The State appeals the trial court's dismissal of a paternity action, contending: (1) there was sufficient evidence to decide which of two men was the father of the minor child; (2) the court should have considered the physical similarities between the alleged fathers and the minor child; and (3) the court considered irrelevant evidence. We affirm.
*185Vickie Flieger (now White) is the mother of Greg, who was conceived sometime during August 1972. At or near the time of conception, Vickie White had intercourse with Richard Hendrickson, Dave Vooge, Tony Hendrickson, and at least two other males. At the time of conception, Vickie White was 15 years of age, Richard Hendrickson was 16 years of age, and Dave Vooge was 15 years of age. Richard and Dave are cousins, being the sons of sisters. Greg was born on May 31, 1973. The father was not identified on the child's birth certificate.
The petition for adjudication of paternity, brought by the State pursuant to the Uniform Parentage Act (RCW 26.26) and filed in August 1977, named only Richard Hendrickson as the alleged father. The petition sought to establish paternity and to obtain child support and reimbursement for public assistance provided by the State of Washington Department of Social and Health Services under the Aid to Families with Dependent Children program. After Richard Hendrickson submitted to blood tests, Tony Hendrickson and Dave Vooge, both cousins of Richard Hendrickson, were joined as additional parties. On the basis of additional blood tests, Tony Hendrickson was excluded as the father.
At trial, Vickie White testified Richard Hendrickson had made statements to her acknowledging his paternity and expressing a desire to purchase items for the child. Vickie White's mother also testified Richard Hendrickson had made similar statements to her. Richard Hendrickson admitted to having intercourse with Vickie White sometime during the summer of 1972, but denied making the above statements to Vickie White or her mother. He stated that the first time he became aware that anyone was alleging he was the father was when he was served the summons and petition.
Vickie White's mother also testified that Richard Hendrickson's mother, Elsie Hendrickson, admitted to her that Greg was her grandson. However, Elsie Hendrickson testified she never admitted her son, Richard Hendrickson, was *186the father of the child. She also stated her son never admitted to her he thought he was the father.
Dave Vooge testified he had intercourse with Vickie White several times around the time of conception. He also testified he had a conversation with Richard Hendrickson a short time later in which each of them admitted to having intercourse with Vickie White within a 3-week period.
The State called two expert witnesses, Margaret Brooks and Terry Houtz, both employees of the Baltimore Rh Typing Laboratory in Baltimore, Maryland. Margaret Brooks is the supervisor of the Red Cell Laboratory and Terry Houtz is the supervisor of the HLA Laboratory. The two experts testified to the results of several blood tests and human leukocyte antigen (HLA) tissue typing tests involving the parties in this action and Vickie White's parents.
The first test, conducted on August 17, 1978, involved Richard Hendricks.on, Vickie White, and Greg, and resulted in a determination that Richard Hendrickson could not be excluded as the father. The test did not calculate the plausibility of Richard Hendrickson's paternity.
The second test, conducted on October 18,1979, involved Dave Vooge, Vickie White, and Greg, and resulted in a determination that Dave Vooge could not be excluded as the father. The plausibility of his paternity was calculated to be 98.32 percent.1
The third test, conducted on December 13, 1979, involved Tony Hendrickson, Vickie White, and Greg, and resulted in a determination that Tony Hendrickson could be excluded as the father.
The Baltimore Lab then calculated a plausibility of *187paternity for Richard Hendrickson of 98.70 percent,2 based on the tests of August 17, 1978 and October 18, 1979. The lab later combined the data from a Minneapolis lab with its data from the earlier testing and calculated a plausibility of paternity for Dave Vooge of 99.84 percent and 99.95 percent for Richard Hendrickson.
Ms. Brooks testified that on the basis of the blood tests conducted, neither Richard Hendrickson nor Dave Vooge could be excluded as the father. This was a very unusual case; in the vast majority of cases only one man remains unexcluded. In only one of 14,000 cases with which Ms. Brooks has been involved did the two nonexcluded males have such high plausibility of paternity percentages; those males were brothers. If either of these men was the only alleged father, in the absence of any testing of the other, he would very likely be determined to be the father. Ms. Brooks admitted she could not say that a sperm from either Richard Hendrickson or Dave Vooge impregnated the egg of Vickie White to produce the child. Furthermore, this type of testing does not allow such a determination; the highest plausibility they report is 99.99 percent.
Ms. Brooks testified that in a rare case with two nonexcluded males with high plausibility percentages, it is necessary to make a direct comparison of the two males using a formula developed by Dr. Konrad Hummel of West Germany and reported in his book, Biostatistical Opinion of Parentage Based Upon the Results of Blood Group Tests (1971). She explained how the formula is used to draw a comparison between two nonexcluded males. Prior to using the formula, a logarithm value is calculated and assigned to each nonexcluded male. According to Dr. Hummel, if there is not a difference of more than plus or minus one between the two logarithm values, the results of the test have no biostatistical significance.
*188Ms. Brooks said that after using Dr. Hummel's formula, it is also necessary to arrive at a plausibility of paternity in excess of 90 percent before reaching the "predicate probable." In other words, if after using the formula, neither male has a plausibility of paternity exceeding 90 percent, Dr. Hummel, again, would question the biostatistical significance of the result.
When Ms. Brooks calculated the logarithm values for Richard Hendrickson and Dave Vooge, she found the difference between the two values to be .478, i.e., less than 1. Despite this finding, Ms. Brooks proceeded to use Dr. Hummel's formula to draw a comparison between Richard Hendrickson and Dave Vooge. Ms. Brooks calculated the plausibility of paternity for Richard Hendrickson to be 76.5 percent and 23.5 percent for Dave Vooge. Both of the values were less than the predicate probable of 90 percent required by Dr. Hummel. While Dr. Hummel would question the results of the tests, Ms. Brooks used his formula anyway because the results were used in conjunction with other information and it was another way of looking at the data. While she did not disagree with Dr. Hummel's statistical prerequisites for the use of his formula, she felt there was something to be gained from the comparison, i.e., "it is another indication that Mr. Hendrickson continues to be higher than Mr. Vooge."
Mr. Houtz is an expert in the area of HLA testing. He testified it was "practically proven" that either Richard Hendrickson or Dave Vooge was the father. Further, based on the plausibility of paternity percentages, it was more likely Richard Hendrickson was the father. He also relied on Dr. Hummel's formula to directly compare Richard Hendrickson and Dave Vooge, despite the preliminary statistical tests not being met.3 Using Dr. Hummel's formula, he corroborated Ms. Brooks' testimony that it was three *189times more likely that Richard Hendrickson was the father than Dave Vooge. Mr. Houtz admitted that when he originally did the HLA typing tests, the results indicated Dave Vooge was more likely the father than Richard Hendrickson. That test, however, was based on only one system of HLA; based upon all the testing, Mr. Houtz stated Richard Hendrickson was more likely the father than Dave Vooge.
At the request of the State, the trial was continued and a final blood test was conducted involving Vickie White's parents. The results, which were admitted into evidence, established a plausibility of paternity for Richard Hendrickson of 99.31 percent and a plausibility of paternity of 99.09 percent for Dave Vooge.
Following the second hearing, the trial judge held the State had not proved to the satisfaction of the court that either Richard Hendrickson or Dave Vooge was the father. The court expressed an uneasiness with its decision:
[N] either individual here has saw [sic] in their heart and mind to come forward and take the responsibility for this child, which is their inalienable right, you have no comfort in the matter and feel no responsibilities in the matter, then the Court must make that determination for you. It's not unique in that regard because we have some hundreds of cases pending in the prosecutor's office now where the unequivocal father is not supporting their child, and we try to do something about that.
As a consequence, the Court feels very restrained to force an unwilling father upon a child. Am I giving this child such a benefit in life with an unwilling father? I don't know. I think he'd probably be better off with only a loving, caring mother than an unwilling father dictated by .05 percent. . . .
I'm not happy with this decision. I'm not quite happy with the people that are before me, but I'm going to find that neither of the parents have been proved to the satisfaction of this Court, neither of these two individuals before the Court have been proved as satisfactory to this Court that they are solely and singly the father of this child as between the two of them. And as I understand from Miss Brooks, this is the first experience she's had in *190this category in 20 years, and I hope that I can say the same 20 years from now.
This appeal followed.
The State first contends the trial court erred in finding the State had not sustained its burden of proof showing it was more likely than not Richard Hendrickson was the father of Greg. The State argues it met its burden of proof by a preponderance of the evidence, State v. Howe, 44 Wn. App. 559, 723 P.2d 452 (1986), in that: (1) both experts testified that based on blood tests, it was more likely than not that Richard Hendrickson was the father; and (2) most of the lay witnesses testified the child more closely resembles Richard Hendrickson than Dave Vooge.
The determination of the credibility of witnesses and the weight to be given to their testimony is peculiarly within the discretion of the trial court and will not be disturbed on appeal. In re Watson, 25 Wn. App. 508, 511, 610 P.2d 367 (1979); Rognrust v. Seto, 2 Wn. App. 215, 222, 467 P.2d 204, review denied, 78 Wn.2d 993 (1970). The trial court may refuse to accept uncontradicted expert testimony as long as it does not act in an arbitrary or capricious manner. Brewer v. Copeland, 86 Wn.2d 58, 74, 542 P.2d 445 (1975); Swenson v. Lowe, 5 Wn. App. 186, 191, 486 P.2d 1120 (1971).
The trial court had before it considerable evidence. Several lay witnesses testified whom the child more closely resembled. The trial court had viewed the minor child in chambers. It heard and considered the testimony of two expert witnesses. Although admitting Dr. Hummel would attach no significance to the results in this case because certain statistical safeguards had not been met, the experts used the formula anyway.
From the trial judge's view, to say this was a very difficult case to decide is an understatement. Although the two experts offered uncontradicted opinions based upon scientific data, the court chose not to accept their opinions because the plausibilities of paternity percentages were so close and because the experts attached significance to the *191results of the Hummel formula, when Dr. Hummel would have ignored those very results. The court also did not give any weight to the opinions of the lay witnesses that the child more closely resembled Richard Hendrickson than Dave Vooge, because the experts did not attach significance to the physical characteristics of the parties and because several members of Dave Vooge's family had similar characteristics to the child.
Because the plausibilities of paternity percentages were so close and because of the questionable significance of the results of the Hummel formula, we cannot say the court's refusal to accept the uncontradicted testimony of the two experts was arbitrary or capricious. After duly considering all the evidence in this extremely unusual case, the court found the evidence was too close to choose between the two men and, thus, found the State had not met its burden of proof. It is not our role to substitute judgment for that of the trial court. Retail Clerks Health & Welfare Trust Funds v. Shopland Supermarket, Inc., 96 Wn.2d 939, 943, 640 P.2d 1051 (1982). Therefore, we find no error.
The State next contends the trial court erred in not considering the similarity in physical appearance between Richard Hendrickson and the minor child, Greg. The State directs the court's attention to comments made by the court in its oral decision.
The State's assignment of error is not proper for two reasons. First, an assignment of error to an oral statement of the trial court is not a proper assignment of error. Jones v. National Bank of Commerce, 66 Wn.2d 341, 344, 402 P.2d 673 (1965); Edward L. Eyre & Co. v. Hirsch, 36 Wn.2d 439, 446, 218 P.2d 888 (1950). Second, the trial court, after viewing the minor child and hearing the testimony of several witnesses concerning the physical similarity between the child and Richard Hendrickson, merely expressed its opinion that while such similarity may have been determinative of paternity in the past, it was not helpful in deciding paternity now in light of scientific improvements and the expert testimony in this case. Further, the court stated *192in finding of fact 3: "The Court's viewing the child was not helpful to the Court." The contention that the court did not consider the child's physical characteristics is not well taken; the observation of the child did not resolve the issue.
Last, the State contends the trial court erred by considering irrelevant evidence in the determination of paternity. The State directs our attention to the oral comments made by the trial judge at the conclusion of the case and concludes the judge considered irrelevant evidence. This is not a proper assignment of error. Jones, at 344; Edward L. Eyre & Co., at 446. Also, the comments merely expressed the judge's concern for the child.
The judgment is affirmed.
Thompson, J., concurs.The plausibility percentage is the likelihood that the tested male is the father of the child compared to a random male. Put another way, it is 59 times more likely that a mating of Vickie White and Dave Vooge would result in a child with the type of blood Greg has than if Vickie White mated with a random male. Converted to a percentage, the plausibility is 98.32 percent (59 to 1 = 59/60 = .9832).
Put another way, it is 76 times more likely that a mating of Vickie White and Richard Hendrickson would result in a child with the type of blood Greg has than if Vickie White mated with a random male (76 to 1 = 76/77 = .9870).
As stated earlier, there was neither a difference of more than plus or minus 1 between the logarithm values computed for each male, nor did either male have a plausibility of paternity percentage in excess of 90 percent.