After testing positive for the Human Immunodeficiency Virus (HIV), Staff Sergeant Brandon M. Dacus engaged in sexual inter*236course with female partners other than his wife without informing them of his medical condition. He was charged with two specifications of attempted murder. Daeus entered pleas of not guilty to attempted murder but guilty to the lesser included offense of aggravated assault. He also entered pleas of guilty to two specifications of adultery and was convicted consistent with his pleas. The United States Army Court of Criminal Appeals affirmed in a summary disposition. United States v. Dacus, No. ARMY 20050404 (A.Ct.Crim.App. Apr. 19, 2007) (unpublished).
If an accused sets up a matter inconsistent with his plea at any time during a proceeding on the plea, the military judge must either resolve the apparent inconsistency or reject the plea. See United States v. Shaw, 64 M.J. 460, 462 (C.A.A.F.2007); Rule for Courts-Martial (R.C.M.) 910(h)(2). We granted review to consider whether Dacus made statements during the providence inquiry or introduced evidence at sentencing that are in substantial conflict with his pleas of guilty to the aggravated assault specifications. 65 M.J. 335 (C.A.A.F.2007). We hold that there was no substantial conflict and affirm.
BACKGROUND
Daeus is HIV-positive. He was counseled by medical personnel and ordered by his commander to inform his sexual partners about his HIV status and to wear a condom during sexual intercourse. Dacus engaged in sexual intercourse with a female partner, HG, on one occasion, during which he wore a condom. He also had sexual intercourse with a different female partner, CH, approximately eleven times while not wearing a condom. Dacus did not inform either of the women that he was HIV-positive.
Stipulation of Fact and Providence Inquiry
At Dacus’s trial, a stipulation of fact was entered into evidence after Dacus testified that the information contained in the stipulation was true and correct. The stipulation of fact specifically addressed how Dacus’s conduct met each of the elements of aggravated assault with a means likely to cause death or grievous bodily harm. The stipulation provided, in part:
a. The accused did bodily harm to [HG] by having sexual intercourse with her while the accused was in an HIV-positive status. This act of sexual intercourse while HIV-positive without informing [HG] constitutes an offensive touching with another.
b. The sexual intercourse was done by unlawful force. That is, the accused had the sexual intercourse without legal justification or excuse, and without the lawful consent of [HG] because the accused had not informed [HG] of his HIV-positive status.
c. The natural and probable result of exposing [HG] to the HIV virus is death or grievous bodily harm. HIV is the virus that causes AIDS, a deadly disease. By having sexual intercourse with [HG], the accused put her at risk of contracting the HIV virus. The probability of passing the infection was more than a mere fanciful, speculative, or remote possibility.
The stipulation of fact went on to acknowledge that even though HG consented to sexual intercourse, she would not have consented if she had known that Dacus was HIV-positive. The stipulation of fact provided the same reasons to explain why Dacus’s conduct constituted the elements of assault against CH.
During the providence inquiry, the military judge reviewed the elements and definitions of assault with a means likely to produce death or grievous bodily harm with Dacus. Dacus also described the offenses in his own words. He stated, in part, as follows:
HIV and AIDS is a bad thing. I know it. I am not here to dispute it and sit here and mislead you or anybody that is here right now. Actually I have been part of this since 1996; I know what I have done and I am willing to accept what I’ve done. HIV is bad because it can cause bodily harm at one point in time of your life. It can cause death and it could in fact — it will change your whole life. And, like I said, I know*237ingly and willingly — I did that, and I am willing to face what is about to happen.
The military judge then questioned Dacus on a number of matters. When he asked Dacus whether he had sexual intercourse with HG, Dacus answered, “To a certain degree.” He stated that it was only a slight penetration, that his penis did not get erect, and that he was wearing a condom. The military judge explained that under the law, sexual intercourse means any penetration, however slight, of the female sex organ by the penis and that ejaculation was not required. Dacus then admitted to having sexual intercourse with HG “[according to the letter of the law.” Dacus also admitted to having sexual intercourse with CH about eleven times and not wearing a condom.
As to both women, Dacus’s testimony at trial was generally consistent with the statements in the stipulation of fact. The military judge elicited from Dacus his understanding and agreement that the women possibly could have contracted HIV through sexual intercourse with him even if he did not ejaculate and even if he wore a condom. The military judge accepted his guilty pleas and convicted him of two specifications of assault with a means likely to produce death or grievous bodily harm.
Testimony of Captain Mark Wallace, M.D., on Sentencing
During the sentencing phase, the defense called Captain Mark Wallace, M.D. (Dr. Wallace), an expert in the field of Acquired Immune Deficiency Syndrome (AIDS) and infectious medicine. Dr. Wallace had personally examined Dacus on one occasion and had reviewed his chart. He testified that even without treatment, Dacus was one of those rare individuals whose immune system was able to shut down viral replication on its own. This resulted in Dacus having an extremely low “viral load.”1 Dr. Wallace testified that Dacus would probably live his normal life span without getting ill from HIV.
Dr. Wallace explained that the possibility of transmitting HIV from one person to another is a function of the viral load of the infected individual. He testified that it was “[Unquestionably” possible that Dacus could transmit the virus but the likelihood was “[ejxtremely low” due to his low viral load. Dr. Wallace also testified that using a condom would reduce the risk of transmitting the virus even further.2 He went on to discuss instances where individuals with undetectable viral loads had sexually transmitted HIV to other individuals. He also noted that there were a small number of cases where pregnant women whose viral loads were kept below the level of detection during pregnancy delivered HIV-infected children. Dr. Wallace’s conclusion was that “there is no question, he could have transmitted HIV [through sexual intercourse], but that it would be very, very unlikely.”
On cross-examination, Dr. Wallace was asked whether an individual infected by a person with a low viral load would also have a low viral load or whether that person’s disease could progress in a totally different way. Dr. Wallace responded that the newly infected individual’s disease could progress in a totally different way. Dr. Wallace summarized by stating that “anything could happen. They could progress slowly, or very, very rapidly.”
ANALYSIS
We review a military judge’s decision to accept or reject a guilty plea for abuse of discretion. United States v. Harrow, 65 M.J. 190, 205 (C.A.A.P.2007). “‘Once the military judge has accepted a plea as provident and has entered findings based on it, an appellate court will not reverse that finding and reject the plea unless it finds a substantial conflict between the plea and the accused’s statements or other *238evidence of record.’ ” Show, 64 M.J. at 462 (quoting United States v. Garcia, 44 M.J. 496, 498 (C.A.A.F.1996)). The “mere possibility” of such a conflict is not enough to overturn the plea on appeal. Id.
Dacus entered pleas of guilty to two specifications of “assault with a dangerous weapon or other means or force likely to produce death or grievous bodily harm” under Article 128(b)(1), Uniform Code of Military Justice (UCMJ), 10 U.S.C. § 928(b)(1) (2000). The Manual for Courts-Martial lists four elements for this offense:
(i) That the accused attempted to do, offered to do, or did bodily harm to a certain person;
(ii) That the accused did so with a certain weapon, means, or force;
(iii) That the attempt, offer, or bodily harm was done with unlawful force or violence; and
(iv) That the weapon, means, or force was used in a manner likely to produce death or grievous bodily harm.
Manual for Courts-Martial, United States pt. IV, para. 54.b.(4)(a) (2005 ed.) (MCM).
Dacus’s challenge focuses on the fourth element. He argues that certain portions of his testimony and the testimony of Dr. Wallace substantially conflict with the fourth element’s requirement that the means was used in a manner likely to produce death or grievous bodily harm. The MCM provides an explanation of the word “likely” in the context of the aggravated assault offense: ‘When the natural and probable consequence of a particular use of any means or force would be death or grievous bodily harm, it may be inferred that the means or force is ‘likely’ to produce that result.” MCM pt. IV, para. 54.c.(4)(ii).
We further discussed the fourth element in United States v. Weatherspoon:
The standard for determining whether an instrumentality is a “means likely to produce death or grievous bodily harm” is the same in all aggravated assault eases under Article 128(b)(1). The concept of likelihood has two prongs: (1) the risk of harm and (2) the magnitude of the harm. The likelihood of death or grievous bodily harm is determined by measuring both prongs, not just the statistical risk of harm. Where the magnitude of the harm is great, there may be an aggravated assault, even though the risk of harm is statistically low.
49 M.J. 209, 211 (C.A.A.F.1998) (citations omitted). In explaining the first prong, we relied upon the “risk of harm” definition developed in several HIV assault cases and stated that the “risk of harm” need only be “ ‘more than merely a fanciful, speculative or remote possibility.’” Id. (quoting United States v. Johnson, 30 M.J. 53, 57 (C.M.A. 1990)); see also United States v. Joseph, 37 M.J. 392, 396-97 (C.M.A.1993); United States v. Klauck, 47 M.J. 24, 25 (C.A.A.F. 1997).
In Joseph, this court stated that “we do not construe the word, ‘likely1 ... as involving nice calculations of statistical probability.” 37 M.J. at 396. “[T]he question is not the statistical probability of HIV invading the victim’s body, but rather the likelihood of the virus causing death or serious bodily harm if it invades the victim’s body.” Id. at 397. Relying on language from an earlier HIV assault case, we concluded in Joseph that the “probability of infection need only be ‘more than merely a fanciful, speculative, or remote possibility.’” Id. (quoting Johnson, 30 M.J. at 57). This standard was reiterated in a subsequent HIV assault case, Klauck, 47 M. J. at 25.3
Addressing the second prong, “magnitude of harm”, we stated in Weatherspoon: “The test for the second prong, set out in the Manual for Courts-Martial, is whether death or grievous bodily harm was a natural and probable consequence.” 49 M.J. at 212.
Consistent with Weatherspoon, the military judge explained the concept of “likelihood” to Dacus as follows:
[T]he likelihood of death or grievous bodily harm is determined by measuring two fac*239tors. Those two factors are: one, the risk of harm; and two, the magnitude of the harm. Evaluating risk of the harm, the risk of death or grievous bodily harm must be more than merely a fanciful, speculative, or remote possibility. In evaluating the magnitude of the harm, the consequence of death or grievous bodily harm must be at least probable and not just possible. In other words, death or grievous bodily harm would be a natural and probable consequence of your acts.
Dacus argues that he admitted facts and introduced evidence that are substantially inconsistent with both the risk and magnitude prongs. As to “risk of harm,” Dacus points to his own testimony that during sex with HG his penis was not erect, his penis barely penetrated HG’s vagina, and he wore a condom. He also points to Dr. Wallace’s testimony that due to Dacus’s low viral road, the risk of his transmitting HIV is extremely low and drops further when a condom is used. Dacus argues that taken together this evidence makes the risk of transmitting HIV to either of the women merely fanciful, speculative, and remote.
As to “magnitude of harm,” Dacus points to his own testimony that it was only “possible” that both women could contract HIV. He also points to Dr. Wallace’s testimony that there is a very small portion of the population who have an immune system that can almost completely suppress the virus on their own. Dacus contends that this testimony substantially conflicts with the requirement, as stated by the military judge, that the “consequence of death or grievous bodily harm must be at least probable and not just possible.”
The Government responds that Dacus did not admit any facts or introduce any evidence that was inconsistent with his pleas. The Government contends that Dacus admitted that there was more than a fanciful, speculative, or remote possibility that both HG and CH were at risk for acquiring HIV and that the magnitude of harm that both women faced was immense. The Government also argues that Dr. Wallace’s testimony was, in fact, consistent with Dacus’s pleas.
We turn first to the second prong and address the “magnitude of harm.” Under the facts of this case, the question we consider is whether death or grievous bodily harm is a natural and probable consequence if HIV were transmitted by sexual intercourse. See Weatherspoon, 49 M.J. at 211-12.
Dacus’s argument that either woman might be able to naturally suppress HIV replication is not supported by the record. Dr. Wallace testified that there is a “very, very small number of people” — less than one percent of the population — who have an immune system that can almost completely suppress the virus on their own. The record here contains no evidence that either HG or CH are among the less than one percent of the population who can suppress the virus without medicine. In contrast, Dr. Wallace explained that for “most people,” “if you didn’t treat them, then they would probably get sick and die in 8, or 10, or 12, or 14 years.” In fact, Dacus himself testified: “HIV is bad because it can cause bodily harm at one point in time of your life. It can cause death and it could in fact — it will change your whole life.”
Dr. Wallace also testified that the manner in which the disease progresses in an individual is not dependent on the viral load of the person who transmitted the disease: “So, if a person with a [low] viral load ... has infected somebody else, anything could happen. They could progress slowly, or very, very rapidly.” Contrary to Dacus’s argument, the evidence in this record does not raise a substantial conflict with the “magnitude of harm” prong.
Turning now to the first prong, we address “risk of harm” and consider whether risk of HIV infection is “more than merely a fanciful, speculative, or remote possibility.” Weatherspoon, 49 M.J. at 211; Joseph, 37 M.J. at 397. The testimony in the record established that although the risk of transmitting the virus was low and therefore arguably “remote,” the risk was certainly more than fanciful or speculative. Dacus admitted that the presence of HIV in pre-ejaculate fluid makes transmission of the virus during sexual intercourse possible, even while wear*240ing a condom. This testimony is consistent with the testimony of Dr. Wallace that “there is no question, he could have transmitted HIV” through sexual intercourse. While Dr. Wallace indicated that it was “very unlikely” that Daeus would transmit the virus on account of his low viral load, he noted instances where individuals with viral loads similar to Daeus’s transmitted HIV by means of sexual intercourse.
In Weatherspoon we noted that “[t]he likelihood of death or grievous bodily harm is determined by measuring both prongs, not just the statistical risk of harm. Where the magnitude of the harm is great, there may be an aggravated assault, even though the risk of harm is statistically low.” 49 M.J. at 211. While the risk here may have been low, the magnitude of harm was significant. As such, we conclude that neither Dacus’s nor Dr. Wallace’s testimony was in substantial conflict with the “risk of harm” prong.
DECISION
The decision of the United States Army Court of Criminal Appeals is affirmed.
. Dr. Wallace explained that "viral load” is a measure of how much HIV virus is in the blood. He testified that while Dacus was infected with HIV, his viral load was so low that it was not detectable with existing technology.
. Dr. Wallace testified that while this was a controversial political and medical issue, he thought the best study suggested that using a condom would reduce the risk of transmission by eighty to ninety-five percent.
. Neither party has asserted or argued that this "risk of harm” standard of "more than merely a fanciful, speculative, or remote possibility” is inconsistent with the language of Article 128, UCMJ, and we therefore do not address that issue.