Rationale for the rule excluding similar fact evidence
The underlying rationale for the rule excluding similar fact evidence is that to allow it in every instance is to risk the conviction of an accused not on the evidence relating to the facts but because of past behaviors or disposition towards crime  . Such evidence without doubt has a prejudicial effect against the accused. At the root of the exclusionary rule is the risk of prejudice to the accused – the concern is that the trier of fact would be unduly influenced by the fact that the accused has been guilty of similar crimes in the past and thereby succumb to the deductive reasoning that the accused has greater propensity to commit the present offence
However, at times, similar facts can be so probative of guilt that to ignore it via the imposition of a blanket rule would hamper the search for truth.  The basic dilemma confronting the reception of similar fact evidence is that while it constitutes extremely probative evidence, it is also thought to be capable of engendering substantial unfavorable sentiment against the accused. The development of law concerning similar fact evidence has been focused on these 2 main considerations.
The approach adopted by the EA is, simply, to face prejudicial evidence head-on at the stage of admissibility. In Singapore, similar fact evidence is principally dealt with under Section 14 and 15 of the EA. S14 renders admissible similar fact evidence where it is relevant to the state of mind, which is an issue.  Section 15 makes admissible similar fact evidence (which is part of a series of similar occurrences) or with a certain intention or knowledge, where that is in issue.  The local courts have applied the Boardman balancing test within the framework of the EA where similar fact evidence is sought to be adduced to prove one of the matters identified in Sections 14 and 15. 
However, as it will be established later, the prejudicial effect of evidence is simply not eliminated by not admitting it. In fact, once it has been presented to the judge, it is admitted into the mind of the judge. The prejudicial effect remains, even if the evidence is not admitted physically.Further, in focusing on admissibility, the need for safeguard after the evidence has been admitted is overlooked. 
The most effective way of reforming this area of law would be to shift the focus away, from the admissibility stage to implementing measures for counteracting prejudice once evidences had been admitted.  This will have the effect of subjecting judges to less pressure and they will not have to seek refuge in special rules.  However, since the focus of this paper is to propose reforms in the area of similar fact within the EA, implementation of any other measures to counter-act prejudice would have to be the subject of another paper.
As such, this paper will proceed with the aim of perfecting the similar fact rule envisioned within Sections 14 and 15 of the Evidence Act, which will allow the exclusionary rule to perform the function that it was meant to.
There is a very close nexus between the prohibition on bad character evidence and the presumption of innocence, essentially since the prohibition aims to ensure that the accused is not pre-judged by evidence of his past behavior or disposition.
Obviously, the focus of this paper is on evidence, which reveals past similar misconduct or offence and thus proving the disposition or propensity to conduct himself in a particular manner so as to prove the present charge(s) against him (“propensity evidence"). Propensity reasoning refers to the line of reasoning that goes from the premise that an accused has committed such an offence before or has a propensity to commit crime or is otherwise of bad character, to the conclusion that he is likely to have committed or has committed the offence for which he is charged (“propensity reasoning").
Propensity evidence is now generally understood as a category of evidence in which “similar fact evidence"," background evidence" and “identity evidence" are subcategories  .
Why guard against evidence of prior acts?
Prejudicial effect on the trier of fact, conclusion of guilt will be unfairly reached
Accused ought to be tried only for his present crimes and not for his past mistakes.
Section 14 and 15 fails to exclude the prejudicial effect of an inadmissible evidence
Sections 14 and 15,which were drafted to reflect the similar fact rule, fails to carry out the very purpose it was created for. It was meant to simply exclude evidence of prior acts because of the prejudicial effect. [find out stephen’s real intention in including the similar fact provision].
A learned academic has criticized the approach  that the judge should hear the similar fact evidence and determine if it ought to be admitted.  It was argued that this approach is improper in a non-jury context, as it assumes that the evidence does not prejudice the judge  . According to this assumption only persons “who have not been trained to think judicially" yield to prejudice.  Simply put, a judge who is prejudiced by the evidence, is expected to give a fair verdict as he has instructed himself to exclude it and perform the very kind of mental gymnastics which Lord Cross had warned against in Boardman. 
Is it a far-fetched claim to make, when one says that judges are still human? Does a legal training and vast legal experience, help in fortifying the mental resilience of a judge and thus making him immune to influence by the prejudicial element of the evidence?
It is argued that judges, in common with most other human beings, are susceptible to moral influence, though their threshold of susceptibility may differ. Conversely, most citizens share, with the judiciary, the moral sentiment that punishment must be confined to the transgression in question and the belief that proof beyond reasonable doubt is the morally acceptable risk of mistake in punishment.
Even in a scenario, where the evidence has been admitted because of it has some other form of relevancy, aside from propensity reasoning, there remains the danger that the judge will nevertheless go down such a path. How do we protect the accused from such pitfalls and yet ensure that the principles of our criminal justice system remain intact?
IMPOSE A PROCEDURAL DUTY via THE EVIDENCE ACT
It is suggested that the Evidence Act be modified to impose a procedural duty onto a trial judge: to cast aside prejudice and ensuring that he consciously alerts himself against bias when he assesses and deals with similar fact evidence, which are highly probative (yet with a certain prejudicial element). His awareness of risk ought to be expressed in writing, say, in his notes of evidence or in his grounds of judgment. In his judgment, it ought to specify precisely what part of the similar fact evidence has played in his line of reasoning and in what way it has contributed to his findings in both the intermediate and the ultimate issues.
Such an approach would compel a judge to
Recognize the lingering effect of excluded propensity evidences and
While also acknowledging the prejudicial effect of highly probative evidence, which have been admitted. As such, this deals with prejudice, not merely at the admissibility stage but also throughout the trial, by imposing a judicial duty to avoid reasoning pitfall in the process of his findings of fact.
Evidence Act allows for the circumvention of the similar fact rule by allowing admission of highly prejudicial evidence via other relevancy provisions
Similar facts of a type declared to be relevant by any of the relevancy sections are admissible notwithstanding that they may be relevant via propensity reasoning. Evidence of prior acts is generally admissible under the EA via provisions such as S 6 (res gestae) or S9 (establishing a relationship). As Cross and Tapper points out  , this approach is potentially dangerous because it can be used to smuggle in similar fact evidence which would otherwise be inadmissible
[add info on res gestae]
Below are a few examples of common law cases wherein evidence of prior acts were admitted and it is argued that the evidences are admitted under the general relevancy provisions in the Evidence Act.
Admissibility under the EA
R v Ball
Evidence of a brother and sister’ previous sexual intercourse was admitted to prove a separate charge of incest
Ceylon Court of Appeal opined that evidence of the was admissible under S 9 of the EA to prove the relations (existence of a sexual relation) between the defendants.
R v Straffen
Defendant was charged with the strangulation of a girl. Confessions that he had previously strangled 2 other girls were admitted. All 3 murders had similarities
Arguably, admissible under S 9 of the EA to prove the murderer’s identity, by reasoning that since the same person killed girls A, B and C and X killed A and B, thus X killed C too.
Background evidence 
tThe Queen (1989) 167 CLR 590 at p. n strictly followed by our courts.f a fact.against the prejudicial effect of sch evidences is admitted because it is “so closely entwined and involved with the evidence directly relating to the facts in issue that it would amount to distortion to attempt to edit it out".  It provides the trier of fact in the general picture about the characters involved in the action and the run up to the offence.  However, highly prejudicial evidence can be admitted as “background" without any adequate assessment of its prejudicial effect. 
The chief distinction between similar fact evidence and background evidence is that in the latter, it is not tendered for the purpose of revealing that the accused has on previous occasions engaged in conduct which is similar to that now alleged and hence it is not directly relevant to fact in issue.  It is merely indirectly relevant to the facts in issue; by explaining, for example, the complainant’s lack of resistance or complaint or why the accused was so brazen in his or her conduct towards the complainant. 
The danger inherent in admitting such evidences as illuminating the background of a crime was recognised by the Singapore courts in PP v Don Promphinit & Ors (CC 89/1993) that the courts applied the Boardman balancing test in admitting the background evidence.  However, our courts have never strictly followed such an approach.
To quote McHugh J:
“Great care needs to be taken … in determining whether evidence, disclosing other criminal conduct, is evidence concerning the res gestae or is merely circumstantial evidence. By applying labels…evidence which is in truth purely circumstantial improperly avoids the tests of admissibility which the modern cases expound." 
Such an approach comes into direct conflict with the proposition laid down in Karam Singh v R  2 MLJ 75, in which the court ruled that although motive was a relevant fact and admissible pursuant to s 8, the provision had to be read subject to the subsequent provisions in the Act governing the exceptions to the hearsay rule (s 32 in this case). In other words, the EA did not allow the admission of evidence, which was not admissible under a specific relevancy provision but admissible under a general relevancy provision. Moreover, since admission under the general relevancy provisions does not require the evidences to be subjected to the threshold balancing test and thus there are no safeguards against the prejudicial effect of such evidences against the accused. Such an approach by the courts defeats the very purpose of the similar fact rule and as a result S 14 and 15 are rendered otiose consequentially.
However, it would seem that the courts aren’t to blame entirely for such a flawed approach. As it will be seen later, S 14 and 15 does not allow the admission of evidences documenting prior acts that proves actus rea. Such an unnecessary restriction forces the judges to secure the admission of such evidence through other means, especially when the evidences are highly probative of a fact.
Strictly speaking, background evidence has explanatory value, not probative value. Under the present law such evidence is admitted as “background" evidence without any assessment of its potentially prejudicial effect.
Leave of court ought to be required in allowing the adduction of background evidence
Where the evidence falls outside the central facts, however, and it is suggested that leave be required before it can be admitted. It should not be exempted from the principles applicable to such prejudicial evidence merely because it falls within one of the general relevancy provisions. Leave should not be given unless the evidence has substantial explanatory value.
The court must be satisfied that, without the evidence, the court would find it impossible or difficult properly to understand other evidence in the case and that its value for understanding the case as a whole is substantial. This test applies whether the evidence relates to a defendant or anyone else.
‘Interests of Justice’ requirement
It is also suggested that an additional test ought to be satisfied, namely that the interests of justice require the evidence to be adduced notwithstanding any prejudicial potential it might have.
The only difference from the balancing test to be satisfied for similar fact evidences is that the value of the background evidence is not, strictly speaking, “probative", in that it does not directly prove anything; it explains other evidence. It is therefore not the probative value of the evidence that must be weighed against the risk of prejudice, but its value for understanding the case as a whole.
It is to be noted that the Law Reform Committee of England concerned with the reform of Evidence of Bad Character in Criminal proceedings has recommended a similar approach. 
Balancing exercise implicit within Section 14 and 15 is flawed
S 14 and 15 creates an illogical distinction between similar fact evidence used to prove MR or AR.
All evidence of prior acts, if of a type declared to be relevant, may be adduced to prove the actus reus of the offence. The admission of similar fact evidence to prove that the accused person committed the actus reus of the offence is problematic within the framework of Sections 14 and 15 of the Evidence Act which only makes reference to the state of mind of the accused.
[Reasons for distinction within the EA?]
The point is that if similar fact evidence is admissible because of it high probative value, then there is no reason why this is limited to simply MR.There is no reason to maintain an artificial distinction between similar facts, which are probative of intention, and those, which are probative of acts done by the accused.
This has however not stopped the local courts from admitting similar fact evidence for proving the actus reus of the offence where the Boardman test is satisfied. In Lee Kwang Peng, the courts found the distinction between AR and MR ‘artificial’ and admitted the evidence, proof of actus rea by virtue of s 11(b).
[why is admitting through s 11(b) bad]
Admission via s 11(b) is a wrong approach
It is admitted that the courts were correct in rejecting the distinction between adducing evidence to establish the AR and MR, however, the usage of s 11(b) to justify the admissibility of similar fact evidence proving AR has attracted much criticisms. Despite so, this approach has been followed in many cases. 
S 11(b) was never intended to cover facts arising from previous misconduct, unconnected with the facts in issue.  It was intended to be a residuary provision for facts which did not fall within any of the general relevancy provisions (Sections 6 to 10)  .Yet such admissions were allowed by the courts  despite the contrary indication by the draftsmen, Sir James Fitzjames Stephen, in his digest:
“It may possibly be argued that the effect of the second paragraph of s 11 would admit proof of facts such as these (including similar fact evidence)…This was not the intention of the section, as is shown by the elaborate provisions contained in the following part of the Chapter II (S 12-39)" 
Jeffrey Pinsler has also pointed out  that s 122(5) of the EA,which provides for cross examination on such conduct where they are admitted via sections 14 and 15, does not make reference to S 11(b).He argues that this is a strong indicator of the fact that S 11(b) was never intended to admit similar fact evidences.
The court justified this approach on its view that the words “highly probable or improbable" are representative of the balancing mechanism of probative force against prejudicial effect in Boardman. Although the court acknowledged that the use of s 11 (b) would be contrary to the scheme of the Act as conceived by the draftsman,111it declared its willingness to ignore this concern in the interest of giving effect to the common law principle.
. According to the scheme of the Act, similar fact
evidence is only admissible pursuant to ss 14 and 15 (and only then) to establish mens rea or a mental element, not actus reus. Stephen himself pointed this out when criticising a case in which hearsay evidence was admitted pursuant to s 1 l(b).112
Stephen never intended s 1 l(b) to be used as a supplementary provision to admit evidence not encompassed by the ositively formulated exceptions to the exclusionary rules in ss 14–57. The danger of using s 11 (b) in this manner is that it undermines the scope of admissibility set by those later provisions, even to the extent of rendering them redundant. Section 1 l(b) is intended to be a residuary provision for the purpose of admitting a non-relevant fact (ie, a fact not declared relevant by ss 6–10) which may be relevant when considered in conjunction with other relevant or non-relevant facts.
It is suggested that S 14 be amended to allow the admission of facts proving the actus rea of a crime. It is argued that no distinction ought to be made between probative similarities of intention, knowledge or states of mind and probative similarities of action. This would similarly remove the need to resort to s 11(b), which arguably causes much difficulty.
It is submitted that two other possible tests for the admissibility of similar fact evidence arise out of the case law, which have never been adequately distinguished. 
a. The test of irrelevance:
Under the test based on irrelevance, propensity reasoning is seen as being always forbidden. According to Lord Hailsham:
"If the inadmissible chain of reasoning is the only purpose for which the evidence is adduced as a matter of law, the evidence itself is not admissible. If there is some other relevant, probative purpose than for the forbidden type of reasoning, the evidence is admitted..."
b. The test of prejudice:
Under the test based on prejudice, however, this line of reasoning is seen as sometimes relevant and is thus admissible provided it fulfils certain conditions. This view can be found in dicta of Lord Cross [FN10]:
"The reason for this general rule is not that the law regards such evidence as inherently irrelevant but that it is believed that if it were generally admitted jurors would in many cases think that it was more relevant than it was, so that, as it is put, its prejudicial effect would outweigh its probative value (“balancing test"). Circumstances, however, may arise in which such evidence is so very relevant that to exclude it would be an affront to common sense."
Lord Cross went on to say:
"The question must always be whether the similar fact evidence taken together with the other evidence would do no more than raise or strengthen a suspicion that the accused committed the offence with which he is charged or would point so strongly to his guilt that only an ultra-cautious jury, if they accepted it as true, would acquit in the face of it."
Difference under the two tests:
Propensity reasoning under the test of relevancy is completely prohibited while under the latter (prejudice test), it is allowed if it is highly probative.
The Singaporean courts have adopted the prejudice test as obvious from case law:
“It would be subverting established jurisprudence to allow conviction based on the particular disposition of the accused – based on the contention that the accused is that type of person who would commit such an offence. On the other hand, there may be cases where the interest of justice clearly outweigh any prejudicial dangers inherent in the evidence."
The courts have accepted that according to the ordinary relevance benchmark, reasoning from bad disposition is sometimes relevant. [FN20] It deals with undue prejudice by requiring that the probative value of the evidence is sufficiently great to make it just to admit it, notwithstanding its prejudicial effect.
Approach under the prejudice based test:
Under the prejudice-based test, however, the evidence is only admissible if its probative value is greater than the prejudicial effect (“balancing test"). There can therefore logically be no further discretion to exclude the evidence on the basis that the prejudicial effect outweighs the probative value. This reflects the fact that undue prejudice is effectively dealt with by the prejudice-based test. There is no room for any discretion, which entails the weighing of probative value against prejudicial effect, because that is itself the test of admissibility.45
Contrary approach taken in Singapore case law:
The ability to carry out the balancing exercise has always been credited to be due to the availability of discretion to exclude technically admissible evidence  . In Tan Meng Jee, the courts concluded that the discretion to exclude evidence was similar to the carrying of the balancing test  . This was the basis on which the common-law’s probative value/prejudicial effect balancing test was superimposed on Sections 14 and 15 of the EA.
In a subsequent case of Law Society of Singapore v Tan Guat Neo Phyllis, the HC held that a discretion to exclude evidence on the basis of unfairness to the accused (probative < prejudice), as recognized in Cheng Swee Tiang, was not in line with the EA. As such, the courts did not have the residual discretion to exclude evidence, which was technically admissible.
It has to be noted that the implicit formulation of the balancing test in Tan Meng Jee was based on the recognition that allowing the balancing test was proper as the courts had the right to exercise their discretion for highly prejudicial evidence. Since, it has been clarified by the HC in Phyllis Tan that no such a discretion exist, it would logically mean that the carrying out of a balancing exercise could not be part of the EA too.
It has been argued that due to the nebulous nature of the key concepts (probative value and prejudicial effect), this test provides little practical guidance and in fact is so broad that it resembles ‘the exercising of a discretion rather than the application of a principle’
S 14 and 15 is unreflective of the position in common law and the imposition of Boardman over s 14 and 15 would be contrary to the rigidity of a code
A very literal reading of Sections 14 and 15 makes it clear that the draftsmen had intended for a categorization approach to be implemented.  . These categories characterize the state of the law  governing similar fact evidence at the end of the 19th century. 
This strict categorization approach has never been wholly ‘strict’  as the courts  had held that there was a judicial discretion to exclude prejudicial evidence, despite its technical admissibility. The courts in Tan Meng Jee went further to fuse the Boardman’s formulation within s 14 and 15, such that it was no longer an exercise of discretion but rather the test of admissibility.
Such a fusion of the Boardman principles onto the provisions (which adopt a strict categorization approach) has been declared to be wrong because the courts have wrongly assumed that the two principles (categorization approach vs. the balancing approach) were similar. 
To apply the Boardman principles would be to change the law in s 14 and 15 and thereby offend the established principle  that the EA ‘…cannot be construed in the light of a decision which has changed the law’.
As mentioned in Jayasena v R  AC 618 at 625, ‘the common is malleable to the extent that a code is not’. As the EA is a code, it should be a self-contained piece of legislation to be construed within its four walls and will not have the same degree of malleability which the common law has.
Reference can be made to Lord Shaw of Dunfermline in Mohamed Sydeol Ariffin v Yeo Ooi Gark  AC 575 at 581:
“the rule and principle of the Colony must be accepted as it is found in its own Evidence Ordinance, and that the acceptance of a rule or principle adopted in or derived from English law is not permissible if thereby the true and actual meaning of the statute under construction be varied, or denied effect."
The EA has been declared to be a facilitative  and non exhaustive statue which allows the application of common law rules of evidence to fill any lacuna in the legislation. The language in Section 2(2) preserves so much of the common law that is not inconsistent with the code. The challenge in every case is therefore to work out whether the common law principle sought to be applied is consistent with the code.
Reference can be made to Phyllis Tan which makes such a declaration and found that the common law rule as to the existence of a judicial discretion was at odds within the ambit of the Evidence Act, which provided that all relevant evidence is admissible unless specifically expressed to be inadmissible.
It is to be noted that while, in Lee Kwang Peng, Yong Pung How CJ(as he was then) opined that Stephen drafted the Indian Evidence Act long before the first authoritative expression of judicial opinion on this subject in Makin v. Attorney-General for New South Wales  AC57 and thus concluded that the distinction between similar facts which are probative of intention (or other states of mind) and similar facts which are probative of acts done by the accused, was inappropriate.
It can be seen that in the case of Lee Kwang Peng, the CJ was merely extending the effect of S 14 and 15 to govern the admissibility of Actus Reus, despite its express reference to the intention of the party in question. The effect was not to change the way similar evidences would be admissible (i.e. the principles encapsulated within Sections 14 and 15).While, in Tan Meng Jee, the application of the Boardman principle cannot be said to be an extension of any principle laid down in S 14 and 15.In fact,this imposition of a balancing test provided a new threshold for admissibility of similar fact evidences which was strikingly different from the current admissibility regime (which was focused on categorization approach).In so far as the import of the common law is allowed, case laws cannot be used to wholly change the meaning or the approach that has been codified.
[Add examples where case law been used to guide or supplement but never to change the regime]
Under the categorization approach as encapsulated by Sections 14 and 15,if the evidence can be subsumed under one of these enumerated categories [mental elements like intention, knowledge, good faith etc], it is ipso facto admissible. However, the Boardman formulation is one of broad principle. Admissibility is to be determined on a case-by-case basis, by balancing the two varying factors of probative value and undue prejudice. To determine the probative value, Boardman suggests that one looks not just at the issues to which the evidence is directed (as with the categorization approach here), but also at every aspect of the evidence. Boardman says that one then counter-balances the result with prejudicial effect. 
By comparison, it is nowhere provided in ss 14 and 15 that one could look at prejudicial effect of the evidence. The categorization approach under ss 14 and 15 inclines towards the admissibility of more similar fact evidence because once the prosecution manages to attach the evidence to some category. This has always justified the for the existence of a judicial discretion to exclude such technically admissible evidence.However, in light of the HC pronouncement in Phyllis Tan,it would seem that no such residual discretion exists.If an evidence is declared to be relevant by S 14 and 15, it would be admissible.
Since, the imposition of the Boardman principle, would go starkly contradictory to the admissibility regime encapsulated in S 14 and 15, it is therefore wise to amend the statue to expressly reflect the position in the common law. Moreover, the clear imposition of a balancing test within the words of the statute would steer away from the harshness of the strict words that form a categorization approach, as implicit in the Boardman balancing formulation is the freedom to look to the unfairness against the accused in adducing those evidence. This negates the need for any residual discretion to exclude evidences.
Amended provision which is framed in the manner that encapsulates the Boardman principles
Express guidelines as to when the evidence would be highly probative and higly prejudicial.This is to starve away the nebulous definitions of concepts such as probative and prejudicial.
In one stroke of the pen, the court converted the code into a source of law capable of judicial development unrestricted by concerns as to inconsistency (and, therefore, contrary to s 2(2)) and regardless of the statutory framework. 
The EA only codified the law of evidence existing at the time of its enactment; therefore, new rules of evidence can be given effect to only if they are not inconsistent with the provisions of the EA or their underlying rationale. The EA is exhaustive only to the extent that all the rules which are not saved by statute and which are inconsistent with it (the EA) are inapplicable. However, this is not an unrestricted licence to import 21st century notions of the common law into a 19th century code.
In Lee Kwang Peng, Yong Pung How CJ declared that in the future the Act should be treated as a facilitative statute rather than a code.125
As Jeffrey Pinsler has eloquently pointed out, s 14 and 15 aptly deals with a fixed categorization approach which allows admission if evidences were to fall within the words of those section while the principle in Boardman lays emphasis on the degree of probity of evidence,irrespective of the purposes for which that evidence is being adduced for.Thus,propensity evidence would be admissible under the Boardman formulation and not s 14 and 15.