OPINION OF THE COURT
SMITH, Judge:The appellant was convicted, contrary to his pleas, of raping a female airman in violation of Article 120, UCMJ, 10 U.S.C. § 920. The approved sentence included a dishonorable discharge, confinement for 1 year and 3 months, forfeiture of $705.50 pay per month, and reduction to E-l. The appellant raises the following errors: (1) The evidence is factually insufficient to support the conviction; (2) The military judge failed to provide a limiting instruction to the court members after Special Agent (SA) Kelly Harrison of the Air Force Office of Special Investigations (AFOSI) was permitted to act as a “human lie detector”; and (3) The approved sentence should be modified to reflect that the forfeitures are to only run for one month and be expressed in a whole dollar amount. Finding no material prejudice to the substantial rights of the appellant, we affirm the findings. The sentence is approved as modified below.
Factual Sufficiency
A general court-martial consisting of officer and enlisted members convicted the appellant of raping Airman First Class DA in her dormitory room on 4 July 2001. The central factual issue at trial was whether the victim consented to sexual intercourse with the appellant. After drinking heavily at a cookout that afternoon and evening, her memory of subsequent events was clouded, but she had three distinct memories between the time she left the cookout at about 2045 hours and the time she woke up the next morning in her dormitory room: (1) She recalled falling out of the appellant’s car, although she did not recall where that happened; (2) She recalled seeing the appellant’s face right in front of her; and (3) At some point, she felt “him inside of me.” A number of witnesses testified for the government and the defense about the events leading up to the time the victim returned to her dormitory room with the help of the appellant and the victim’s female suitemate. Only the appellant and the victim were present in her room when the rape occurred.
At first, the victim thought she may have dreamt about the sexual intercourse, but later she believed it was too “real” to be a dream. On the following Monday, 9 July 2001, she met with her training supervisor and told him that she thought something had happened with the appellant on the night of the cookout. The supervisor pressed her for details, asking whether the intercourse that apparently occurred was consensual, to which the victim replied, “I don’t know.” After further discussion, the victim decided to report the matter to AFOSI. On 23 July 2001, two AFOSI agents hid in the victim’s bedroom during a pretext meeting between the victim and the appellant. She asked the appellant if they had intercourse on 4 July 2001, to which the appellant replied they did. When she asked if he recalled how drunk she was, he said, “Yes, you were really $ í í í up ^
Over the course of the next nine months, the appellant made a number of oral and written statements to different AFOSI agents. The appellant made a written statement on 24 July 2001, three written statements on 28 February 2002, and a final written statement on 2 April 2002. The oral statements were contemporaneous with the various written statements and, in some instances, added detail not captured in the appellant’s longhand versions. The appellant’s explanation changed over time in a couple of significant respects. First, while he consistently maintained that the victim solicited the sexual intercourse, his account of where that solicitation occurred varied. Second, although he maintained that the victim consented to intercourse, in separate interviews with AFOSI the appellant described the victim as a “7” on a drunkenness scale of “10” and recalled that, “[w]hen I put my penis in her she was asleep.”
There was no objection at trial to the admission of the appellant’s statements to *966AFOSI; however, on appeal, he urges us to disregard those statements as we conduct our factual sufficiency review pursuant to Article 66(c), UCMJ, 10 U.S.C. § 866(c). In a broad attack on the reliability of the statements, the appellant challenges AFOSI policy (because the interviews were not recorded, there is no way to confirm what he actually told the agents) and the agents’ methodology (inconsistencies and omissions between the agents’ interview notes, their testimony, and the appellant’s written statements). At trial, and on appeal, the appellant argues that his written statements were the product of overzealous interrogation tactics designed to secure a confession at any cost. While conceding that the evidence is legally sufficient to sustain the conviction, the appellant invites us to sever his oral and written statements from our factual analysis because, unlike court members, we are not susceptible to being “dazzled and misled” by military investigators. Without his statements, the appellant contends, the evidence is not factually sufficient to support the conviction.
The test for factual sufficiency is whether, after weighing the evidence and making allowances for not having personally observed the witnesses, this Court is convinced of the appellant’s guflt beyond a reasonable doubt. United States v. Reed, 54 M.J. 37, 41 (C.A.A.F.2000). We are convinced of the appellant’s guilt beyond a reasonable doubt. The evidence, including the appellant’s statements, establishes his guilt. The appellant’s concern about the weight to be given his statements is understandable, because the inconsistent explanations he offered regarding the victim’s consent were pivotal and damning.1
“Human Lie Detector” Testimony
SA Harrison was an important government witness. He interrogated the appellant over the course of about five hours on 28 February 2002. No one else was involved in the questioning, which produced three of the appellant’s five written statements.2 At the time of trial, SA Harrison had been in law enforcement on active duty and as a civilian agent for over 40 years. By his recollection, he had conducted several thousand subject interviews during his career.
The issue of whether a lay or expert witness has offered “human lie detector” testimony implicates Mil. R. Evid. 608 (evidence of character, conduct, and bias of witness), 701 (opinion testimony by lay witnesses), 703 (bases of opinion testimony of experts), and 704 (opinion on ultimate issue). Our superior court has succinctly identified the essential concern with such testimony: While a party may introduce opinion evidence regarding a person’s general character for truthfulness, a witness may not offer “an opinion as to whether [a] person was truthful in making a specific statement regarding a fact at issue in the case.” United States v. Kasper, 58 M.J. 314, 315 (C.A.A.F.2003). See also United States v. Petersen, 24 M.J. 283, 284 (C.M.A.1987) (“We are skeptical about whether any witness could be qualified to opine as to the credibility of another.”). Such an opinion invades the province of the trier of fact, whom we presume to be quite capable of resolving matters of credibility, guilt, and innocence. We first examine SA Harrison’s specific testimony as it developed at trial, then consider the effect of his answers.
1. Direct examination
The trial counsel’s direct examination of SA Harrison was relatively straightforward, designed to explain the interrogation session and pinpoint the main inconsistencies in the appellant’s accounts. The only part of the direct examination the appellant now claims to be “human lie detector” testimony was asked and answered without objection. SA Harrison recalled that the appellant told him the victim was asleep when he penetrated her. The appellant also told him that, after he ejaculated, she awoke and said, “Thank you.” The trial counsel then asked SA Har*967rison how he responded to that, to which he replied, “I kept my composure sir.” From that exchange, and the trial counsel’s follow-on question, the appellant contends the government initiated the “human he detector” testimony by implying the appehant was lying.
2. Cross-examination
Throughout the proceeding, the appellant’s trial defense counsel raised the issue of “false confession,” suggesting the appellant’s incriminating statements were the products of a variety of influences and pressures exerted by AFOSI. The false confession theory pervaded the defense case and formed a significant part of the cross-examination of SA Harrison. In scrutinizing the techniques and tactics employed by SA Harrison, the assistant trial defense counsel elicited the following:
Q: And usually these interrogations are only going to be conducted when the investigator, in this ease yourself, is reasonably certain of the subject’s guilt?
A: Yes.
Q: So no doubt in your mind this was interrogation?
A: No doubt.
Q: And that’s because Airman Jones was a suspect or the subject or how ever you want to call it?
A: He was a suspect.
Q: Okay. Well did you have any other subjects that you were aware of other than Airman Jones?
A: Not based on what the case facts pointed to sir.
Q: Okay. So he was the only guy there as far as a suspect goes?
A: Yes.
Q: How many of the cases that you were a part of and that were prosecuted, how many of .those were actually where there was a false confession that was elicited?
A: In my eases, none that I have ever know[n] of sir.
Q: So out of the thousands of cases, no eases that you’re aware of?
A: Well not that many that went to court-martial sir.
Q: Okay. How many would you say went to a court-martial, what number?
A: I believe in a combination of Article 32s, administrative hearings for AAFES [Army and Air Force Exchange Service] and U.S. military courts-martial and civilian trials, I have probably testified about two hundred times.
Q: So of the two hundred cases that went to trial, how many of those were where a false confession was elicited?
A: None.
Q: And that’s based on what? How do you know?
A: Sir, let me assure you, you know when a man is telling you the truth. It’s backed up by corroborated evidence and jury findings.
Q: So for instance out of the two hundred cases that you’ve been a part of you probably haven’t read the complete record of trial in these cases, have you?
ATC [Assistant Trial Counsel]: I object, Your Honor, as to relevance at this point.
ADC [Area Defense Counsel]: Well Your Honor, this certainly is a witness’ testimony as to technique and certainly the issue of false confessions. Maybe I can lay a foundation as to whether something that’s in existence based on his training and experience.
After questioning SA Harrison about the “Reid Technique”3 for interrogations, the assistant trial defense counsel continued to focus on the agent’s method of interviewing the appellant:
Q: So the first one that we talked about or that you talked about is that you let Airman Jones know that you believed that he was guilty?
*968A: No sir. I told him I didn’t believe what he was telling me.
Q: Okay. And you told him that one time, more than one time?
A: Several times sir.
Q: So let me get this straight, Airman Jones comes in, gives you a written statement, you asked for it, he gives it to you, he got a chance to review, you went back and said, I don’t believe you.?
A: Yes sir.
Q: Isn’t it true that also part of the strategy to obtain the confession was to let him know that he was guilty? You had to do that. I mean, isn’t that true?
A: Yes sir I did.
Q: And part of that is—I guess have you heard words to this effect that ‘You had to bait the hook so that you could catch the fish?”
A: Yes sir.
Q: Where have you heard that before?
A: Sir, that’s common sense.
Q: But have you heard that phrase before?
A: Yes sir.
Q: Okay. Where did you hear that before?
A: Sir, my Dad’s told me that, my Granddaddy’s told me that, I’ve said that.
Q: And you’ve actually said that to us. Remember you told us that last night? A: Oh yes. Absolutely sir.
Q: So in this case you had to bait the hook to catch the fish?
A: Yes sir. If I went in there and said, “I believe everything you’re telling me” we’d been through with that thing.
There was no re-direct examination by the trial counsel. No limiting instructions were requested by counsel for either side, nor were any given by the military judge sua sponte.
3. Questions from, the members
The members were actively engaged throughout the proceeding (the record includes 30 question sheets from the members, many with multiple questions). Three of the members submitted a total of six questions for SA Harrison. Two of the questions were objected to by both counsel and not asked, and another question dealt mainly with the nature and duration of the interviews SA Harrison conducted. The remaining three questions concern “human lie detector” evidence.
SA Harrison was asked if “baiting the hook” is a legal way to obtain a confession. He responded:
It depends on the context of what you’re referring to. If you’re referring to the fact that I told Airman Jones “I don’t believe your story”, what I’m putting him on notice is that I don’t believe your story. And as long as I engage in did you do it, no, did you do it, no, we’re not getting anywhere. But Jonesy, I don’t believe your story, now did you do this thing, he’s on notice that he has to respond to me differently. He’s either got to tell the truth, lie to me or get up and walk out. Those are his options as a criminal suspect under rights advisement. So you can put it in the context of baiting the hook, priming the pump or putting the suspect on notice, I don’t believe what you’re telling me.
SA Harrison was then asked if he had ever interrogated an innocent person. He replied:
Absolutely, and they will not make a confession. Of course we have to do that sometimes. You just have to say, “Look, did you shoot your wife? We need to know that.” And the man is absolutely innocent. But guess what, we’re going to ask the questions that if he says yes to us it’s going to be incriminating. So in a way the interrogatory is an interrogation. Sure, you don’t get anywhere with people who are telling you the truth and did not do the crime.
Finally, SA Harrison was asked if his interviews had ever led to exoneration. He testified:
Absolutely. Absolutely. You take a statement, you go over everything with the individual and sometimes they will give you that piece of information that will vali*969date their innocence. Or they will give you what we call a lead. Obviously the best thing in the world is an alibi. If hey, I wasn’t there. Well, where were you? I got a gas ticket. I was two hundred miles away at a service station. I can prove it. Well obviously a man can’t be at two places at the same time. So yes, a lot of these so called “did you do it interview” I don’t believe let’s work this story out a little bit better will absolutely lead to ‘Yeah, I believe you know [sic].” And you terminate the interview and you shake his hand and say thank you for your cooperation. We’re sorry we had to put you through this.
These exchanges occurred without objection and the trial counsel asked no additional questions. No limiting instructions were requested or given.
k. Discussion
We review the admission of evidence for an abuse of discretion. United States v. Johnson, 46 M.J. 8, 10 (C.A.A.F.1997). The question of whether the members were properly instructed is a question of law that we review de novo. Kasper, 58 M.J. at 318.
There is no litmus test for determining whether a witness has offered “human lie detector” evidence. Not surprisingly, the outcomes in reported eases have hinged on their particular facts. See generally Kasper, 58 M.J. at 314 (impermissible opinion testimony initiated by the prosecution constituted prejudicial plain error); United States v. Whitney, 55 M.J. 413 (C.A.A.F.2001) (special agent’s testimony that the appellant did not respond to a challenge to his truthfulness was error of constitutional proportion, but was harmless given the remedial action by the military judge and the strength of the evidence); United States v. Schlamer, 52 M.J. 80 (C.A.A.F.1999) (while perhaps improper, it was not plain error for trial counsel to ask a special agent if the appellant’s confession was false); United States v. Eggen, 51 M.J. 159 (C.A.A.F.1999) (admission of doctor’s opinion of victim’s truthfulness during a counseling session was not plain error, but was invited error in any event).
From the case law, we can identify several, nonexclusive factors germane to an assessment of whether “human lie detector” testimony has occurred and, if it has, its prejudicial impact. These factors include: (1) the role of the government counsel in initiating or furthering objectionable testimony (Kasper, 58 M.J. at 314); (2) the role of the defense counsel, particularly if it appears the defense initiated the testimony for strategic reasons (Schlamer, 52 M.J. at 80); (3) the defense’s failure to object or request cautionary instructions (United States v. Halford, 50 M.J. 402, 404 (C.A.A.F.1999)); (4) whether the witness has been asked for specific conclusions or their opinion about the truth or falsity of another’s statements or allegations, or about whether a crime occurred (United States v. Anderson, 51 M.J. 145 (C.A.A.F.1999); United States v. Birdsall, 47 M.J. 404 (C.A.A.F.1998); United States v. Marrie, 43 M.J. 35 (C.A.A.F.1995)); (5) whether the testimony in question is on a central or peripheral matter (Kasper, 58 M.J. at 314; United States v. Robbins, 52 M.J. 455 (C.A.A.F.2000); Birdsall, 47 M.J. at 404); (6) whether the trial was before members or by military judge alone (Robbins, 52 M.J. at 455; United States v. Raya, 45 M.J. 251 (C.A.A.F.1996)); and (7) the remedial action, if any, taken by the military judge (Eggen, 51 M.J. at 159; United States v. Stroh, 46 M.J. 643 (A.F.Ct.Crim.App.1997)).
The appellant contends that the trial counsel opened the “human lie detector door” during direct examination. We disagree. Even if we infer a disbelief or disdain on the part of SA Harrison for the appellant’s contention that the victim thanked him for having sexual intercourse with her, the exchange between the trial counsel and SA Harrison was too ambiguous to warrant a sua sponte instruction. Since the trial defense counsel did not object or request an instruction, the issue is waived absent plain error. See United States v. Powell, 49 M.J. 460, 463 (C.A.A.F.1998). We do not find plain error.
During cross-examination, the assistant trial defense counsel attempted to elicit support for the defense theory that SA Harrison obtained a false confession. Consistent with that theory, the exchanges between counsel and the witness provided details on the circumstances of the interview *970and the evolution of the appellant’s accounts to demonstrate, from the defense perspective, how the agent’s techniques led to untrustworthy admissions. In our view, SA Harrison did not express an opinion on either the truthfulness of the appellant’s different accounts or his guilt or innocence. Nevertheless, to the extent SA Harrison’s cross-examination testimony can be construed as including his personal belief in the truthfulness of the appellant’s accounts, we review for plain error, recognizing that a witness may not opine that an out-of-court declaration is true. Schlamer, 52 M.J. at 86. Even if we construe SA Harrison’s cross-examination testimony as implying an opinion about the appellant’s out-of-court declarations, the appellant’s substantial rights were not materially prejudiced by the lack of a limiting instruction to the members at this stage of the proceeding. Article 59(a), UCMJ, 10 U.S.C. § 859(a); Mil. R. Evid. 103. Further, any error was invited by the defense as part of their trial strategy and does not provide a basis for relief. Eggen, 51 M.J. at 162; Raya, 45 M.J. at 254.
The assistant trial defense counsel’s cross-examination of SA Harrison triggered the court members’ questions and opened the door for potential “human lie detector” testimony to clarify the circumstances of the interrogation. The appellant’s trial defense counsel did not object to the questions, presumably because he expected the answers to support the defense’s false confession theory. While the appellant should not benefit from an error he occasioned, we must consider whether SA Harrison’s answers to the members’ questions, combined with the nature of his testimony up to that point, should have prompted the military judge to issue an instruction, on his own motion, reminding the members of their ultimate responsibility to assess credibility and determine guilt or innocence. See Kasper, 58 M.J. at 319.
The members’ questions excerpted above are not objectionable on their face. SA Harrison was not asked, and did not directly express, his opinion on the appellant’s guilt or whether a rape had in fact occurred. But, in answering the three questions, SA Harrison, in effect, provided an executive summary of his professional—and personal— view of the ease: that he did not believe the appellant was telling him the truth, or the whole truth, initially; innocent people will not confess and “they will give you that piece of information that will validate their innocence”; and the appellant confessed.4 At this point in the proceeding, SA Harrison’s testimony included an implicit opinion as to the appellant’s truthfulness, and it could be construed as providing his opinion on guilt or innocence. Cast either way, his testimony became objectionable “human lie detector testimony.” As an opinion of the appellant’s truthfulness, it usurped the members’ exclusive function to weigh evidence and determine credibility. Birdsall, 47 M.J. at 410. As opinion testimony on the guilt or innocence of an appellant, it was unhelpful to the trier of fact, because factfinders are “perfectly” capable of assessing credibility. United States v. Wagner, 20 M.J. 758, 761 (A.F.C.M.R.1985). If “human lie detector” testimony is offered, the military judge must issue “prompt cautionary instructions” to the members. Kasper, 58 M.J. at 315. In the appellant’s case, the military judge erred by not issuing an instruction specific to SA Harrison’s testimony. The remaining issue is whether the appellant’s substantial rights were materially prejudiced by the absence of such an instruction. Article 59(a), UCMJ. We do not find material prejudice.
Significantly, the government did not open the door to “human lie detector” testimony. Once it was opened on cross-examination, there was no re-direct examination by trial counsel and no follow-on questions to the members’ questions. The trial counsel’s findings argument included fair comment with respect to SA Harrison’s testimony and *971was not objected to on “human lie detector” grounds.
The answers elicited from SA Harrison were part of the defense strategy to portray the AFOSI agents, and SA Harrison in particular, as in aggressive pursuit of a confession without regard to the truth. SA Harrison’s answers supported this strategy, given his confidence in his interrogation tactics and ability to obtain confessions.5 Even so, the line between standard interrogation tactics and “human lie detector-type” opinion was sometimes blurred, as was reflected in the following exchange between SA Harrison and the assistant trial defense counsel after one of the member’s questions:
Q: So again, why—You could have just let Airman Jones go after he provided the first statement to you?
A: I think I would have been derelict in my duty if I would have just let that young man go.
Q: Okay. And that’s because based on the information that you had, you thought that he wasn’t telling the truth. Isn’t that right?
A: I wanted to see how that interview went sir.
Q: You talked about in the testimony that you kept reminding him that you didn’t believe him.
A: Yes sir. That is just standard police procedure.
All of the appellant’s statements were properly admitted and considered by the members. The appellant’s inconsistent explanations on central issues (e.g., when and how the victim asked him to have sex with her) were obvious. Two other AFOSI agents testified about the inconsistent stories offered by the appellant, and the evidence presented on the victim’s condition portrayed a person who most likely lacked the capacity to consent at the time the offense occurred. That likelihood was substantiated with the appellant’s admission that the victim was asleep when he penetrated her, an admission he made not only to SA Harrison, but also to two other AFOSI agents only one month later.
Although the military judge did not provide a tailored cautionary instruction, he did properly instruct the members on their responsibilities. He told the members during preliminary instructions that it was “your duty to hear the evidence and determine whether the accused is guilty or not guilty,” and “that the final determination as to the weight of the evidence and the credibility of the witnesses in this case rests solely upon you.” Prior to closing for deliberations, he reminded the members that, “each of you must resolve the ultimate question of whether the accused is guilty or not guilty based upon the evidence presented here in court,” and “[t]he final determination as to weight or significance of the evidence and the credibility of the witnesses in this case rests solely upon you.” In the absence of evidence to the contrary, the members are presumed to follow the military judge’s instructions. United States v. Holt, 33 M.J. 400, 408 (C.M.A.1991) (citing United States v. Ricketts, 23 C.M.A. 487, 490, 50 C.M.R. 567, 1 M.J. 78 (C.M.A.1975)); United States v. Dewrell, 52 M.J. 601, 605 (A.F.Ct.Crim.App.1999), aff'd, 55 M.J. 131 (C.A.A.F.2001). There is insufficient evidence in this case to overcome that presumption.
Based on the circumstances of SA Harrison’s testimony, the appellant’s “false confession” theory, the totality of the evidence, and the standard instructions given, the failure to give a specific cautionary instruction did not materially prejudice the substantial rights of the appellant. See Article 59(a), UCMJ; Reed, 54 M.J. at 41.
*972 The Approved Forfeitures
The adjudged sentence included forfeiture of all pay and allowances. In his action, the convening authority approved only “forfeiture of $705.50 pay per month.” The appellant asserts two errors with the convening authority’s action: The amount forfeited must be stated in whole dollars and the duration of the forfeitures must be specified. While Rule for Courts-Martial (R.C.M.) 1003(b)(2) requires adjudged forfeitures to be expressed in whole dollar amounts (unless total forfeitures are adjudged), R.C.M. 1107(d)(1) does not require the convening authority to do likewise in approving the sentence. There was no error in the action with respect to the amount of forfeitures.
It is appellate practice that, if the duration of forfeitures is not specified in the action, their duration shall not exceed one month. United States v. Burkett, 57 M.J. 618 (C.G.Ct.Crim.App.2002); United States v. Foster, 39 M.J. 846 (A.C.M.R.1994). Therefore, we affirm a sentence of a dishonorable discharge, confinement for 1 year and 3 months, forfeiture of $705.50 pay per month for 1 month, and reduction to E-1.
The approved findings and sentence, as modified, are correct in law and fact, and no error prejudicial to the substantial rights of the appellant occurred. Article 66(e), UCMJ; Reed, 54 M.J. at 41. Accordingly, the approved findings and sentence, as modified, are
AFFIRMED.
. “Proof beyond a reasonable doubt ... does not mean that the evidence must be free of conflict.” United States v. Lips, 22 M.J. 679, 684 (A.F.C.M.R.1986).
. SA Harrison was a polygrapher and his interviews with the appellant were in conjunction with a polygraph examination of the appellant. This information was not disclosed to the court members.
. A commonly used interview and interrogation practice. See generally Fred E. Inbau, et al., Criminal Interrogation and Confessions (4th ed.2001); Kasper, 58 M.J. at 322 (Crawford, C.J., dissenting).
. Implicitly, from SA Harrison's perspective, the appellant's statement was a truthful statement. The defense made some effort at trial to distinguish between "confessions” and "admissions,” not conceding that any of the appellant's statements were confessions. SA Harrison clearly considered the appellant’s subsequent statements to be "confessions,” particularly when the appellant characterized the victim as asleep when he penetrated her.
. One study suggests that police detectives are only slightly better able to distinguish between true and false stories than a sampling of college students, and that both groups have an accuracy rate only slightly better, statistically, than making the call on a coin toss. The study further suggests detectives have a high degree of confidence in their ability to discern the truth. Saul M. Kassin, Effective Screening for Truth Telling: Is it Possible? Human Judges of Truth, Deception, and Credibility: Confident But Erroneous, 23 Cardozo L.Rev. 809 (2002) (Transcript of an edited version of a presentation given by Professor Kassin at the Benjamin N. Cardozo School of Law Symposium, The Cooperating Witness Conundrum: Is Justice Obtainable? (2000)).