507 U.S. 997, 113 S.Ct. 1616, 123 L.Ed.2d 176 (1993). At least in theory, ambiguity in an essential term or a mutual mistake about the meaning of such a term can invalidate it. See id.; United States v. Atkinson, 979 F.2d 1219, 1222-23 (7th Cir.1992). When the government proposes a plea agreement, when the defendant accepts it and when the district court enforces it, there must be a meeting of minds on all of its essential terms. See Fed. R.Crim.P. 11; see also Santobello v. New York, 404 U.S. 257, 260-63, 92 S.Ct. 495, 497-99, 80 L.Ed.2d 427 (1971). The nature of the charge to which the defendant pleads, the factual basis for the plea and the limits of the district court’s sentencing authority provide the essential parameters of all guilty pleas. See Santobello, 404 U.S. at 260-63, 92 S.Ct. at 497-99; see also North Carolina v. Alford, 400 U.S. 25, 34-39, 91 S.Ct. 160, 165-68, 27 L.Ed.2d 162 (1970). These factors determine the nature and severity of a defendant’s culpability, which, in turn, is the subject of the agreement between the government and the defendant.
The precise nature of the punishment that will be imposed after a plea is effected is not always an essential term of the plea agreement.' Although punishment is related to culpability, punishment itself is not a subject of every plea agreement. In some situations, of course, it can be. When the government and a defendant enter into a plea agreement under Federal Rule of Criminal Procedure 11(e)(1)(C), the punishment is specified as an essential term of the agreement because the guilty plea is exchanged for the imposition of a specific sentence, and the district court that enforces the agreement is bound by this term. See United States v. Bennett, 990 F.2d 998, 1002 (7th Cir.1993).
On the other hand, when the government and a defendant agree on a guilty plea regulated by Rule 11(e)(1)(B), the punishment is not an essential term of the plea agreement. The agreement is complete when the parties agree on the nature and extent of the defendant’s culpability. They leave the determination of a sentence to the discretion of the district court, as guided by the sentencing guidelines and by the applicable criminal statutes. See Bennett, 990 F.2d at 1001. A proper plea colloquy informs the defendant of the contingent nature of the