OPINION OF THE COURT
FERGUSON, Senior Judge:Convicted by a general court-martial in accordance with his pleas of multiple serious offenses, the appellant was sentenced to a dishonorable discharge, confinement at hard labor for 10 years, forfeiture of all pay and allowances, and reduction to the lowest enlisted grade. The convening authority approved the findings and sentence as adjudged.
In its review of the record, the Navy Court of Military Review determined that the appellant had been subjected to 56 days of illegal pretrial confinement1 which, to that point, had been neither countenanced nor remedied in any fashion.2 That court then reassessed the sentence as the reparation for this pretrial wrong and reduced the appellant’s adjudged sentence to confinement for 10 years by 2 months, to 9 years 10 months. At issue before us is the legality and the adequacy of the approach used by the Court of Military Review to cure the error committed at the trial level.
There would appear to be two possible procedural correctives at the trial or appellate level for illegal pretrial confinement served by an accused. The first is a reduction of what otherwise would be the sentence appropriate for the given accused and his offenses. If the trial judge recognized the illegality, then the sentence adjudged would expressly reflect appropriate judicial “credit” on what otherwise would have been the sentence inflicted upon the accused. If the pretrial error were not noted until review by the convening authority or appellate courts, the sentence then would be reassessed to render this “credit.” This approach of judicial reduction of the appropriate sentence is what was followed in this case by the intermediate court to remedy the error.
The second alternative procedural device is to adjudge and to affirm an otherwise appropriate sentence, but to judicially order an administrative “credit” thereon for the number of days served illegally in pretrial confinement. For the reasons discussed below, it is our judgment that this latter course is the only legal and fully adequate remedy for an accused who finds that the time he spent confined prior to trial was imposed upon him unlawfully.
I
LEGALITY OF THE REMEDY
In looking at the authority of a Court of Military Review to review an adjudged sentence, we have had occasion to observe:3
Congress desired intermediate appellate authorities to look again at the penalty adjudged and reduce the severity of its impact until it was deemed appropriate. Code, supra, Articles 64, 66.
It is axiomatic that exercise of appellate authority on the sentence may not increase its impact.
Under the circumstances of the instant case, we believe that the action of the court below, in purporting to reduce the sentence adjudged, operated in fact to increase the potential impact of that sentence; as such, that court’s action on the sentence was unlawful.
The sentence initially imposed included confinement at hard labor for 10 years. Under the Navy’s so-called graduated *373schedule of good time,4 he was entitled to earn 10 days per month reduction in his sentence “for good conduct and faithful observance of all rules and regulations.”5 However, when the Court of Military Review reduced his sentence to one of confinement for 9 years 10 months, administratively the appellant’s good conduct rate was changed thereby, to 8 days per month.6 Therefore, this “reduction” in sentence actually had the consequence of increasing the period of confinement to which the appellant was potentially subject. The appellant’s normal release date, computed using the formula prescribed in paragraph 1015, SECNAVINST 1640.9 (June 19, 1972), would in fact be 196 days earlier on the 10-year sentence than on the “lesser” sentence of 9 years 10 months. This phenomenon results from the use of a graduated scale of earning good conduct time credits and the administrative practice in paragraph 1015 of calculating the normal release date by subtracting the total number of good conduct days that could be earned during the entire term of confinement from the full-term release date. Put another way, when the good conduct rate per month changes from 8 days per month on a sentence of 9 years 10 months, to 10 days per month on a 10-year sentence, the accelerated earning of good conduct time at the higher rate produces an earlier release date, although the nominal sentence to confinement actually is longer. Obviously, because of the administrative regulations in effect, the anomaly is that by reducing the appellant’s sentence by 2 months, the court’s action had the practical effect of subjecting the appellant to possible incarceration for 196 days — over 6 months — longer than possible under the adjudged sentence. We have no doubt that, by causing the appellant to be exposed to the possibility of this eventuality, the court’s action actually increased the impact of the sentence adjudged. As such, that action was contrary to the dictates of the law.7
On the other hand, a judicially ordered administrative credit on the adjudged sentence is a legal method of remedying the error. A sentence of a given number of years, being “appropriate” to the offenses and the accused, is not magically rendered “inappropriate” by the illegal character of the pretrial confinement perpetrated upon the accused. There is no relationship between them. Therefore, while it is true, as the Government has urged, that the Court of Military Review may not affirm an “illegal” sentence,8 the sentence to 10 years which was adjudged is within the maximum sentence imposable, and it is, therefore, not “illegal” in any sense. The sentence is specifically authorized by law and any “illegality” is a fiction. See Lee v. United States, 400 F.2d 185 (9th Cir. 1968). By ordering an administrative credit, therefore, the “appropriate” sentence — no less viable because of a pretrial illegality — retains its integrity.
II
ADEQUACY OF THE REMEDY
Further, this scheme of credit is the only one which truly affords full credit for time actually served. The theory behind any crediting of a sentence for illegal pretrial confinement is that the illegitimate nature of that period of incarceration somehow converts it into confinement served pursuant to the sentence eventually adjudged. That being the case, 1 month served in illegal pretrial confinement ought to count the same as 1 month served after trial.
*374Yet, as it is highly unlikely that any given accused will ever serve the full sentence to confinement, simply reducing the adjudged sentence proportionately for time actually served is not a full remedy. For instance, if a given convicted accused normally would serve approximately two-thirds of his sentence to confinement, accused “A” would serve 4 months of a 6-month sentence if there had been no unlawful confinement before trial. If, however, accused “B” had been confined illegally for 2 months prior to trial, and he received the same 6-month sentence, under the concept used by the court below, only a 4-month sentence should be approved. Pursuant to that 4-month sentence, “B” would serve 2% months post-trial which, when added to the 2 months pretrial, would result in service of 4% months on the 6-month sentence, as opposed to “A’s” service of 4 months under an “identical” sentence. Under the approach we require by this decision, the appropriate 6-month sentence would be approved for “B”, and an administrative credit of 2 months confinement served would be applied thereto, with the practical result that “B” would serve 2 months pretrial and 2 months post-trial. That is, in our opinion, full and adequate credit.
The concept of our approach is not without precedent. In Lee v. United States, supra, the accused had been confined pursuant to a sentence to 10 years, which would have entitled him to earn 10 days per month good time under a schedule9 virtually identical to that found in the Navy regulation. However, he claimed that he was entitled under the Bail Reform Act10 to a credit of 68 days for confinement served while awaiting trial on the charges ultimately resulting in his sentence. In making such a claim, however, he stated that if all the court could do was to give him credit for the 68 days by reducing his sentence accordingly, he did not want his remedy for it would result in the same anomaly present in the instant case. The Circuit Court of Appeals opined that if the sentence to 10 years were an illegal one because credit had not been given therein for the pretrial confinement, then it was not open to the accused’s choice whether to accept or reject his remedy, for the court was impelled to correct an illegal sentence even if such action results in a defendant forfeiting part of the good time credits he would have earned otherwise.
However, the court ruled that the Bail Reform Act, which provides for the Attorney General to furnish sentence credit for confinement served prior to trial in connection with an offense for which the sentence was given,11 involved an administrative computation of credit, not a judicial one. Such credit is to be applied on the sentence imposed by the court and cannot be implemented by tampering with the sentence itself, the court ruled. The court indicated that under such circumstances, the adjudged sentence- — otherwise a legal one— ought be affirmed and that the appellant must be credited administratively with the period of pretrial confinement; his remedy for the failure to so credit him lay in a writ for habeas corpus.
Similarly, in Bostick v. United States, 400 F.2d 449 (5th Cir. 1968), cert. denied, 393 U.S. 1068 (1969), rehearing denied, 409 F.2d 5 (5th Cir. 1969), cert. denied, 396 U.S. 890 (1969), the defendants contended that the trial judge had erred in failing to credit them for 7 months 23 days in pretrial confinement. The court found that the defendants were entitled to this credit, but noted that they would be adversely affected if their sentences merely were decreased by *3758 months because of their resulting loss of good conduct credit time. Therefore, rather than remanding the case to the trial court for an 8-month reduction in the sentence, the court held that the Attorney General, through the Bureau of Prisons, should compute the defendants’ good conduct time at the rate applying to their adjudged sentences of 10 years, and give them credit administratively for the 8 months spent in pretrial confinement. We are in general agreement with this approach.
The decision of the United States Navy Court of Military Review is affirmed in part and reversed in part.12 The sentence as affirmed is set aside. The record of trial is returned to the Judge Advocate General of the Navy for resubmission to the Court of Military Review with directions to issue the orders necessary to effect an- administrative credit of 2 months on the appellant’s adjudged sentence to confinement.13
. Article 13, Uniform Code of Military Justice, 10 U.S.C. § 813; United States v. Nixon, 21 U.S.C.M.A. 480, 45 C.M.R. 254 (1970).
. Upon defense motion, the trial judge had ruled that a portion of appellant’s pretrial confinement was illegal, but incorrectly had held that the 56 days in question were lawful.
. United States v. Johnson, 12 U.S.C.M.A. 640, 643, 31 C.M.R. 226, 229 (1962). See United States v. Christensen, 12 U.S.C.M.A. 393, 30 C.M.R. 393 (1961); United States v. Russo, 11 U.S.C.M.A. 352, 29 C.M.R. 168 (1960).
. Paragraph 1009.1, Department of the Navy Corrections Manual, SECNAVINST 1640.9 (May 31, 1973).
. Id.
. Id. This change in good time rate is dictated by paragraph 1009.7, SECNAVINST 1640.9 (June 19, 1972), which provides:
When a change in the period of confinement . . . [such as] a decrease in the sentence by reviewing authorities . . . changes the rate of earning good conduct time, the sentence will be recomputed at the new rate.
. See United States v. Johnson, supra; Articles 64 and 66, UCMJ, 10 U.S.C. §§ 864 and 866; paragraph 88a, Manual for Courts-Martial, United States, 1969 (Rev.).
. See United States v. Christensen, supra at 397, 30 C.M.R. at 397.
. 18 U.S.C. § 4161.
. 18 U.S.C. § 3568.
. As the military has no similar provision despite absence of opportunity for bail, the convicted accused in our system is not entitled by right to credit on his sentence for pretrial confinement. See United States v. Blackwell, 19 U.S.C.M.A. 196, 199 n. 2, 41 C.M.R. 196, 199 (1970). He must, of course, be credited on his ultimate sentence in a rehearing or new trial with time served pursuant to an initial sentence set aside on appellate review, paragraph 89c (8) (a), MCM; North Carolina v. Pearce, 395 U.S. 711, 718-19, 89 S.Ct. 2072, 23 L.Ed.2d 656 (1969), and with time served illegally in pretrial confinement.
. Our resolution of this issue renders moot any decision on whether this appellant was denied equal protection of the law by having applied to him an improperly graduated scale of good conduct time. However, it appears to us that while there may exist valid incentive reasons for application of a graduated schedule, no legitimate policy facially suggests itself supporting this particular schedule, under which an accused sentenced to confinement for 9 years 11 months is liable in fact to serve longer than an accused sentenced to 10 years.
. Quite clearly, in this opinion we have drawn a meaningful legal and practical distinction between judicial reduction of a sentence and judicial ordering of an administrative credit thereon. To the extent that paragraph 1011, SECNAVINST 1640.9 (June 19, 1972), implies otherwise, it is without lawful foundation or authority. Therefore, we declare to be illegal and of no application its directive that in cases wherein a “reviewing authority approves a sentence that includes confinement and then states that ‘credit shall be given for time spent in confinement prior to the date the sentence was adjudged’ . . . it is preferable for the convening or reviewing authority to reduce the sentence in the amount of pretrial confinement served rather than to ‘credit’ pretrial confinement, per se.”