Defendant appeals from a judgment of conviction on four counts of criminal mistreatment in the first degree, ORS 163.205. Defendant contends that the trial court improperly admitted expert testimony, asserting that that testimony impermissibly commented on the credibility of the juvenile complainant. Defendant also challenges the sentences imposed by the trial court. We conclude that the testimony was properly admitted but conclude that the sentences imposed by the trial court exceed the statutory maximum. Consequently, we affirm defendant’s conviction but vacate the sentences and remand for resentencing on all four counts.
The material facts are as follows:1 During the summer of 1997, the complainant, B, and his mother, Kimberly, lived with defendant at defendant’s mother’s apartment. Defendant and Kimberly had a three-year-old son, S, who also lived with them. Defendant, defendant’s mother, and Kimberly used heroin regularly, and testimony at trial indicated that defendant was often violent and controlling toward his mother and Kimberly.
On September 6, 1997, Kimberly took B, who was nine years old at that time, to the Emanuel Hospital Emergency Room for an examination of bruises Kimberly noticed on B. In addition to asserting various acts of physical abuse, B told the emergency room physician, Dr. Dean, that defendant had touched B’s penis, forced B to touch his penis, put his penis in B’s mouth, and sexually assaulted B with a broomstick. B told Dean that defendant threatened B with a knife and threatened to kill B if he did not do as defendant told him.
On September 8, 1997, B was interviewed by Katherine Kroeger, a licensed clinical social worker with the Child Abuse Response and Evaluation Service (CARES) at Emanuel Hospital. Diming that interview, B recounted many of the same acts he had described to Dean but added that defendant forced him to suck on his two-year old brother S’s *549penis. B also alleged for the first time that defendant had forced B to lie in a comer of a bedroom for three days without food or drink.
When B made his original allegation, defendant was in jail on a probation violation. A few months later, and before charges were filed in this case, defendant was released from jail. Defendant and Kimberly reunited and decided to take the children to Arkansas. Defendant was subsequently charged in this case with two counts each of sodomy in the first degree and unlawful sexual penetration in the first degree, four counts of sexual abuse in the first degree, and four counts of criminal mistreatment in the first degree.2
In August 1998, shortly before trial, B recanted some of the details from his initial disclosures. Specifically, B stated that the broom handle had not penetrated his anus but, instead went between the “cheeks” of his bottom and that defendant neither touched his penis nor forced him to touch defendant’s penis. Therefore, the district attorney notified defendant’s attorney that B had partially recanted his report and that the prosecution might seek to introduce expert testimony on the “significance of a recantation by a child victim.” In response, defendant filed a motion in limine that sought, inter alia, to exclude (1) “any opinions regarding the truthfulness of any witness,” and (2) expert opinion as to “why child/victims of abuse might recant.” The court deferred ruling on the first request pending development of evidence at trial. The court denied the second request, concluding that expert testimony regarding recantation was not “novel scientific evidence” and therefore did not require an OEC 401 hearing under State v. Brown, 297 Or 404, 687 P2d 751 (1984), and, alternatively, that “if Brown applies, I think this evidence qualifies.”
Immediately before trial, B told the prosecutor that his original report to Dean and CARES was the truth. However, at trial, B once again changed his account. In particular, although his trial testimony was otherwise consistent with *550his original report to Dean and CARES, B testified that, contrary to his original accusations, defendant had neither touched B’s penis nor forced B to touch defendant’s penis. B testified that he had made up those aspects of his initial disclosures, and explained that this was because he was “scared” and “embarrassed.”
After B testified, the state called Dean. Dean testified extensively, and without renewed objection, to the phenomenon of recantation among child sexual abuse victims.3
The state then called Kroeger, the clinical social worker who had interviewed B and Kimberly at CARES. Kroeger first testified as to the history of instability, frequent moves, violence, and drug use among the adults in B’s life. The prosecutor then embarked on a series of hypothetical questions regarding the effect of that disruptive home life on a child’s willingness to disclose abuse by a parent. Defendant objected, unsuccessfully, to this testimony as constituting a comment on B’s credibility.
Because those aspects of Kroeger’s testimony are the focus of this appeal, we reproduce them in their entirety:
“Q [By Prosecutor]: What about a situation where — as we have here, where B[ ] tells on the defendant, he goes to the CARES program, he makes a videotape about what has happened to him, and then his mom tells him, ‘I still love the guy, I want to get back with him,’ basically, ‘tell me it really didn’t happen,’ and he agrees with that, and then the defendant shows up and takes B[ ] off to Arkansas, with the mom, with S[ ], and off they go. How does that impact B[ ]’s sense of security today—
ííífc i¡< }]< ^ ‡
“[Defense Counsel]: Object. He’s asking her to comment on the credibility of this particular witness.
“[Prosecutor]: Not so. I’m simply—
“The Court: Overruled.
*551«‡ :*< 'Jfi
“[Prosecutor]: Okay. Very brief. B[ ] tells at CARES, he goes to CARES, and he tells about what happened. And then he winds up back in the defendant’s power, if you will, with this mother, who wants to be there, and off they go to Arkansas. Then B[ ] has to walk back into a courtroom and talk about stuff that the defendant did. Does that kind of precedent tell you as a professional anything about where this little boy’s mind’s going’s to be as he’s sitting up there on that witness stand?
“[Defense Counsel]: Objection, Your Honor. He’s asking her to comment on the credibility—
“[Prosecutor]: No, I’m talking about—
“The Court: Overruled. * * * Go ahead.
“Kroeger: * * * There are — when a child talks about something, takes it back, talks about it, takes it back, kind of this process this is called recanting or retracting a statement that you’ve made. And when children aren’t supported by their families in the initial telling about it, they’re more likely to take it back. And you know, there are studies that, you know, talk about that. So—
“[Prosecutor]: So recanting is not an uncommon event?
“Kroeger: No.
“[Prosecutor]: And if the offender is in the room with the child, does that — how does that affect the calculus of the child feeling safe to tell?
“Kroeger: If a child were frightened of a person in the room with them, that threat would be there, the same way that, you know, for whatever — well, its hard for me to talk not talking about this child.
“[Prosecutor]: No, please do. You may assume that—
“Kroeger: When I saw this child, which the jury will hear on the tape, he talked to me about being fearful of having contact. So if he felt that way then, when his stepfather wasn’t in the room, I would imagine he would likely feel that way being in the room with his stepfather.
“[Prosecutor]: And does the fact that he’s on several occasions ended up back with that same person enhance that, if you will?
*552“Kroeger: Just increases that sense of, ‘Here we go again,’ or not being able to be protected.
«*#***
“[Prosecutor]: I want to talk about the notion of a consistent report and what the child either adding things or taking them out — in other words, saying sometimes that a certain type or series of events happened and other times deleting some of those and saying no, they didn’t. Is that something that — and you may assume that we’ve got that here, that B[ ] has at times said that something’s happened and other times said that they didn’t. He’s been consistent on some points and not on others. Tell us, is that something you’ve ever experienced before? Do you deal with it commonly? How do you evaluate it?
“[Defense Counsel]: Your Honor, I would object. It’s commenting on the credibility of this particular witness in this case and what this child said.
“[Prosecutor]: This is exactly—
“The Court: Overruled.
“[Prosecutor]: Thank you. You may answer the question.
“Kroeger: Since we usually only see children one time, I don’t see the aftereffects. But sometimes I hear back or, you know, we learn about studies that are done or that kind of thing, and in my field now it’s kind of seen that the process of disclosing sexual abuse can be like a process, it’s not like a one-time thing, but it can be a process where bits come out, maybe come back, come out, come back. And it’s very common. And I think — I mean, I could refer to my notes, but I think when we’re talking about recantation, I think it’s about 25 percent or somewhere around there where that kind of thing happens with kids. It happens more often where there’s threats to the child’s safety.
“[Prosecutor]: And is it always the acts themselves that can be recanted, or can other aspects of the report — for example, threats, the existence of threats, can those themselves be the subject of recantation, in you experience?
“Kroeger: Well, I mean, I — like I said, I don’t see children more than once, so I guess I would just have to say I’m not probably an expert on saying exactly what they recant or don’t recant with the context of their disclosure. But just *553like when children are interviewed, sometimes I see children, they may have been interviewed by a doctor, as in this case, or by a police officer, and I may get different details. They may not be inconsistent, but they’re not identical, and may have told the doctor something about a threat, didn’t tell me, or vice versa, but it doesn’t mean that they aren’t accurate.” (Emphasis added.)
After Kroeger testified, but before the state rested, the state recalled B and elicited further testimony regarding his earlier allegations. B now offered a different explanation for his earlier inconsistent reports. He testified that he had fabricated the fondling allegation to ensure that defendant would remain in jail:
“Because I was afraid that [defendant] would come out soon and get me because he beat my mom up before, and he came out of jail, and he then beat her up again and then went to jail. And I was afraid that [defendant] wouldn’t stay in there for that long, and when he would come out he would hurt me really bad. I was afraid that he was going to hurt me.”
While admitting that the fondling allegations were false, B reiterated that the other allegations of sexual abuse (i.e., the allegations of sodomy and sexual penetration) and all the allegations of physical abuse were true.
Despite B’s insistence that the remaining sex abuse-based allegations were true, the state withdrew all of the sex abuse-related counts4 immediately after B testified. Thus, only the physical abuse charges remained before the jury.
Defendant then presented his case-in-chief. During the defense case, the second, and much briefer, interchange underlying defendant’s “comment on credibility,” challenge occurred. During cross-examination of Bunny Stiles, an SCF caseworker, the prosecutor elicited the following testimony:
“[Prosecutor]: In your experience, what’s the effect on a child who tells on an adult about something very serious, and then the child ends up being picked up by that adult and taken out of state?
*554* * * *
“Stiles: Overall I think it’s very bad, you know. This child has — his mother has given him the, either directly or indirectly, that ‘I don’t believe you,’ you know. And that’s a real tough message for a kid who’s been trying to tell the truth, been trying to tell his parent—
“[Defense Counsel]: Objection, Your Honor.
“Stiles: —his protector something.” (Emphasis added.)
Defense counsel subsequently explained that her objection was that the references to “a kid who’s been trying to tell the truth” was a comment on B’s credibility. The court overruled that objection.
In closing argument, defense counsel argued that B was not credible and had lied and continued to lie about the abuse. The prosecutor explained B’s inconsistent reports as generated out of a fear that defendant would hurt him or his mother again:
“The defense got up here and argued to you B[ ]’s just a liar, he just lied about everything, don’t believe anything he says, just ignore it all. Well, I submit that you should simply go through what he said and look at it. And that ultimately nothing that he talked about is a lie, because ultimately we got down to the fact that the one thing he told you that wasn’t true is that the reason he had claimed the touching didn’t — the touching did happen was because he was confused or scared or embarrassed. When, in fact— those are also true, but he wanted to be sure the best he could that the defendant was going to stay away from him and his mom so they wouldn’t get hurt anymore. And that to do that he added. That’s just not acceptable. And the defendant got a huge benefit when you learned about that and when the State learned about it.
‘Now B[ ] told you everything else that happened, and when he came in, fortunately we found out about that. And he came in, and you heard about that, too. He did not tell you on the witness stand that the touching really happened. He said it didn’t. His explanation of why he earlier mentioned it was not 100 percent complete. And that’s enough properly to get rid of those charges. And the defendant ben-efitted from that. But you can still look at what he did and why that child would be so scared.”
*555The jury then convicted defendant of all remaining charges, i.e., the four physical abuse-based charges.
On appeal, defendant assigns error to the admission of the quoted excerpts of Kroeger’s and Stiles’s testimony.5 Invoking State v. Milbradt, 305 Or 621, 756 P2d 620 (1988), and State v. Keller, 315 Or 273, 844 P2d 195 (1993), defendant argues principally6 that the testimony impermissibly commented on B’s credibility, particularly to the extent that the witnesses related their general opinions to B’s individual circumstances. The state responds that, under State v. Middleton, 294 Or 427, 438, 657 P2d 1215 (1983), expert testimony that explains the reactions of typical abuse victims and relates the behavior of a particular victim to the typical pattern is admissible so long as it does not contain a direct and explicit, as opposed to an indirect or inferential, comment on a witness’s credibility. The state further argues that neither Kroeger nor Stiles offered a direct opinion as to B’s credibility, but, instead, merely “provided general testimony about the reasons that child victims of sexual abuse make inconsistent statements.”
There is much truth to each party’s position: This case has elements of both Middleton and Milbradt (and Keller). Thus, our task is to derive and apply a principle common to those cases.
Middleton involved a classic “recantation” scenario. There, the complaining witness, a 14-year-old girl, initially reported that her father had raped her. Two months later she claimed she had lied about the rape so she could move out of her parents’ house. At trial, however, she testified that the rape did occur as originally reported. The trial court admitted testimony from two social workers who had interviewed the *556child and who testified as to “the behavior of the type of children who have reported a claim of rape by a family member and whether the [complaining witness’s] behavior was consistent with what [was] described as typical behavior.” Middleton, 294 Or at 433. That testimony included statements from the expert that the expert “found [the victim’s] behavior very much in keeping with children who have complained of sex molestation at home,” id. at 432 n 5, and that the victim’s behavior of retracting a report was “a very common kind of thing to happen * * * very typical for a teenage sex abuse victim.” Id. at 433-34 n 7.7
On appeal, following his conviction, the defendant argued that the expert’s testimony was improperly admitted because it “was a direct effort to support the credibility of the complaining witness.” Middleton, 294 Or at 435. The Supreme Court disagreed:
“Neither of the experts directly expressed an opinion on the truth of the victim’s testimony. Much expert testimony will *557tend to show that another witness either is or is not telling the truth * * *.
“* * * It would be useful to the jury to know that not just this victim but many child victims are ambivalent about the forcefulness with which they want to pursue the complaint, and it is not uncommon for them to deny the act ever happened. Explaining this superficially bizarre behavior by identifying its emotional antecedents could help the jury better assess the witness’s credibility.
“* * * Defendant does not deny that this form of behavior may exist following familial sexual abuse, or that the experts described it correctly. It is information that the jury did not have. * * * Because the jurors said they had no experience with victims of child abuse, we assume they would not have been exposed to the contention that it is common for children to report familial sexual abuse and then retract the story. Such evidence might well help a jury make a more informed decision in evaluating the credibility of a testifying child.” Id. at 435-37 (footnotes omitted; emphasis added).
The Middleton court expressly noted that it did “not go so far” as to allow counsel to ask an expert whether the expert “believed the child” complainant. Id. at 437 n 11. In doing so, it confined the permissible scope of this type of expert witness testimony:
“Perhaps the jury itself would have been capable of deciding whether the daughter’s behavior actually fit the pattern described by the experts. However, as said in 4 Weinstein’s Evidence 702[02] (1981) there is no bright line separating issues within the comprehension of the jurors from those that are not. * * * If a qualified expert offers to give testimony on whether the reaction of one child is similar to the reaction of most victims of familial child abuse, and if believed this would assist the jury in deciding whether a rape occurred, it may be admitted.
«‡ %
“We expressly hold that in Oregon a witness, expert or otherwise, may not give an opinion on whether he believes a witness is telling the truth. We reject testimony from a witness about the credibility of another witness, although we recognize some jurisdictions accept it.
*558“We hold that if a witness is accepted as an expert by the trial court, it is not error to allow testimony describing the reaction of the typical child victim of familial sexual abuse and whether a testifying victim impeached by her prior inconsistent statement reacted in the typical manner when she made that inconsistent statement.” Middleton, 294 Or at 437-38 (footnotes omitted; citation omitted; emphasis added).
Thus, the court in Middleton approved not only general testimony about the phenomenon and dynamics of recantation but also testimony as to whether the particular complainant's conduct was consistent with that general phenomenon— i.e. whether “a testifying victim * * * reacted in the typical manner when she made that inconsistent statement.” Id. at 438. While admitting the former, “general phenomenon” testimony comports with the court’s rationale of “explaining * * * superficially bizarre behavior by identifying its emotional antecedents,” id. at 436, admission of the latter, “applying the general to the specific” testimony is not necessarily so justified. That is, it is one thing to educate the jury about an unusual phenomenon bearing on credibility, but it quite another to “connect the dots” explicitly by eliciting testimony relating that phenomenon’s specific application to a particular witness’s testimony. See id. at 437 (“Perhaps the jury itself would have been capable of deciding whether the daughter’s behavior actually fit the pattern described by the experts.”). The Middleton court appreciated that distinction and attempted to strike a middle ground: The expert could give an opinion as to whether the particular complainant/witness’s circumstances were consistent with the general dynamics of recantation but could not explicitly state whether he or she believed that the complainant/ witness was telling the truth.8 At least the last “dot” must be left “unconnected.”
One other, perhaps self-evident, aspect oí Middleton bears emphasis: Middleton addressed recantation — scad recantation, unlike some other behaviors associated with sex abuse, is innately a matter of truth and falsehood. The *559essence of the expert testimony in Middleton (and in this case) is that, given the dynamics of child abuse, the “recantation” was false — and, concomitantly, the original accusation was true. It is one thing to permit an expert, without adequate foundation, to testify that acting out sexually with others is a common manifestation of sex abuse and that the particular child’s behavior was consistent with that phenomenon — which may well indirectly buttress the complainant’s credibility. See State v. Dale, 75 Or App 453, 455, 706 P2d 1009, rev den 300 Or 451 (1985) (admitting testimony regarding typical behavior of sex abuse victims, including, inter alia, acting out sexually with others). It is quite another to testify that falsely withdrawing accusations is a general phenomenon and that the particular child’s “recantation” accords with that phenomenon. And yet, that is precisely what Middleton approved.
No subsequent Oregon Supreme Court case has gone so far in permitting testimony so explicitly and directly relating to a witness’s credibility.9 Indeed, in Milbradt, and again in Keller, the court, without qualifying or disavowing Middleton, precluded opinion testimony describing “expert” assessments of child sex abuse complainants’ demeanor.
In Milbradt, the challenged expert was a psychologist who had interviewed two severely mentally retarded young women whom the defendant had allegedly sexually abused. The psychologist testified, over objection, that neither exhibited “evidence of deception.”10 Milbradt, 305 Or at *560625. The psychologist further testified, over objection, that one of the complainant’s mental condition rendered her so unsophisticated as to be incapable of“plan[ning] a systematic or motivated deception, or [of] carrying] it through.” Id. at 627. The Supreme Court reversed defendant’s conviction, holding that the disputed evidence constituted an impermissible comment on the credibility of a witness because “an opinion that a person is not deceptive, could not fie without being tripped up, and would not betray a friend” is “tantamount to” an opinion that the expert believes a witness “is telling the truth.” Id. at 630.11
In Keller, a physician with CARES who had examined and interviewed the complaining child witness testified that she had assessed the child’s demeanor in formulating her diagnosis that the child had been sexually abused. In particular, the physician had attempted to ascertain whether the child had been “coached” and had concluded that there was no evidence of such “leading” or “coaching.” The witness later testified that, during the interview, the child was “obviously telling you about what happened to her body.” Keller, 315 Or at 279.12 The court found that testimony inadmissable:
*561“[The physician] testified on direct examination during the state’s case-in-chief that ‘[t]here was no evidence of leading or coaching or fantasizing’ during the interview at CARES and that, in that interview, the child was ‘obviously telling you about what happened to her body.’ Each of those statements amounts to testimony that the child was credible.” Id. at 285.
The court emphasized that such testimony is inadmissable regardless of whether the witness is commenting on the credibility of another witness’s trial testimony or addressing statements made by that witness on some other occasion. Id.
Both Milbradt and Keller rest on the principle that an expert’s assessment of a particular witness’s demeanor is inadmissible: “An opinion that a person is not deceptive * * * is tantamount” to saying that the person is telling the truth. Milbradt, 305 Or at 630. To be sure, an expert’s testimony that there was no “evidence, or indicators of deception,” Milbradt, 305 Or at 626, and “no evidence of leading or coaching or fantasizing,” Keller, 315 Or at 278, is not necessarily “tantamount” to saying that the subject was telling the truth — a person who does not manifest “deception” or “coaching” still may be lying. Nevertheless, the risk of the jury so construing the expert comment is sufficiently great that such testimony is impermissible.
Thus, the line between Middleton and Milbradt is fine indeed. Middleton permits experts to comment on whether a particular witness/complainant’s varying accounts are “similar to” the “common” phenomenon of false recantation by abuse victims. Conversely, Milbradt and Keller preclude expert comment on whether a witness/complainant’s account bears features of “deception” or “coaching” that are common in abuse cases. Both recantation, because it is *562innately a matter of truth and falsehood, see 173 Or App at 558, and demeanor are directly related to credibility.
The line, albeit fine, is principled.13 Both Middleton and Milbradt/Keller preclude an expert from explicitly stating that he or she believes that the witness/complainant is truthful. That is, they agree that that “ultimate” question cannot be answered. Where they differ is in their treatment of the “penultimate” question — i.e., the expression of an opinion as to whether the specific complainant’s account comports with more general phenomena or dynamics bearing on credibility. Middleton allows the expert to “connect that dot”; Milbradt sand Keller do not.
That difference, in turn, flows from the Supreme Court’s fundamentally differing assessments of the underlying subject matters. In Middleton, the court characterized recantation in sex abuse cases as involving a complex and “superficially bizarre” phenomenon outside the experience of most jurors — with expert testimony about that phenomenon concomitantly assisting the jury “to better assess the witness’s credibility.” 294 Or at 436. Conversely, in Milbradt and Keller, the court, at least implicitly, regarded indicia of “deception” or “coaching” as straight-forward matters well within jurors’ common experience. Thus, in Middleton, answering the “penultimate” question was permissible as assisting the jury; in Milbradt and Keller it was not.
In sum, the common principle underlying Middleton, Milbradt, and Keller is that expert testimony must assist, not supplant, the jury’s assessment of credibility. The jury’s function is not impinged upon when expert testimony does no more than provide jurors with useful, nonconclusive information from which inferences as to credibility may be drawn.
Here, the challenged testimony did not run afoul of that principle. We note, at the outset, that the colloquy with the experts here was disjointed. Some questions were framed generally (“What’s the effect on a child * * *”) and answered *563specifically (“This child has * * *”). See 173 Or App at 553-54 (recounting Stiles’s testimony). Other questions were framed specifically (“Does that kind of precedent tell you * * * anything about where this little boy’s mind is going to be?”) and answered generally (“When a child talks about something * * * when children aren’t supported by their families * * *”). See 173 Or App at 551 (recounting Kroeger’s testimony). While such “slippage” is regrettable and sometimes confusing, it is also inevitable in trial practice. Thus, our analysis turns not on such isolated and selective semantic slippage but on an assessment of the totality of the experts’ testimony.
We further recognize that Kroeger and Stiles both responded in general terms to questions that, while couched as hypothetical, left little room for responsive answers applicable to anyone other than the particular child in this case. Such witness-specific questions should generally be avoided as they risk eliciting answers that cross the line into direct commentary on credibility.
Nevertheless, when viewed in their entirety, the content of Kroeger’s and Stiles’s challenged testimony did not impermissibly comment on B’s credibility. Rather, the witnesses provided the jury with possible reasons for B’s behavior of giving inconsistent prior reports of the abuse.
For example, Kroeger testified that children are more likely to recant when they are not supported by their families when they initially disclose the abuse, that recanting is not an uncommon event, that the process of disclosing abuse often involves inconsistent reports, and that recanting happens “more often when there’s threats to the child’s safety.”
Similarly, Stiles testified that a parent’s disbelief or lack of support sends a “tough message” to a child. However, neither expert rendered an explicit or direct opinion as to how B’s feelings of fear of insecurity had affected his testimony or past accounts of the abuse. Nor did either witness comment on B’s demeanor or mannerisms while describing the alleged abuse.
In sum, the jury was left with the task of determining which of B’s various accounts to believe, how much *564weight to give to B’s own explanation of why he lied about certain details in the earlier reports, and, most importantly, how his partial recantation of those details of the sexual abuse affected their assessment of his testimony regarding the physical abuse and food deprivation charges.14 Ultimately the challenged testimony did no more than provide the jury with information useful in making those assessments, and the trial court did not err in admitting it.
We note, moreover, that even if any aspect of the challenged testimony could be characterized as a marginally impermissible comment on credibility, the likelihood of any prejudice to defendant was extremely remote. That is so because of three peculiar aspects of this case.
First, although the challenged testimony here pertained to recantation, this is not a case, like Middleton, where the complainant recanted without explanation. Rather, B did offer an explanation: He had fabricated the fondling allegations out of fear, to ensure that defendant would be out of his family’s life.
Second, B’s explanation of his recantation and Kroeger’s and Stiles’s opinion testimony were “two ships passing in the night.” The premise of Kroeger’s and Stiles’s testimony, like that in Middleton, was that a child abuse complainant who has made truthful accusations may, from fear or guilt, later withdraw some or all of those accusations. Thus, the dynamic of recantation that Kroeger and Stiles described was one in which the original accusation was “true,” and the recantation was “false.” That is irreconcilable with B’s explanation, in which the original accusation was, in part, “false,” and the recantation was “true.” If the jury credited B’s explanation, Kroeger’s and Stiles’s recantation testimony was immaterial.
*565Third, following B’s final testimony, the state withdrew not only the fondling charges, but also all of the sex abuse-based charges. That is, the state did not proceed with the theory that B’s recantation was false and, concomitantly, that the original sex abuse charges were true. Bluntly: The charges that were the subject of the recantation were not submitted to the jury. Consequently, even if Kroeger’s or Stiles’s disputed testimony were somehow deemed to be an improper comment on B’s credibility, any possible prejudice to defendant would be remote and indirect, as, at most, collaterally influencing the jury’s assessment of B’s credibility on the physical abuse charges.
The trial court did not err in admitting Kroeger’s and Stiles’s challenged testimony.
Defendant also assigns error to the imposition of 72-month upward durational departure sentences on three of the four counts. The court imposed sentences of 60 months on Count 9, 72 months concurrent on Counts 10 and 11, and 72 months consecutive on Count 12, with three year’s post-prison supervision on each count. Defendant did not object at sentencing.
A defendant’s sentence for each crime may not exceed the statutory maximum indeterminate sentence. Criminal mistreatment in the first degree is a Class C felony, for which the maximum indeterminate sentence is five years. ORS 161.605(3). The state concedes that the 72-month sentences imposed here exceed that statutory maximum.
That error is apparent on the face of the record, ORAP 5.45(4)(b), because it is an obvious legal error that is beyond reasonable dispute, and we “need not go outside the record or choose between competing inferences to find it.” State v. Brown, 310 Or 347, 355, 800 P2d 259 (1990). We thus exercise our discretion to review it,15 and, finding error, remand for resentencing. ORS 138.222(5); State v. Umtuch, *566144 Or App 366, 369, 927 P2d 142 (1996), rev den 324 Or 654 (1997) (finding error apparent on the face of the record where applicable statutory maximum was 240 months and court imposed sentence of 260 months).16
Conviction affirmed; sentences vacated and remanded for resentencing.