Karakatsanis J. -
 Sexual assault trials raise unique challenges in protecting the integrity of the trial and balancing the societal interests of both the accusedand the complainant. Parliament and the courts have responded to these challenges by setting out rules of evidence tailored to this context.
 Parliament enacted s. 276 of the Criminal Code, R.S.C. 1985, c. C-46,to govern the accused’s right to introduce evidence regarding the complainant’s prior sexual activity. Such evidence is never admissible to support the twin myths that the complainant is less worthy of belief or more likely to have consented to the sexual activity in question. In order to respect the presumption of innocence, evidence may be adduced for other relevant purposes but must satisfy rigorous criteria to ensure it does not undermine the integrity of the trial or the complainant’s dignity and privacy.
 The issue in this case is how these requirements apply where the Crown introduces evidence relating to the complainant’s sexual activity and the accused seeks to challenge that evidence by cross-examining the complainant.
 Here, the complainant testified that she was a virgin at the time of the assault. The Crown introduced evidence of her subsequent pregnancy and the approximate date of conception to support the complainant’s testimony that she was sexually assaulted by the accused. The accused denied the allegations and sought to question the complainant as to whether anyone else could have caused the pregnancy.
 The application judge ruled that the accused was not permitted to ask whether the complainant had engaged in any other sexual activity because the accused had no evidence of “specific instances of sexual activity” — one of the requirements of s. 276(2) of the Criminal Code. The accused was, however, permitted to cross-examine the complainant about her claim that she was a virgin at the time of the assault.
 I conclude that the application judge erred. The cross-examination sought to establish that the pregnancy was caused by sexual activity other than the alleged assault. The Crown-led evidence implicated a specific sexual act, namely activity capable of causing pregnancy within a particular timeframe. The accused’s request satisfied the “specific instances” requirement of s. 276(2) because it was sufficiently detailed to permit the judge to apply the regime.
 The Crown clearly intended to rely on evidence of the pregnancy to establish the actus reus. The presumption of innocence requires the accused to be permitted to test such critical, corroborating physical evidence before it can be relied on to support a finding of guilt. Given the accused’s denial of any sexual contact with the complainant, and the lack of other evidence of paternity, the ability to cross-examine the complainant was fundamental to his right to make full answer and defence.
 Nonetheless, permitting an accused to question a complainant about such matters treads on dangerous ground, raising both dignity and privacy concerns. Judges must tightly control such cross-examination to minimize those risks. The accused’s right to make full answer and defence must be balanced with other interests protected in s. 276(3). Here, balancing those interests would have required any cross-examination to be narrow in scope.
 That said, I am of the view that no miscarriage of justice occurred in this case. The cross-examination that was permitted and actually occurred allowed the defence to test the evidence with sufficient rigour. I would allow the appeal and restore the conviction.
 The accused and the complainant are cousins. During the Canada Day weekend of 2013, they went camping with several members of their extended families. At the time, R.V. was 20 years old and the complainant was 15. The complainant testified that R.V. sexually assaulted her in the early morning hours of July 1st.
 The complainant testified that on the families’ last night together, she played cards and hung out with her cousins. The parents in the group went to bed around 2:00 a.m. while the complainant stayed awake with several of her cousins. She said that around 4:00 a.m., R.V. suggested the cousins go to the beach for a swim. After briefly venturing into the lake, the cousins headed back to their campsites. R.V. borrowed the complainant’s phone to use as a flashlight on the walk back.
 The complainant explained that she returned to her tent, realized R.V. still had her phone and went to his tent to retrieve it. Upon her arrival at his tent, R.V. told the complainant he needed to speak to her in private. He took her by the wrist and led her into a men’s washroom near the beach.
 Once inside the washroom, the complainant said R.V. tried to kiss her and remove her shirt, which she resisted. Then he told her to lay down on the floor. She complied out of fear and he pulled down her pants and underwear as well as his own. R.V. then placed himself on top of her. She believes he tried to penetrate her vagina with his penis but her memory of this moment is blank. Her next memory is of R.V. asking her if she heard one of the cousins calling for her. After warning the complainant not to tell anyone what had happened, R.V. got up, put his clothes on and left. After he left, the complainant went to the women’s washroom to clean up. She said the area outside her vagina felt wet and sticky and she felt disgusted. She then returned to her tent and fell asleep.
 R.V. denied the allegations and the complainant’s account of the evening. He acknowledged being around the bonfire with the family. But, he stated, he went to bed at around the same time as the parents and had no further interaction with the complainant that evening.
 Initially, the complainant did not tell anyone about the assault. In late August, she went to see a doctor, complaining of abdominal pain and nausea. When asked, she denied being sexually active. A urine test taken during a physical exam on August 29th subsequently confirmed that she was pregnant. Based on an ultrasound performed on September 18th, the doctor estimated that conception had occurred at the end of June or the beginning of July.
 The doctor asked the complainant whether she had any interactions capable of causing pregnancy around the date of conception. During this discussion, the complainant told the doctor about the incident with R.V. Because the complainant was underage, later that day the doctor relayed this information to the Children’s Aid Society, which in turn contacted the police.
 After consulting with her doctor again on September 19th, the complainant terminated the pregnancy on September 21st. The clinic disposed of the fetal remains that day, making a DNA paternity test impossible. The police contacted the complainant the following week and took a statement from her on September 24th. The complainant told the police (and testified at trial) that she was a virgin at the time of the assault. The police charged R.V. with sexual assault and sexual interference.
A. The Voir Dire: Baker J.
 In light of the medical evidence about the date of conception, the Crown tendered evidence of the pregnancy to support the complainant’s testimony that the alleged assault was the cause. During pre-trial proceedings, R.V. applied to question the complainant “about her prior sexual activity, with the [accused], or any other individual, that may have occurred between June 1st and July 1st, 2013” (A.R., vol. II, at p. 3). Because the Crown intended to rely on the complainant’s pregnancy as evidence of sexual contact with the accused, R.V. argued his right to make full answer and defence entitled him to inquire into “whether any other individual could have impregnated the complainant” (A.R., vol. II, at p. 4).
 Baker J. dismissed the s. 276 application. In her view, R.V. had failed to point to specific instances of sexual activity. Instead, she found, the request was “more in the nature of a fishing expedition” (A.R., vol. I, at p. 15).
 The application judge also observed that other means could be used to challenge the inference that R.V. caused the complainant’s pregnancy. First, the gestational age of the fetus might not align with the date of the alleged offence. She based this conclusion on the voir dire submissions of R.V.’s counsel, who suggested the date of conception was approximately June 14th, 17 days prior to the assault.
 Second, R.V. could question the complainant on her understanding of the term “virgin” and the truthfulness of her statement that she was a virgin at the time of the alleged offence. The application judge held that questions regarding virginity do not fall within s. 276. However, R.V.’s right to ask these limited questions did not “give the defence carte blanche to cross-examine the complainant on any sexual activity she may have undertaken in the month immediately preceding the alleged offence” (A.R., vol. I, at p. 15).
 The app
Please Login To View The Full Judgment!
lication judge accepted that R.V. was not intending to rely on the cross-examination evidence to further the twin myths. Nevertheless, the complainant’s personal dignity and right of privacy outweighed the “highly uncertain probative value of the proposed evidence” (A.R., vol. I, at p. 16).
B. The Trial: Gee J.
 After the voir dire and prior to trial, the application judge invoked s. 669.2 of the Criminal Code and the trial continued before another judge. At the outset of the trial, Gee J. declined R.V.’s request to re-litigate the s. 276 application. He held that where a trial is continued by another judge, s. 669.2 does not provide for the re-consideration of pre-trial motions decided by the previous judge. In any event, he concluded there was no reason to re-hear the application because no change of circumstances had occurred.
 The complainant, the complainant’s doctor and the accused testified at trial. During the doctor’s testimony, it became clear that defence counsel’s calculation of the conception date at the voir dire was wrong. Based on the ultrasound, conception would have occurred between June 21st and July 5th, 2013.
 The trial judge rejected R.V.’s account of the night in question. In assessing R.V.’s credibility, he considered inconsistencies in R.V.’s trial testimony and interview with the police, as well as his admission that he had lied during both. The trial judge ultimately concluded that R.V. was not a credible witness and did not accept his evidence.
 In contrast, the trial judge found the complainant to be “a very compelling witness” (A.R., vol. I, at p. 46) whose account was detailed and precise. He accepted her testimony that prior to July 1st she had never had intercourse. He also concluded that her pregnancy was “compelling evidence supportive... of her allegations” (A.R., vol. I, at p. 47).
 R.V. was convicted of sexual interference and received a four-year custodial sentence.
C. The Court of Appeal: MacFarland, Watt and Paciocco JJ.A.
 Writing for the Court of Appeal for Ontario, Paciocco J.A. allowed the appeal and ordered a new trial. He held that it was patently unfair for the Crown to rely on the pregnancy as confirming the complainant’s story while preventing the accused from challenging this inference. While the request to cross-examine the complainant about her sexual activity “that may have occurred between June 1st and July 1st, 2013” was “extravagant”, the accused should still have had the opportunity to ask relevant questions: 2018 ONCA 547, 141 O.R. (3d) 696, at paras. 27 and 29-30.
 Paciocco J.A. held s. 276(2) requires the accused to adequately identify targeted evidence and proposed lines of questioning so as to permit the application judge to perform the balancing required by s. 276. Here, the probative value of the cross-examination was clear — it had the potential to neutralize the Crown’s reliance on the pregnancy. Uncertainty about whether a line of questioning will succeed does not eliminate its potential probative value. Thus, the application judge erred in requiring an evidentiary foundation for the proposed cross-examination.
 Paciocco J.A. also concluded that the application judge had erred in characterizing the requested cross-examination as a “fishing expedition”. In his view, the proposed questioning was “responsive to an important plank in the Crown’s case”, and the alternative methods of challenging the complainant’s claim of virginity did not adequately protect the accused’s right to a fair trial: paras. 69-84.
 According to the Court of Appeal, the impact of the application judge’s errors was compounded by the trial judge’s incorrect conclusion that he was bound by the initial s. 276 ruling. A new trial was therefore warranted because the trial judge’s refusal to reconsider the application effectively prevented future meritorious applications from being considered.
 The question in this case is how s. 276 operates when the accused seeks to cross-examine the complainant to challenge sexual history evidence led by the Crown. Section 276 requires that the accused’s right to make full answer and defence be balanced with the dangers that cross-examination may pose to the complainant’s privacy and dignity and to the integrity of the trial process. This analysis applies with equal force regardless of whether the accused seeks to introduce evidence to establish a defence or to challenge inferences urged by the Crown.
A. Challenging the Crown’s Evidence Within the Limits of Section 276
 Testifying in a sexual assault case can be traumatizing and harmful to complainants: see E. Craig, Putting Trials on Trial: Sexual Assault and the Failure of the Legal Profession (2018), at pp. 4 et seq. Questions about a complainant’s sexual history are often irrelevant, serving no purpose other than supporting the “twin myths” — that a complainant’s past sexual acts make her less worthy of belief or more likely to have consented to the sexual activity in question. Historically, wide-ranging and intrusive inquiries into the complainant’s sexual history were used to distort the trial process and essentially put the complainant on trial: see R. v. Goldfinch, 2019 SCC 38, at para. 33.
 In an effort to abolish “outmoded, sexist-based use of sexual conduct evidence”, Parliament, in 1982, enacted a blanket ban on all evidence of a complainant’s prior sexual activity, subject to three limited exceptions: R. v. Seaboyer,  2 S.C.R. 577, at p. 625. One exception covered evidence “that rebuts evidence of the complainant’s sexual activity or absence thereof that was previously adduced by the prosecution”: An Act to amend the Criminal Code in relation to sexual offences and other offences against the person and to amend certain other Acts in relation thereof or in consequences thereof, S.C. 1980-81-82-83, c. 125, s. 246.6(1)(a) (later s. 276(1)(a)). In Seaboyer, the Court struck down the 1982 provision as unconstitutional because it was too restrictive — it had the potential to exclude relevant evidence crucial to a fair trial: Seaboyer, at p. 625. The Court emphasized that the relevance of all evidence, including sexual history evidence, must be assessed on a case-by-case basis: p. 609.
 In response, Parliament amended s. 276, codifying the guidelines outlined in Seaboyer: R. v. Darrach, 2000 SCC 46,  2 S.C.R. 443, at para. 20. The modern version of s. 276 seeks to preserve the integrity of the administration of justice, and the trial, by striking a balance between the rights of the accused and those of the complainant:
276 (1)In proceedings in respect of [various sexual offices], evidence that the complainant has engaged in sexual activity, whether with the accused or with any other person, is not admissible to support an inference that, by reason of the sexual nature of that activity, the complainant
(a) is more likely to have consented to the sexual activity that forms the subject-matter of the charge; or
(b) is less worthy of belief.
(2) In proceedings in respect of an offence referred to in subsection (1), no evidence shall be adduced by or on behalf of the accused that the complainant has engaged in sexual activity other than the sexual activity that forms the subject-matter of the charge, whether with the accused or with any other person, unless the judge, provincial court judge or justice determines, in accordance with the procedures set out in sections 276.1 and 276.2, that the evidence
(a) is of specific instances of sexual activity;
(b) is relevant to an issue at trial; and
(c) has significant probative value that is not substantially outweighed by the danger of prejudice to the proper administration of justice.
(3) In determining whether evidence is admissible under subsection (2), the judge, provincial court judge or justice shall take into account
(a) the interests of justice, including the right of the accused to make a full answer and defence;
(b) society’s interest in encouraging the reporting of sexual assault offences;
(c) whether there is a reasonable prospect that the evidence will assist in arriving at a just determination in the case;
(d) the need to remove from the fact-finding process any discriminatory belief or bias;
(e) the risk that the evidence may unduly arouse sentiments of prejudice, sympathy or hostility in the jury;
(f) the potential prejudice to the complainant’s personal dignity and right of privacy;
(g) the right of the complainant and of every individual to personal security and to the full protection and benefit of the law; and
(h) any other factor that the judge, provincial court judge or justice considers relevant.
 Before evidence of a complainant’s sexual history may be introduced under s. 276(2), the court must carefully scrutinize the potential evidence. First, the accused must set out in writing the “detailed particulars of the evidence that the accused seeks to adduce” and its relevance “to an issue at trial”: s. 276.1(2). If the judge is persuaded that the evidence is “capable of being admissible under subsection 276(2)”, a voir dire is held: s. 276.1(4). Evidence adduced to support the twin myths is categorically barred. And, even where it has some relevance for another purpose, evidence may still be excluded if admitting it would endanger the “proper administration of justice”: s. 276(2)(c).
 In many cases, when the accused applies to adduce evidence pursuant to s. 276, they have an evidentiary basis for known sexual activity (i.e. “detailed particulars” of “specific sexual activity”). In this case, however, while R.V. maintained that other activity necessarily occurred because the complainant became pregnant, he had little knowledge of the particulars of that activity. He thus sought to adduce evidence of the other sexual activity by cross-examining the complainant.
 Individuals charged with criminal offences are presumed innocent until proven guilty. As a result, an accused has the right to call the evidence necessary to establish a defence and to challenge the prosecution’s evidence: R. v. Osolin,  4 S.C.R. 595, at p. 663. “Full answer and defence” is a principle of fundamental justice, protected by s. 7 of the Canadian Charter of Rights and Freedoms. In Seaboyer, McLachlin J. explained, at p. 608:
The right of the innocent not to be convicted is dependent on the right to present full answer and defence. This, in turn, depends on being able to call the evidence necessary to establish a defence and to challenge the evidence called by the prosecution.
In short, the denial of the right to call and challenge evidence is tantamount to the denial of the right to rely on a defence to which the law says one is entitled.
 Generally, a key element of the right to make full answer and defence is the right to cross-examine the Crown’s witnesses without significant and unwarranted restraint: R. v. Lyttle, 2004 SCC 5,  1 S.C.R. 193, at paras. 1 and 41; Osolin, at pp. 664-65;Seaboyer, at p. 608. The right to cross-examine is protected by both ss. 7 and 11(d) of the Charter. In certain circumstances, cross-examination may be the only way to get at the truth. The fundamental importance of cross-examination is reflected in the general rule that counsel is permitted to ask any question for which they have a good faith basis — an independent evidentiary foundation is not required: Lyttle, at paras. 46-48.
 However, the right to cross-examine is not unlimited. As a general rule, cross-examination questions must be relevant and their prejudicial effect must not outweigh their probative value: Lyttle, at paras. 44-45. In sexual assault cases, s. 276 specifically restricts the defence’s ability to ask questions about the complainant’s sexual history. By virtue of s. 276(3), full answer and defence is only one of the factors to be considered by the trial judge; it must be balanced against the danger to the other interests protected by s. 276(3). These additional limits are necessary to protect the complainant’s dignity, privacy and equality interests: Osolin, at p. 669; see also R. v. Mills,  3 S.C.R. 668, at paras. 61-68. They also aim to achieve important societal objectives, including encouraging the reporting of sexual assault offences: s. 276(3)(b).
 Thus, the fact that the accused’s ability to make full answer and defence requires that the complainant be cross-examined is not the end of the analysis. The scope of the permissible questioning must also be balanced with the danger to the other interests protected by s. 276(3), including the dignity and privacy interests of the complainant.
 Here, the Crown introduced evidence of the complainant’s pregnancy and virginity to corroborate her testimony that the assault occurred. In his s. 276 application, R.V. sought to challenge that inference by questioning the complainant about her sexual activity from June 1st to July 1st, 2013 in order to determine “whether any other individual could have impregnated the complainant” (A.R., vol. II, at pp. 3-4).
 When the accused seeks to inquire into the complainant’s sexual history, the three subsections of s. 276 work together to achieve the provision’s objectives.
 Section 276(1) sets out an absolute bar against introducing evidence for the purpose of drawing twin-myth inferences. Here, R.V.’s request to challenge the inference that the pregnancy resulted from the alleged assault did not engage the twin myths. As such, the application judge correctly concluded that the cross-examination was not barred by s. 276(1).
 If, as in this case, evidence that the complainant has engaged in sexual activity is to be adduced for another purpose, it is presumptively inadmissible unless the accused satisfies s. 276(2)(a) and (b) by identifying specific instances of sexual activity, relevant to an issue at trial. Where the accused applies to cross-examine the complainant about her sexual history to challenge Crown-led evidence, the analysis will often turn on the balancing exercise mandated by s. 276(2)(c). This third step involves weighing the factors set out in s. 276(3) to determine whether the probative value of the cross-examination is significant enough to substantially outweigh the dangers of prejudice to the proper administration of justice. This provision requires judges to determine the permissible scope of cross-examination in light of the competing rights of the accused and the complainant and the other interests set out in s. 276(3). Where the right to full answer and defence requires some cross-examination, judges should tailor their rulings to best safeguard the other interests protected by s. 276(3).
 I now turn to the application of s. 276(2) to the facts of the case.
(1) Section 276(2)(a):“Specific Instances of Sexual Activity”
 Broad exploratory questioning is never permitted under s. 276. Open-ended cross-examination concerning a complainant’s sexual history clearly raises the spectre of the impermissible uses of evidence that the provision was intended to eliminate. Section 276(2)(a) requires the accused to identify “specific instances of sexual activity” to avoid unnecessary incursions into the sexual life of the complainant.
 That said, the words “specific instances of sexual activity” must be read purposively and contextually. They limit admissible evidence to discrete sexual acts, and protect against the misuse of general reputation evidence to discredit the complainant and distort the trial process: see R. v. L.S., 2017 ONCA 685, 40 C.R. (7th) 351, at paras. 79-80. The “specific instances” requirement is buttressed by the procedural aspects of a s. 276 application, which require the accused to set out “detailed particulars” of the evidence to be adduced: s. 276.1(2). By requiring “detailed particulars,” the Criminal Code ensures that judges are equipped to meaningfully engage with the s. 276 analysis and that defence evidence does not take the Crown or complainant by surprise: Darrach, at para. 55, Goldfinch, at para. 51; see also L.S., at paras. 82-85.
 Section 276(2)(a) does not always require an accused to come before the court armed with names, dates and locations. As counsel for the intervener Criminal Lawyers’ Association of Ontario pointed out, requiring such details may, in some cases, be unduly intrusive, defeating one of the provision’s most important objectives. Rather, as Doherty J.A. observed in L.S. and this Court affirmed in Goldfinch, the degree of specificity required depends on the circumstances of the case, the nature of the sexual activity that the accused seeks to adduce and the use to be made of that evidence: L.S., at para. 83; Goldfinch, at para. 53.
 Here, R.V. proposed to cross-examine the complainant about other sexual activity that could have caused her pregnancy, without knowing what her answers would be. Because his defence was a bare denial, he obviously sought to establish that the pregnancy was the result of some other activity. Is this enough to qualify as “detailed particulars” of “specific instances of sexual activity” in the circumstances of this case? Does it provide sufficient notice to the complainant and Crown and equip the judge to apply s. 276?
 The Crown argues that it does not. In its view, s. 276 requires the accused to identify “concrete particulars” and to provide an evidentiary basis in support of a s. 276 application. The Crown agreed with the application judge that R.V.’s request amounted to nothing more than a “fishing expedition” — precisely the kind of free-ranging cross-examination s. 276 aims to prevent.
 In my view, the application judge erred when she concluded that the accused failed to identify evidence of specific instances of sexual activity. As Paciocco J.A. rightly noted, this requirement must be interpreted purposively and the objectives of the requirement would have been satisfied here.
 R.V. sought to cross-examine the complainant on a specific instance of sexual activity — the activity that caused her pregnancy — evidence of which was introduced by the Crown. The pregnancy itself demonstrated only that sexual activity capable of impregnating the complainant took place around July 1st. The existence of such activity was not speculative. But the fact of pregnancy here did not reveal exactly when or with whom that sexual activity occurred. The proposed cross-examination was directed at challenging the inference that R.V. caused the pregnancy.
 During oral arguments before this Court, Crown counsel submitted that a bare denial cannot satisfy the “specific instances” requirement. However, this submission turns the presumption of innocence on its head. The Crown’s assertion that the pregnancy arose from the sexual activity that formed the subject-matter of the charge cannot prevent the accused from leading evidence to suggest that the pregnancy was caused by someone else or by some other sexual act. The presumption of innocence requires that R.V. be allowed to challenge the Crown’s evidence that he committed a sexual assault. Of course, the trier of fact may ultimately reject the accused’s denial. But, as Paciocco J.A. emphasized, it would be unfair for the Crown to rely on the complainant’s testimony that the accused caused the pregnancy while at the same time preventing the accused from challenging the complainant’s account.
 Moreover, pregnancy is evidence of sexual activity that can be situated within a particular timeframe. R.V.’s s. 276 application set out a specific, albeit broad, timeframe of one month. At trial, the medical evidence established a rough two-week window during which conception would have occurred. In either case, the clearly identified time period, along with the specific nature of the activity — activity capable of causing pregnancy — was sufficiently specific to satisfy s. 276(2)(a).
(2) Section 276(2)(b):“Relevant to an Issue at Trial”
 The accused’s s. 276 application must also identify the relevance of the evidence to be adduced. As a matter of logic, evidence tendered to rebut Crown-led evidence implicating the accused will be relevant to the accused’s defence. As noted above, even in the 1982 iteration of s. 276, Parliament carved out an exception for evidence rebutting Crown-led evidence of the complainant’s sexual activity or absence thereof.
 InSeaboyer, McLachlin J. noted that the complainant’s other sexual activity “may be relevant to explain the physical conditions on which the Crown relies to establish intercourse or the use of force, such as semen, pregnancy, injury or disease”: p. 614. L’Heureux-Dub J., writing in dissent, agreed that where the Crown contends that physical consequences such as pregnancy were caused by an assault, the defence may adduce sexual history evidence in rebuttal: p. 682.
 In this case, the Crown suggests that because the answers to R.V.’s questions were unknown, the relevance of those questions was speculative. I cannot accept this proposition. The relevance of the proposed questioning was clear. The Crown relied on the pregnancy as corroborating the complainant’s account. Regardless of her answers, the complainant’s testimony would be relevant. If she denied the existence of other sexual activity, this could strengthen the Crown’s case. But if other sexual activity could have occurred during the relevant time period, the probative value of the pregnancy would be significantly reduced.
 Given the clear relevance of challenging Crown-led evidence, in cases like the present appeal, the outcome of the analysis will generally turn on balancing the probative value of the evidence with its potential prejudice to the complainant and the proper administration of justice.
(3) Section 276(2)(c): Balancing Competing Interests
 Even where proposed evidence is sufficiently specific and relevant, cross-examination about a complainant’s sexual history is only allowed if the proposed line of questioning has “significant probative value that is not substantially outweighed by the danger of prejudice to the proper administration of justice”: s. 276(2)(c). This balancing requires judges to pay careful attention to the factors listed in s. 276(3) in assessing the potential impact of the evidence on the accused, the complainant and the administration of justice.
 The application judge appears to have concluded that the s. 276(3) factors also militated against granting R.V.’s application. At the voir dire, defence counsel suggested that the evidence indicated that conception occurred mid-June. This in itself weakened the Crown’s inference that the pregnancy resulted from the July 1st assault. In addition, the application judge was of the view that R.V. could question the complainant about her claim of virginity without engaging s. 276. Accordingly, she held that the probative value of any further questions about the complainant’s sexual activity was highly uncertain. In her assessment, the complainant’s dignity and privacy interests outweighed the “speculative” probative value of the proposed questioning.
 While R.V. did not know the answers to the questions he sought to ask, I agree with Paciocco J.A. that “uncertainty of result does not deprive a line of questioning of its probative value”: para. 64. The application judge should not have considered the probability that R.V.’s questioning would be successful, but rather whether the answers would be probative. Because the answers had the potential to undermine or confirm important Crown evidence, their probative value was high. In my view, two factors related to the probative value of the evidence required that some form of cross-examination of the complainant be allowed:
(a) the interests of justice, including the right of the accused to make a full answer and defence; [and]
(c) whether there is a reasonable prospect that the evidence will assist in arriving at a just determination in the case.
 InR. v. Crosby,  2 S.C.R. 912, L’Heureux-Dub J. wrote that “[s]ection 276 cannot be interpreted so as to deprive a person of a fair defence”: para. 11; see also Darrach, at para. 43; Seaboyer, at p. 616. Thus, in weighing how the accused may respond to Crown-led evidence, the judge must ensure the accused is not denied the right to make full answer and defence.
 Simply put, the more important evidence is to the defence, the more weight must be given to the rights of the accused. For example, the need to resort to questions about a complainant’s sexual history will be significantly reduced if the accused can advance a particular theory without referring to the complainant’s sexual history. But in other circumstances — where challenging the Crown’s evidence of the complainant’s sexual history directly implicates the accused’s ability to raise a reasonable doubt — cross-examination becomes fundamental to the accused’s ability to make full answer and defence and must be allowed in some form: Mills, at paras. 71 and 94.
 This is such a case. Here, there was no independent evidence of paternity. In light of R.V.’s denial, the only way he could challenge the inference urged by the Crown was by cross-examining the complainant with respect to other sexual activity. In these circumstances, the complainant’s privacy must yield to cross-examination in order to avoid convicting the innocent.
 In addition, where the accused’s defence involves challenging Crown-led evidence, cross-examination assists the trier of fact in arriving at a just determination: Seaboyer, at p. 609; see also S. Ozkin, “Balancing of Interests: Admissibility of Prior Sexual History under Section 276” (2011), 57 C.L.Q. 327, at pp. 331-32. Because the window for conception overlapped with the date of the alleged assault, the pregnancy could corroborate the complainant’s account. However, the trial judge’s ability to rely on the pregnancy depended on the extent to which he could rule out the possibility that other sexual activity had caused it. Thus, both ss. 276(3)(a) and 276(3)(c) required permitting some form of cross-examination of the complainant.
 This does not, however, open a door to wide-ranging inquiries. The right to a fair trial does not guarantee the most favourable procedures imaginable: R. v. Quesnelle, 2014 SCC 46,  2 S.C.R. 390, at para. 64; Darrach, at para. 24; Mills, at para. 75. Even where the right to a fair trial requires cross-examination of the complainant, it does not entitle an accused to pursue the most expansive cross-examination. The scope of the permissible questioning must also be determined by balancing the accused’s rights with the other rights and interests protected by s. 276(3), including:
(f) the potential prejudice to the complainant’s personal dignity and right of privacy; [and]
(g) the right of the complainant and of every individual to personal security and to the full protection and benefit of the law.
The trial judge must therefore narrow the scope of the questioning to minimize the impact on the complainant, while maintaining the accused’s ability to answer the charges.
 Inquiries into any individual’s sexual history are highly intrusive. The threat to the complainant’s dignity and privacy is even higher when the proposed questions surround the conduct of a 15-year-old. Both the length of the relevant time period and the degree of detail to be adduced impact the potential prejudice to the complainant: see, e.g., R. v. Nkemka, 2013 ONSC 2121, at paras. 10-20 (CanLII). Open-ended questioning about an individual’s sexual activity, even during a particular timeframe, risks devolving into the very type of inquiry that s. 276 was intended to prevent. Caution must be exercised where the proposed inquiry captures a broad range of sexual activity and is limited only by a specified timeframe.
 Determining the boundaries of permissible cross-examination will always be a challenging and fact-specific task. In the present case, where the complainant maintained she was a virgin and had no boyfriend at the time of the assault, only limited questioning was appropriate. In other cases, more latitude may be warranted.
 In summary, R.V.’s right to make full answer and defence required some means of challenging the Crown’s reliance on the pregnancy. In my view, a correct balancing of the interests set out in s. 276(3) would have allowed R.V. to inquire into: (i) the complainant’s understanding of the types of sexual activity capable of causing pregnancy and (ii) whether she engaged in any such activity at the end of June and the beginning of July. However, to minimize the impact on the complainant’s privacy and dignity, that inquiry needed to be limited. As I discuss below, even when a s. 276 application is granted, the trial judge must continue to keep these interests in mind throughout the trial.
(4) The Trial Judge’s Gatekeeper Role
 This Court recently emphasized the important “gatekeeper” role trial judges play in sexual assault cases: R. v. Barton,2019 SCC 33, at paras. 68 and 197; Goldfinch, at para. 75. The procedural and evidentiary context of this case illustrates two aspects of this role: (i) the importance of remaining alive to the objectives of s. 276 as the trial unfolds by actively supervising cross-examination and adapting s. 276 rulings as necessary when new evidence comes to light; and (ii) the advantages of assessing the evidence of other sexual activity to be adduced by both the defence (as required by s. 276) and the Crown (in light of s. 276(1) and the common law Seaboyer principles).
(a) Monitoring Cross-Examination and Re-Visiting Section 276 Applications Throughout the Trial
 Section 276 continues to operate even after an initial evidentiary ruling has been rendered. Trial judges must therefore remain vigilant in ensuring the objectives of the provision are upheld as the trial unfolds. Cross-examination about the complainant’s sexual history, where permitted, should be closely monitored to ensure it remains within the permissible limits. And as evidence emerges, it may become necessary to re-consider prior s. 276 rulings.
 First, where targeted cross-examination of the complainant is permitted, trial judges must strike a delicate balance between giving counsel sufficient latitude to conduct effective cross-examination and minimizing any negative impacts on the complainant and the trial process. Proposed questions should be canvassed in advance and may be re-assessed based on the answers received. In certain cases, it may even be appropriate to approve specific wording: see, e.g., Nkemka, at para. 18; R. v. Akumu, 2017 BCSC 533, at paras. 26-31 and 35-54 (CanLII).
 Second, as a general rule, an order related to the conduct of trial may be varied or revoked if there is a material change of circumstances: C.A. reasons, at paras. 98-103; see also R. v. Adams, 4 S.C.R. 707, at para. 30; R. v. Calder,  1 S.C.R. 660, at para. 21; R. v. La,  2 S.C.R. 680, at para. 28. As evidence emerges at trial, both the probative value and potential prejudice of proposed evidence may change. If a material change of circumstances occurs, either party may request that a previous evidentiary ruling be re-visited.
 In this case, the application judge refused R.V.’s request to question the complainant on her sexual history. At the outset of the trial, the trial judge subsequently held that because the trial had been continued before him under s. 669.2 of the Criminal Code, the pre-trial motions could not be re-litigated and he was bound by the application judge’s s. 276 ruling. But as the Court of Appeal correctly stated, s. 669.2 does not displace the general rule that a trial judge has discretion to re-consider rulings made earlier in the proceedings if there is a material change of circumstances: paras. 98-108.
 In this case, the trial judge also observed, correctly, that no material change of circumstances had occurred between the s. 276 ruling and the start of trial. Nevertheless, given that the trial judge held that he could not re-consider the ruling — which included the erroneous conclusion that the proposed cross-examination did not qualify as a specific instance of sexual activity — I accept that counsel for the accused may have thought it would be futile to apply for a re-consideration, even if the circumstances changed during the trial.
 Indeed, the evidentiary foundation did shift in this case. At the voir dire, counsel for R.V. suggested that the date of conception was approximately June 14th — 17 days prior to the alleged offence. However, the doctor’s testimony at trial established that conception would have occurred between June 21st and July 5th. As the potential conception date coincided more closely with the date of the assault, the probative value of the pregnancy as evidence of the assault increased. The narrower timeframe also decreased the potential prejudice to the complainant. Both are factors that would likely have provided grounds for re-considering the s. 276 ruling.
(b) The Admissibility of Crown Evidence and the Possibility of Joint Assessments
 While s. 276(2) applies only to evidence “adduced by or on behalf of the accused”, s. 276(1) and the common law principles apply to Crown-led evidence of a complainant’s sexual history: Barton,at para. 80. In Seaboyer, McLachlin J. emphasized the importance of the trial judge’s gatekeeper role in ensuring that sexual history evidence “possesses probative value on an issue in the trial... [that] is not substantially outweighed by the danger of unfair prejudice flowing from the evidence”: p. 635. Irrespective of which party adduces evidence of the complainant’s sexual history, the trial judge must guard against twin-myth reasoning as well as prejudice to the complainant, the trial process and the administration of justice.
 Where, as in this case, the accused’s s. 276 application relates to Crown-led evidence, it would be prudent to consider both the Crown’s proposed use of the evidence and any challenges proposed by the accused at the same time. A view of how both sides intend to use the evidence would allow trial judges to more accurately assess the impact of admitting such evidence and appropriately tailor the ways in which it may be adduced. Further, the Crown’s decision to adduce evidence, or even to call a particular witness, is a matter of prosecutorial discretion: Darrach, at para. 69. If the manner in which the evidence may be challenged is clear from the outset, the Crown can make an informed decision about whether the interests of justice are served by adducing the evidence in the first place.
 Here, the Crown introduced evidence of the complainant’s sexual history. In direct examination, the Crown asked the complainant if she was a “virgin” at the time of the assault and when that “physical state” changed. The complainant responded that she was a virgin on July 1st and indicated that she had sexual intercourse for the first time on September 2nd. The complainant’s doctor also testified to conversations regarding the complainant’s sexual activity. It is not clear on the record why the Crown adduced this evidence in this way. Whether these statements were admissible, as well as how the defence would be permitted to challenge them, should have been decided in advance.
 Whether sexual inactivity is captured by either s. 276 or the Seaboyer principles is not directly at issue before this Court. There is appellate authority stating that s. 276 does not prevent the complainant from testifying as to virginity: R. v. Pittiman (2005), 198 C.C.C. (3d) 308 (Ont. C.A.), aff’d 2006 SCC 9,  1 S.C.R. 381, on a different point, at para. 33; R. v. Brothers (1995), 169 A.R. 122 (C.A.), at paras. 26-29. However, these cases also recognize that admitting evidence of virginity raises further questions, including: (i) the inferences the finder of fact may be asked to draw from the fact of the complainant’s virginity and (ii) how the accused may challenge this claim: see Pittiman, at paras. 34-37; Brothers, at paras. 30-35. While I leave this issue for another day, I agree with Paciocco J.A. that it would be incongruous to hold that the statement “I am a virgin” does not engage s. 276 while an answer to the contrary would clearly be a reference to sexual activity: para. 79.
 Nonetheless, questions regarding when the complainant ceased to be a virgin undoubtedly fell within the ambit of s. 276 and the Seaboyerprinciples. In this case, the Crown presumably sought to confirm that the complainant had not engaged in sexual activity during the timeframe when conception could have occurred. How the Crown intended to adduce this evidence — and whether discussion of her activity on September 2nd,well beyond the conception timeframe, was necessary — should have been evaluated in advance and considered alongside R.V.’s s. 276 application.
B. Section 686(1)(b)(iii): Has a Miscarriage of Justice Occurred?
 The application judge erred in adopting an overly restrictive approach to s. 276. Asking the complainant whether she had engaged in other sexual activity that could have resulted in pregnancy during the relevant timeframe was sufficient to satisfy the “specific instances” requirement of s. 276(2). Some cross-examination on other possible causes of the pregnancy was warranted to safeguard R.V.’s ability to defend himself on the charges. Further, the trial judge erred in concluding, at the outset of the trial, that he did not have the discretion to re-hear the s. 276 application. As I explained above, given the application judge’s refusal to grant the application, the evidence that emerged at trial would likely have constituted a material change of circumstances, justifying a re-consideration. The effect of these errors, however, must be viewed in light of the fact that the trial judge permitted the defence to cross-examine the complainant on the issue of virginity.
 Section 686(1)(b)(iii) of the Criminal Code permits a court of appeal to dismiss an appeal from a conviction where “no substantial wrong or miscarriage of justice has occurred”. In my view, R.V. suffered no substantial wrong because despite these errors, the questions he was permitted to ask allowed him to adequately challenge the inference urged by the Crown.
 The curative proviso set out in s. 686(1)(b)(iii) may be applied where there is no “reasonable possibility that the verdict would have been different had the error... not been made”: R. v. Bevan, 2 S.C.R. 599, at p. 617, aff’d R. v. Khan, 2001 SCC 86,  3 S.C.R. 823, at para. 28. Applying the curative proviso is appropriate in two circumstances: (i) where the error is harmless or trivial; or (ii) where the evidence is so overwhelming that the trier of fact would inevitably convict: R. v. Sekhon, 2014 SCC 15,  1 S.C.R. 272, at para. 53; R. v. Van, 2009 SCC 22,  1 S.C.R. 716, at para. 34; Khan, at paras. 29-31.
 Cross-examination is undoubtedly a key element of the right to make full answer and defence. This Court has held that sometimes “there will be no other way to expose falsehood, to rectify error, to correct distortion or to elicit vital information that would otherwise remain forever concealed”: Lyttle, at para. 1 (emphasis in original); see also Osolin, at p. 663. Thus, as a general rule, counsel “may pursue any hypothesis that is honestly advanced on the strength of reasonable inference, experience or intuition”: Lyttle, at para. 48. Because it is difficult to predict what lines of questioning counsel might pursue and what evidence may have emerged had cross-examination been permitted, a failure to allow relevant cross-examination will almost always be grounds for a new trial: Shearing, at para. 151; Crosby, at para. 20;Osolin, at pp. 674-75.
 The key question at this stage is whether the errors in this case prevented R.V. from making full answer and defence. More specifically, is it clear that R.V. was able to adequately challenge the inference that the pregnancy confirmed his participation in the assault? Here, the Court of Appeal did not specifically address the scope of the permissible cross-examination. However, it held that whatever cross-examination occurred was “not a fair substitute for the cross-examination that should have been allowed”: para. 91.
 I disagree. In this case — where the proposed cross-examination involved the conduct of a 15-year-old who testified she was a virgin — wide-ranging questions would have been inappropriate. Instead, the circumstances of this case warranted tightly-controlled questioning of the complainant. R.V.’s right to make full answer and defence entitled him to test whether someone else could have caused the complainant’s pregnancy. As discussed above, a correct balancing of the interests set out in s. 276(3) would have allowed R.V. to make limited inquiries into: (i) the complainant’s understanding of the types of sexual activity capable of causing pregnancy and (ii) whether she engaged in any such activity at the end of June and the beginning of July. As I shall explain, the defence was allowed to do so — despite the errors in the s. 276 ruling.
 Before this Court, counsel for R.V. advanced the theory that the complainant had engaged in such sexual activity with her boyfriend during June or July. Afraid of the potential repercussions once her family found out she was pregnant, the complainant fabricated the story about the assault. Counsel argued a new trial is required because R.V. was unable to pursue this theory at trial and it is impossible to know whether this strategy would have succeeded.
 I remain unpersuaded. Having reviewed the trial transcripts in light of the questioning that actually occurred as well as that suggested by counsel, I am satisfied that the accused was not precluded from adequately testing the evidence in this case.
 First, the application judge’s ruling permitted R.V. to ask the complainant about her understanding of “virginity” and to challenge whether she was telling the truth about being sexually inactive.
 Defence counsel asked the complainant about her understanding of the types of sexual activity capable of causing pregnancy. He asked, “you knew that sexual intercourse could lead to pregnancy” and “if there was contact between the male genitals and the female genitals, sexual intercourse didn’t have to occur, but you could become... pregnant from that type of sexual conduct?” The complainant responded affirmatively to both questions (A.R., vol. V, at pp. 26-27).
 The complainant consistently maintained (to her doctor, the police and the court) that she was both a virgin and not sexually active prior to September 2013. In challenging this evidence, defence counsel asked about her definition of “virginity” and “sexual activity”. Among a number of questions on these issues, he asked, “your knowledge... of virginity is somebody who hasn’t had actual full sexual intercourse” and she replied, “Yes” (A.R., vol. V, at p. 28). At another point in her testimony, the complainant broadly defined “sexual activity” so as to include even the touching of genitals. Counsel also challenged the complainant’s statement to her doctor during the following exchange:
Q. And, [the doctor] asked you, specifically, if you had been having sexual activities?
A. Yes, she did.
Q. And, you said, no?
Q. And, that wasn’t accurate, what you’re telling the court is true?
A. I’m sorry.
Q. When you told the doctor you were having sexual activity, that was, wasn’t accurate?
A. But, that was true, because I wasn’t have sexual activity before September 2nd.
Q. Well, but the incident itself is a form of sexual activity, wouldn’t you agree?
A. Yes, but, at that time, I, I hadn’t told anyone, so I didn’t…
Q. You didn’t want to tell her?
A. I, I didn’t. I felt super uncomfortable.
Q. Okay. And, so, you said, no, even though you, in your mind, you probably knew that this had happened, right? It didn’t go away, you…
A. Yes, and…
Q. …were thinking about it?
Q. But, when you talked to the doctor on the 22nd of August, the first day, I take it, that when she asked that, I guess you just said, no, because you didn’t want to tell her?
A. Yes, and because I also wasn’t sure if it had occurred exactly that day, or if it didn’t.
Q. Okay. But, contact of the nature that you’ve described is a form of sexual activity, you’d agree?
(A.R., vol. V, at pp. 37-38 (emphasis added))
 At no point did the trial judge prevent defence counsel from further exploring the complainant’s definition of “virginity” or what she meant when she told her doctor she was not sexually active prior to September 2nd.
 During cross-examination, the complainant was also questioned about when she met her boyfriend. She testified that she “met up with him, I think, after July, and we officially started dating on December 26th”. Defence counsel challenged this testimony, suggesting she was “hanging out with him in July and August”. The complainant responded, “[a]t the end of July, beginning of August” (A.R., vol. V, at pp. 15-17). Defence counsel chose not to press her further on this point.
 In short, nothing in the record, apart from speculation, suggests that the 15-year-old complainant was sexually active or even had a romantic partner at any time relevant to challenging the pregnancy evidence, despite cross-examination on both issues. As such, I am persuaded that the scope of permissible cross-examination would not have been any broader than the questioning that actually occurred.
 Second, the s. 276 ruling did not prevent defence counsel from advancing the theory that the complainant lied to protect her relationship with her boyfriend. Counsel asked the complainant to confirm that she told the police in September 2013 that she had a boyfriend she “really, really” liked, which she did. Counsel also asked whether the complainant’s mother had threatened to kick her out of the house if she ever became pregnant — the complainant acknowledged that she had. Counsel further suggested that the complainant wanted to hide the assault and pregnancy so she did not “look bad” to her family — the complainant agreed she had thought about this.
 I acknowledge that this cross-examination may have been less effective because counsel could not ask the final question: “I put to you that it’s actually your boyfriend that you were having sex with in July”. However, the implication of the questions asked at trial was clear and the trial judge could have considered the possibility of a motive to lie. Indeed, before this Court, counsel for R.V. candidly admitted that nothing prevented the defence from further probing the complainant’s testimony about when she began seeing her boyfriend or her motive to lie. Accordingly, it cannot be said that any error with respect to s. 276 prevented the accused from making this argument.
 For these reasons, I conclude that neither the application judge’s interpretive error with respect to s. 276(2)(a) nor the trial judge’s conclusion that he was bound by the previous s. 276 ruling prevented R.V. from making full answer and defence at trial. On the facts of this case, the errors are harmless and there is no reasonable possibility that the verdict would have been different had the errors not been made. R.V. suffered no substantial wrong as a result of these errors.
 I would allow the appeal and restore R.V.’s conviction.
The following are the reasons delivered by
Brown And Rowe JJ. -
 This is the third appeal decided in recent weeks, along with R. v. Barton, 2019 SCC 33, and R. v. Goldfinch, 2019 SCC 38, where the Court is called upon to interpret the purpose, scope, and application of s. 276 of the Criminal Code, R.S.C. 1985, c. C-46. Sexual offence trials are unique among criminal trials in Canada (A.F., at p. 1). Evidence of a complainant’s sexual history is inadmissible where it is tendered by the accused, unless and until the accused meets the admissibility criteria set out in s. 276(2).
 In Barton and Goldfinch, the Court explained how the admissibility criteria operates where the accused seeks to lead evidence of a complainant’s sexual history at trial. This appeal, however, turns on how to apply the admissibility criteria where the Crown leads evidence of a complainant’s sexual history, and the accused seeks to challenge that evidence through relevant cross-examination. We stress that, in this appeal, the Crown was not merely introducing evidence of the complainant’s sexual history for the purposes of providing “background”, or as the necessary link to explain a series of events, but as a central plank of the case incriminating R.V.
 The Crown’s position is that the admissibility criteria remains the same for the accused, regardless of who elicits the evidence. The application judge adopted the Crown’s position. In her s. 276 ruling, the application judge found that R.V. could not pursue certain lines of inquiry because he did not know, in advance of cross-examination, what the complainant’s evidence would be. R.V. could not, therefore, establish “specific instances of sexual activity” as she interpreted s. 276(2). His proposed inquiries were nothing more than a “fishing expedition” into the complainant’s sexual history, and the probative value was “speculative” at best (A.R., vol. I, at pp. 15-16). At trial, the trial judge held that he lacked the jurisdiction under s. 669.2 of the Criminal Code to revisit the application judge’s ruling.
 We are all agreed that the application judge misapplied the admissibility criteria under s. 276 and further, that the trial judge erred in holding that he had no jurisdiction to reconsider the s. 276 ruling in light of the evidence adduced by the Crown. What divides us is not the proper application of the admissibility criteria under s. 276(2) in these circumstances, where we are in agreement with the majority, but the appropriate remedy for the errors of the application and trial judges.
 The majority finds that, notwithstanding the erroneous s. 276 ruling, any impact on trial fairness was of no moment because R.V. managed to effectively conduct the cross-examination that the ruling restricted. While the application judge adopted an “overly restrictive approach” to cross-examination, the majority would not order a new trial since “the scope of permissible cross-examination would not have been any broader than the questioning that actually occurred” and “nothing in the record, apart from speculation, suggests that the 15-year-old complainant was sexually active or even had a romantic partner at any time relevant to challenging the pregnancy evidence, despite cross-examination on both issues” (paras. 83 and 96). The majority therefore invokes the curative proviso (s. 686(1)(b)(iii) of the Criminal Code) to restore R.V.’s conviction.
 We cannot agree. The errors in this case were not “harmless” or “minor”. Nor was the evidence overwhelming. Indeed, one may get the impression after reading the majority’s reasons that all R.V. was deprived of was the ability to ask two additional interrogatories (para. 88). But this is not so. While the immediate effect of the ruling was to prohibit R.V. from cross-examining the complainant on the legitimate theory that a different person was the cause of the complainant’s pregnancy, R.V. was deprived of much more than simply the opportunity to ask two additional interrogatories, or hear certain responses. He was denied an entire process of questioning. This had reverberating effects on all aspects of his defence that renders the majority’s parsing of the transcript highly unpersuasive. Quite simply, as we cannot know what a witness will say in cross-examination, none of us can know the nature of the evidence R.V. may have elicited but for the erroneous s. 276 ruling. And that is what divides us.
I. The Curative Proviso
 We are mindful of the following general principles concerning the proviso. It is the Crown who must show that a conviction should stand despite a finding that there was an error of law. And as this Court has repeatedly held, the proviso can only be invoked where there is no reasonable possibility that the verdict would have been different had the error not been made (see R. v. Sekhon, 2014 SCC 15,  1 S.C.R. 272; R. v. Sarrazin, 2011 SCC 54,  3 S.C.R. 505, at para. 24; R. v. Van, 2009 SCC 22,  1 S.C.R. 716, at paras. 34-36).
 Unsurprisingly, then, the curative proviso is rarely (successfully) invoked, and applies where, and only where, the error is minor or harmless, or the evidence is overwhelming. These are, moreover, separate and distinct preconditions to the application of the proviso; “the seriousness of a trial judge’s error(s) is not balanced against the strength of the Crown’s case” (R. v. Brown, 2018 ONCA 481, 361 C.C.C. (3d) 510, at para. 77, citing Sarrazin, at paras. 22-28). In either circumstance, it is a high bar for the Crown to meet. We have repeatedly refused to lower the bar, even upon express invitation to do so, as was before us in Sarrazin.
 Here, no party argues that the evidence is overwhelming. The application of the proviso therefore turns on whether the erroneous s. 276 ruling was so minor or harmless that it could not have had an impact on the verdict.
 It is often difficult for an appellate court to speculate about how cross-examination might have affected the fact-finding process, had it not been unduly restricted. This is especially so where the Crown relies on circumstantial evidence (see LeBel J. dissenting, but not on this point, in Sekhon, at para. 88), as it did here with the evidence of pregnancy. And where (as here) a case involves not a single error at first instance, but multiple errors, the cumulative impact is assessed to determine whether such errors can truly be said to be minor or harmless (see Brown, at para. 75, citing R. v. Bomberry, 2010 ONCA 542, 267 O.A.C. 235, at para. 79; R. v. Hill, 2015 ONCA 616, 339 O.A.C. 90, at para. 102).
 The high bar on the proviso’s use strongly affirms the need to safeguard the integrity of the criminal justice system from the risk of wrongful conviction. This risk looms large where there has been improper interference with the right of cross-examination, as the right to test the Crown’s evidence through relevant cross-examination is guaranteed by both the common law and the Canadian Charter of Rights and Freedoms as a core element of the right to make full answer and defence(see R. v. Osolin,  4 S.C.R. 595, at pp. 611-12; R. v. Levogiannis,  4 S.C.R. 475; R. v. N.S., 2012 SCC 72,  3 S.C.R. 726, at para. 24; R. v. Schmaltz, 2015 ABCA 4, 593 A.R. 76, at para. 18; Regina v. White(1976), 1 Alta. L.R. (2d) 292, at p. 299). If an accused’s right to test the Crown’s evidence is irremediably impaired through an inability to challenge a crucial part of the case against them, it will be inappropriate to invoke or apply the proviso (R. v. Lyttle, 2004 SCC 5,  1 S.C.R. 193, at paras. 69-70, citing R. v. Anandmalik (1984), 6 O.A.C. 143, at p. 144; R. v. Wallick (1990), 69 Man. R. (2d) 310, at p. 311; see also R. v. Borden, 2017 NSCA 45, 349 C.C.C. (3d) 162, at para. 215).
II. The Importance of Cross-Examination to a Fair Trial
 As a general rule, an accused has the right of cross-examination in the fullest and widest sense of the word as long as that right is not abused (see Lyttle, at paras. 44 and 70). And we stress that this is not a case about an abusive cross-examination. While we share the majority’s concern that this area of cross-examination can fall into abuse if it is not tightly controlled by all actors in the courtroom (see also Binnie J. in R. v. Shearing, 2002 SCC 58,  3 S.C.R. 33, at paras. 121-22), here we all agree with Paciocco J.A.: R.V.’s proposed cross-examination of the complainant as to the cause of the pregnancy pertained to evidence that was relevant and had significant probative value that was not outweighed by substantial prejudice under s. 276 (2018 ONCA 547, 141 O.R. (3d) 396).
 We turn first to the majority’s treatment of the cross-examination that actually did occur in this case. We disagree entirely as to whether meaningful cross-examination occurred, so as to support the conclusion that nothing would have changed had R.V. been able to pursue the inquiries the s. 276 ruling restricted. This appears to resurrect the very error that we all agree the application judge made (majority reasons, at para. 62, adopting Court of Appeal reasons, at para. 64): equating the probative value of cross-examination with its probability of success in eliciting the evidence sought (or equating the right of cross-examination with “the certainty of result”). But this view of cross-examination reduces it to nothing more than a checklist of topics, where the relevance of questioning is contingent on the examiner’s ability to know, in advance, the responses of the opposing witness. And this impoverished view of cross-examination has long ago been rejected by this Court.
 We therefore do not accept that R.V. was denied merely the ability to ask two additional questions, or hear two further responses, as the majority suggests. Rather, he was denied an entire process of questioning. This denial impacted all aspects of his defence at trial in ways that cannot be accounted for by the majority’s strained parsing of the transcript. As we will explain, cross-examination is an organic process — each answer potentially affecting the cross-examiner’s bearing — rather than a (mechanistic) checklist of topics, and, as such, none of us can know whether a properly designed cross-examination culminating in a series of questions (which, again, we all agree should have been permitted) would have elicited evidence favourable to R.V.’s case. Cross-examination, after all, is not canvassed in torpor.
 Of even more fundamental concern to us is that the majority’s reasoning is inconsistent with how this Court has (traditionally) regarded the effects of improper interference with the right of cross-examination on the fairness of a criminal trial. The leading decision of this Court is Lyttle. Although distinguishable on its facts and not a case involving the application of s. 276, the legal issue in Lyttle — the effect an improperly constrained cross-examination has on the fairness of an accused’s trial — bears striking similarity to the case at bar. We therefore recount the reasoning in that case at some length, focusing on two of its foremost principles that bear on the question of whether to invoke the curative proviso in this appeal: (1) a manifestly constrained cross-examination has a prejudicial impact on an accused’s ability to control his or her defence and on the fairness of their trial; and (2) an accused’s fair trial rights include not just the fact of cross-examination, but also control over the rhythm of cross-examination.
 In Lyttle, the victim was found severely beaten. He told police that this was the result of a dispute over a gold chain. Police did not believe him, speculating that he was more likely beaten up as a result of a drug debt gone bad. But the victim denied this, and police came around to his account that it occurred over a gold chain. At trial, Mr. Lyttle’s counsel was prohibited from putting suggestions to Crown witnesses that the beating occurred over the initial police theory, a drug debt gone bad, which would have been exculpatory evidence favouring Mr. Lyttle, unless counsel intended to prove the factual basis for the suggestion. She was unable to do so, absent cross-examination of Crown witnesses, and Mr. Lyttle was convicted in the beating of the victim.
 Before this Court, the legal issue was whether an accused person must have a good faith basis or, a stricter, evidentiary foundation, to cross-examine a Crown witness on a particular matter at issue. In other words, is an accused person limited to questioning witnesses on matters that can be confirmed through other means? This Court answered in the negative. It was found that a question can be put to a witness in cross-examination regarding matters that need not be proved independently, provided that the examiner has a good faith basis for putting the question to the witness. It was observed that it is not uncommon for an examiner to believe what is in fact true, without being able to prove it other than by cross-examination.The risk is that if the examiner gets a denial or some answer that does not suit them, the answer stands against them in evidence (Fox v. General Medical Council,  1 W.L.R. 1017, at p. 1023, per Lord Radcliffe, cited in Lyttle, at para. 49).
 This Court went on to conclude that the trial judge had improperly interfered with Mr. Lyttle’s right to cross-examine the Crown witnesses, having placed unwarranted conditions on legitimate lines of questioning. On whether to invoke the curative proviso, the Court had little difficulty finding that a substantial wrong occurred and an unfair trial resulted, even though nothing on the record suggested that the beating occurred over a drug debt gone bad. Further, the evidence against Mr. Lyttle was compelling: the victim had identified Mr. Lyttle as his “unmasked” attacker in a photographic line-up. Nevertheless, the Court chose not to invoke the proviso, for reasons — and it is this point that the majority does not answer — that a manifestly constrained cross-examination strikes at the heart of the accused’s ability to control his or her defence and on the fairness of their trial, and that, where fair trial rights have been infringed, the appeal should run its natural course.
 We pull from Lyttle that:
Cross-examination may often be futile and sometimes prove fatal, but it remains nonetheless a faithful friend in the pursuit of justice and an indispensable ally in the search for truth. At times, there will be no other way to expose falsehood, to rectify error, to correct distortion or to elicit vital information that would otherwise remain forever concealed.
That is why the right of an accused to cross-examine witnesses for the prosecution — without significant and unwarranted constraint — is an essential component of the right to make full answer and defence. [Emphasis in original; paras 1-2]
Put simply, meaningful cross-examination cannot be undertaken if its scope is manifestly constrained.
 But Lyttle also draws attention to the rhythm of cross-examination as an essential aspect of an accused’s fair trial rights, and not just the fact of cross-examination. Again, cross-examination is not so much a seriesofquestions as a processofquestioning. Recall that in Lyttle, the same argument was made by the Court of Appeal for Ontario as the majority makes here: while the trial judge wrongly interfered with the cross-examination of Crown witnesses, counsel effectively got to ask the questions anyway. But this line of reasoning was rejected at this Court where it was noted that, regardless of the fact that the evidence prohibited through cross-examination was effectively canvassed in direct examination of the witnesses, the ruling had an intimidating effect on defence counsel, disrupted the rhythm of her cross-examination, and placed conditions on a legitimate line of questioning (paras. 3 and 71). Further, unduly restricting Mr. Lyttle’s cross-examination had downstream effects on the fairness of the trial, as Mr. Lyttle had to call Crown witnesses as his own, and therefore gave up his right to address the jury last.
 The Court’s reasoning reflects the fact that cross-examination is an organic process that cannot be considered in isolated pieces. It consists of more than a single question or series of questions. Indeed, it typically consists of a process of questioning in which skilled counsel seek to elicit things that are not immediately apparent that, within strict bounds, tests the credibility of a witness. The rhythm of cross-examination involves putting careful questions to a witness that are designed to explore bit by bit the nature and extent of that witness’s knowledge. Cross-examination is effective only where it is permitted to proceed step by step towards the ultimate point, where the cross-examiner can pose the final question (or questions), knowing by that time what the answer(s) will be, having regard to the earlier evidence elicited (see G.D.E. Adair, On Trial: Advocacy Skills and Practice (2nd ed. 2004), at p. 333). This means that, when cross-examination is unduly constrained, the effects on the fairness of the trial will often reverberate beyond, and cannot be fully appreciated by parsing, the particular words in a transcript.
 In other words, effective cross-examination does not depend merely on what is asked, but also how it is asked. To repeat, in pursuing cross-examination, skilled counsel engage the witness in a process that culminates in a final inquiry, where the answers to all of the previous inquiries, taken together, reveal the examiner’s theory or point. It is only after all of the evidence has been elicited following cross-examination that facts can be found and credibility assessments can be made. It flows logically that where the examiner knows from the outset that a particular line of inquiry cannot be explored, this will necessarily have an impact on the rhythm of the cross-examination, and the examiner is likely to adjust the main theory or point of their cross-examination as a result. That was likely the case here.
 It is for this reason that cross-examination cannot be unduly restricted. It is not sufficient for the examiner to get 90 percent of the way through a line of questioning. That 90 percent may have no purpose or relevance at all if the examiner cannot get to the final 10 percent. Relatedly, only being able to ask a single question, or a final 10 percent, without being able to first lay out the foundation, cannot be said to fulfill an accused’s constitutionally protected right of cross-examination either. Safeguarding sufficient control over the rhythm of cross-examination ensures we do not imprison a person in his or her privileges and call it the Charter (we borrow from the language of Frankfurter J., in Adams v. United States ex rel. McCann, 317 U.S. 269 (1942), at p. 280).
 But we wish to make clear that a cross-examination that is not unduly restricted does not mean a cross-examination that is boundless, and nothing in Lyttle or in our s. 276 jurisprudence suggests otherwise. Lyttle makes clear that the examiner must have a good faith basis for putting questions to a Crown witness and, as the majority notes, Parliament has specifically legislated limits on questioning related to a complainant’s sexual history into the Criminal Code. While, therefore, trial judges always ensure that counsel “[are] not merely taking a random shot at reputation imprudently exposed or asking a groundless question to waft an unwarranted innuendo into the jury box” (Lyttle, at para. 51, citing from Michelson v. United States, 335 U.S. 469 (1948), at p. 481), cross-examination is even more narrowly constrained by the evidentiary corral built by the admissibility criteria in s. 276(2); inferences that, by operation of s. 276(1), will never be relevant; and by the common law principles enunciated in R. v. Seaboyer,  2 S.C.R. 577. When an examiner exceeds those bounds created by s. 276 or the common law, the examiner is no longer asking questions in good faith, and therefore any evidence elicited as a result of those inquiries is inadmissible and should be immediately disregarded.
 But in the case at bar, we repeat: the proposed inquiries would not have contravened s. 276 or the common law principles from Seaboyer and there was a good faith basis for putting the questions to the complainant. While, therefore, we agree that, “the circumstances of this case warranted tightly-controlled questioning of the complainant” (majority’s reasons, at para. 88), our point of disagreement is this: the questioning that did occur at R.V.’s trial was not a fair substitute for what the erroneous ruling restricted. The majority’s whittling down of any error to a deprivation of two interrogatories represents in our respectful view an unfortunate obfuscation of this Court’s previously clear emphasis of the importance of an accused person exercising sufficient control and rhythm over the entire cross-examination strategy, from the lines of inquiry chosen to the order in which questions are asked. All of this was cast aside in R.V.’s trial when he was prevented from putting to the complainant a series of questions that we all agree should have been permitted.
III. The Restriction on Cross-examination Was Not a Minor or Harmless Error
 The foregoing explains why we cannot accept that there was meaningful cross-examination so as to support the majority’s conclusion that there is no reasonable possibility that the verdict could have been different had the errors in this case not been made. It follows that the error in the s. 276 ruling cannot be said to be “minor” or “harmless” when, as submitted by the Criminal Lawyers’ Association of Ontario, at its core it allowed the Crown to claim for itself the right to adduce evidence of a pregnancy as incriminatory of the accused, while insisting that the accused should be barred by the language of s. 276(2) from challenging the very evidence relied on by the Crown (I.F., Criminal Lawyers’ Association of Ontario, at p. 7).
 It does not follow that R.V. coming close to what he would have been allowed to ask following a proper s. 276 ruling means that the error did not meaningfully impact his actual trial. Such an application can only water down the proviso, because it strains common sense for an appellate court to conclude that it knows with certainty what evidence would have been adduced, or what effects such evidence would have had on the trial judge’s assessment of the witnesses, had no such errors been made at trial. None of us can confidently say whether, had R.V. been able to effectively challenge the cause of the pregnancy, he would have elicited something that was sufficient to raise a doubt. This illustrates the importance of cross-examination: it is sometimes the only way to elicit evidence that is often not apparent at the outset.
 The erroneous s. 276 ruling had ricochet effects. As Paciocco J.A. noted (C.A. reasons, at para. 122), the trial judge’s error in believing he could not re-visit the s. 276 ruling rendered future meritorious applications to reconsider futile, which likely would have had an intimidating effect on defence counsel’s ability to respond effectively (see Lyttle, at para. 7) as evidence was elicited during the trial that required a reconsideration of the ruling. We accept Paciocco J.A.’s assessment (at para. 110) that counsel was functionally powerless to do anything about it. We note that, invariably, R.V. would have pursued a different trial strategy had he been permitted to directly challenge the evidence led by the Crown. For instance, R.V.’s decision to testify in this case would likely have been impacted by the ruling. Had R.V. been able to pursue the inquiries prohibited by the ruling, he may not have testified at all. There are simply too many variables that flowed from the erroneous s. 276 ruling for an appellate court to comfortably invoke the proviso.
 Overall, the majority’s parsing of the transcript, and its reduction of any errors to a mere limit on R.V.’s ability to ask two particular interrogatories, provide little comfort. The majority effectively treats R.V.’s trial as if it was error-free, which it certainly was not. The application judge misapplied all three statutory criteria under s. 276(2). And to repeat, her ruling deprived R.V. of his right to proceed step by step through the process of cross-examination leading to the final question (or questions).
 The majority sees no harm in any of this, since other lines of questioning could have been pursued (para. 92). But, and with respect, the fruitfulness of such other lines of questioning was contingent on the possibility that those lines of questioning might culminate in questions that could undermine the probative value of the pregnancy. And the application judge’s erroneous s. 276 ruling foreclosed this possibility, meaning that R.V.’s cross-examination strategy would still have had to be adjusted accordingly. The point is that, but for the application judge’s errors, R.V.’s entire cross-examination may well have been different.
IV. The Relevance of the Evidentiary Record to the Application of the Proviso
 This brings us to the majority’s treatment of the evidentiary record, where the majority says, notwithstanding the restriction on cross-examination, we do know what evidence R.V. would have elicited, based on evidence elicited when trial counsel “skirt[ed] the [s. 276] ruling” (C.A. reasons, at para. 91), and suggested to the complainant that she was not a virgin at the time of the alleged assault. We note parenthetically, as Paciocco J.A. observed, that trial counsel was restricted from probing her answers or engaging in any other interrogatories based on her responses. Nonetheless, the majority would apply the proviso because, in its view, this shows that there is no reasonable possibility that the verdict would have been different but for the application judge’s errors.
 We have several objections to this approach.
 First, we question the majority’s choice to invoke the proviso on the basis of evidence that was inadmissible in R.V.’s trial (given the erroneous s. 276 ruling) and ought to have been immediately disregarded by the trial judge. It is trite to observe that where the law clearly prohibits a line of cross-examination, lawyers should understand that they are not to cross-examine a witness in that area of evidence (see E. Craig, “The Ethical Obligations of Defence Counsel in Sexual Assault Cases” (2014), 51 Osgoode Hall L.J., at p. 456, fn. 147). The majority’s treatment of this evidence is particularly troubling in the context of a sexual assault trial which, as it rightly notes, presents “unique challenges” to the administration of justice and requires that questioning of complainants be “tightly-controlled” (paras. 1 and 88). Indeed, the Court’s posture towards counsel “skirting” rulings that they disagree with in a similar fashion should be one of discouragement, not encouragement. Our concern is with the practical effect of relying on evidence adduced in contravention of the s. 276 ruling, with no admonition of counsel for “skirting” in the first place. Using evidence obtained in contravention of a s. 276 ruling to support the invocation of the proviso may well have the unintended consequence of encouraging bad behaviour. It may signal to counsel that there are incentives to playing fast and loose with rulings they disagree with, since the evidence elicited thereby may be used by appellate courts in their favour. The result is, in this case, particularly troubling, as Paciocco J.A. noted (C.A. reasons, at para. 91), since R.V.’s counsel had no means to probe the complainant’s answers and never actually asked the series of questions all of us agree should have been permitted.
 Secondly, we note that the majority’s reliance on the record appears to resurrect an approach to the curative proviso that this Court rejected in Sarrazin. There, the Court considered whether to adopt an “holistic approach” to the proviso, according to which it could be applied where “an appellate court is satisfied that the evidence of guilt is very strong, although not quite overwhelming, and the legal error or errors, though not insignificant, are highly unlikely to have affected the result” (R. v. Sarrazin, 2010 ONCA 577, 268 O.A.C. 200, at fn. 113 (emphasis in original), cited by this Court in Sarrazin at para. 16). But such an approach was rejected as a watering down of the proviso. The strength of the Crown’s case cannot counterbalance the effect of an error that is neither minor nor harmless. Put another way, unless the evidence against the accused is overwhelming, its strength is effectively irrelevant to the application of the proviso. Appellate courts may therefore not straddle the branches of the proviso, which the majority’s conclusion appears to do. Rather, the harmless error branch focuses on whether the appellate court can confidently conclude that the legal error did not impact the verdict. In our view, this certainty is unavailable.
 Thirdly, the misapplication of s. 276 was compounded by the trial judge’s reliance on the pregnancy as corroborative in finding that R.V. was guilty. This was in error because the trial judge essentially convicted R.V. on evidence that he was unable to challenge. This is distinct from the trial judge’s credibility findings, with which we take no issue: the complainant was credible, and R.V. was not. But the probative value of the pregnancy cannot be disentangled from the credibility assessments. For example, had R.V. been able to pursue the cross-examination that we all agree he was entitled to pursue, he may have chosen not to testify at all. In that case, the trial judge would not have been able to make any negative findings about R.V.’s credibility, irrespective of whether R.V.’s cross-examination elicited evidence helpful to his defence. Again, there are simply too many variables that flowed from the erroneous s. 276 ruling for an appellate court to comfortably invoke the proviso on this evidentiary record.
 Finally, and moving forward, the majority’s approach to the application of the proviso appears to turn the proviso on its head. Its reasons suggest that, in the face of a legal error, a new trial should only be ordered if the appellate court can be certain that the trial would have gone differently. However, where there has been a legal error, the default is to order a new trial; the proviso allows a departure from this default rule only in very narrow circumstances. Indeed, a reviewing court must order a new trial unless it is satisfied that the result could not have been different absent the legal error.
 Given the interests the proviso protects, we cannot invoke it here. The errors were neither minor nor harmless. They were significant. The cumulative effect deprived R.V. not so much of the right to ask particular questions, or hear particular responses, as the majority finds, but to engage in a process of questioning, protected by both the Charter and the common law.
 The right to present one’s case cannot, constitutionally, be curtailed in the absence of an assurance that the curtailment is clearly justified by even stronger contrary considerations (see Seaboyer, at pp. 620-21). This takes on even greater importance when the evidence of the complainant and accused person are diametrically opposed in every material respect, leaving credibility as the central issue at trial (see R. v. Crosby,  2 S.C.R. 912, at para. 12). The truth is that none of us know what evidence R.V. may have elicited had a proper s. 276 ruling been made in advance of trial.
 This is, after all, what divides us. Cross-examination, being the primary vehicle through which R.V. could make full answer and defence, was restricted in a manner not consistent with the purpose behind s. 276. As a result, R.V. was denied a fair trial. We would dismiss the appeal and affirm the Court of Appeal’s order that a new trial be directed.
 The September 18th ultrasound indicated that the complainant had been pregnant for 13 weeks, 5 days. At the voir dire, this was used to advance a conception date of June 14th. However, the doctor explained that this estimate referred not to the date of conception but to the first day of the last menstrual period, with conception occurring roughly 14 days later, plus or minus 5 to 7 days.
 On December 13, 2018, Bill C-51, An Act to amend the Criminal Code and the Department of Justice Act and to make consequential amendments to another Act, 1st Sess., 42nd Parl., received royal assent. This Act makes minor amendments to clarify the application of s. 276. Although they do not apply to this case, they are not inconsistent with the analysis set out here.
 Now s. 278.93(2).
 Now s. 278.93(4).
 We note that on December 13, 2018, s. 276(2) was amended to include a fourth admissibility criteria: evidence of a complainant’s sexual history cannot be adduced by the accused for the purposes of drawing on inferences prohibited by s. 276(1). Moving forward, judges considering an application by the accused to adduce sexual history evidence must consider whether the accused has met all four criteria for admissibility under s. 276(2).
 A. Woolley, Understanding Lawyers’ Ethics in Canada (2nd ed. 2016), at p. 403: “[a] lawyer should cross-examine witnesses within the rules established by the law of evidence” (emphasis added); Craig, at p. 430: “Defence counsel are ethically obligated to restrict their carriage of a sexual assault case (including the evidence they seek to admit, the lines of examination and cross-examination they pursue, and the closing arguments they submit) to conduct that supports findings of facts within the bounds of law” (emphasis added)