The plaintiff, Boston Gas Company (company or utility), doing business as Keyspan Energy Delivery New England, appeals from a reinstated decision of the Appellate Tax Board (board), upholding the fiscal year 2004 assessment by the board of assessors of Boston (assessors), for the company’s *518rate-regulated utility property. The reinstated decision followed a partial remand by the Supreme Judicial Court in Boston Gas Co. v. Assessors of Boston, 458 Mass. 715 (2011) (Boston Gas I), ordering the board to consider further certain aspects of its analysis in valuing the company’s personal property.1 In this appeal, the company argues that the board’s findings of fact and report on remand were not supported by substantial evidence and were erroneous under applicable law.

For background and the issues on remand, we refer to the Supreme Judicial Court’s opinion in Boston Gas I, supra. Briefly, the board had reached an estimate of the fair cash value of the company’s personal property by according equal weight to the net book value of the property and the property’s reproduction cost new less depreciation (RCNLD), a methodology upheld in this case by the Supreme Judicial Court. See Boston Gas I, supra at 729 (“In sum, we conclude that the board relied on sufficient evidence in determining that special circumstances warranted the use of a valuation method other than net book value”). As part of the RCNLD determination, the board relied on the income capitalization approach as a market reference, in order to estimate the economic obsolescence associated with the property, particularly the effect of governmental regulation on value. It was within the board’s discretion to utilize the income capitalization approach for that limited purpose. See, e.g., Boston Edison Co. v. Assessors of Boston, 402 Mass. 1, 17 (1988).

The problem that concerned the court, however, was that, because the property’s RCNLD exceeded the value reached through the income capitalization approach, the board’s methodology called for the RCNLD amount to be reduced to the income capitalization value. Thus, in estimating economic obsolescence by reference to the income capitalization approach, the board essentially adopted that value as the property’s RCNLD. “Given the importance of the income capitalization approach to the board’s final valuation, we conclude that the income approach must itself be sound.” Boston Gas I, 458 Mass. at 731.

The court identified three aspects of the income capitalization approach used by the board that required further consideration. *519First, the court remanded to the board for clarification of its decision not to use a tax factor in the income capitalization approach. Id. at 734, 740. Second, with respect to the board’s analysis of the property’s earnings before interest, taxes, depreciation, and amortization (EBITDA), the court sought clarification of the board’s exclusion of calendar year 2001 from the sample used to determine the property’s average EBITDA. Id. at 732, 740. Last, the court also sought clarification regarding the board’s calculation of the EBITDA multiplier. Id. at 733-734, 740. We are persuaded that, as to each of those issues, the board’s findings of fact in its reinstated decision were supported by substantial evidence. Accordingly, we affirm.

1. Tax factor. The board arrived at a valuation of the property from the income capitalization approach by averaging the revenues of the company for five years between 1997 and 2003.2 In so doing, the board subtracted the property’s operating expenses from the revenues for each of the five years to arrive at the annual EBITDA figure. Included in the operating expenses were the property taxes actually incurred by the company. Noting the board’s preference in prior decisions for the use of a tax factor, rather than the amount of property taxes actually paid — based, as they are, on the disputed assessment —• the court remanded, for further consideration, the board’s decision not to use a tax factor in its calculation under the income capitalization approach. See Boston Gas I, 458 Mass. at 734-735 (“We have recognized that, in using a tax factor rather than the tax expense actually incurred, one avoids including the very tax assessment in dispute in the valuation of the property for the purpose of resolving that dispute”).3

*520“A reviewing court must set aside a finding of the board if ‘the evidence points to no felt or appreciable probability of the conclusion or points to an overwhelming probability of the contrary.’ ” Irving Saunders Trust v. Assessors of Boston, 26 Mass. App. Ct. 838, 841 (1989), quoting from New Boston Garden Corp. v. Assessors of Boston, 383 Mass. 456, 466 (1981). We are satisfied that the board, on remand, adequately explained and supported its decision not to use a tax factor in its income capitalization approach.

Among its stated reasons on remand, the board specifically found that, as a rate-regulated utility, the company was entitled to charge rates that included reimbursement of its property taxes. See Boston Gas I, 458 Mass. at 718 (“The [Department of Public Utilities (DPU)] allows a utility to recover, through the rates charged to consumers, its reasonable operating expenses, taxes, depreciation and amortization, and other costs”), citing Boston Gas Co. v. Department of Telecommunications & Energy, 436 Mass. 233, 234 (2002).4 That finding was supported in the record, as the evidence plainly established that the rates were set by the DPU to allow for the company’s operating expenses, including property taxes, to be recovered from the ratepayers. The board also referred to the fact, also supported in the record, that the utility’s earnings were highly regulated, and the evidence showed that rates were set by the DPU to permit the utility owner to earn a reasonable return on investment. Taken together, we read those findings as explaining the board’s decision to deduct the property taxes actually paid by the company as an expense in the board’s income capitalization approach, as those amounts were recovered from the ratepayers rather than from the utility owner’s profits.

It was appropriate for the board, in valuing the property of the company, a rate-regulated utility, to consider evidence of the regulatory features of the property in deciding to use the actual property taxes as an expense, rather than a tax factor.5 As the record made clear, a significant regulatory feature in valuing the *521company property was rate setting as a means of recovering expenses and ensuring a reasonable rate of return on investment. “[Gjovemmental restrictions on the financial returns of a utility company are . . . relevant to the price which a willing buyer would pay to a willing seller for utility property.” Montaup Elec. Co. v. Assessors of Whitman, 390 Mass. 847, 850-851 (1984). If property is subject to “a governmentally-imposed restriction affecting its value or its earning power, that fact should be considered in any determination of its fair cash value.” Boston Edison Co. v. Assessors of Watertown, 387 Mass. 298, 304 (1982), S.C., 393 Mass. 511 (1984).

Mindful of the effect of governmental regulation on the property’s capacity to generate income, the board properly treated the property taxes actually incurred by the company as an operating expense, based on substantial evidence that recovery of those amounts was provided through the rates set by the DPU. “Substantial evidence is ‘such evidence as a reasonable mind might accept as adequate to support a conclusion,’ taking ‘into account whatever in the record fairly detracts from its weight.’ ” Assessors of Brookline v. Buehler, 396 Mass. 520, 524 (1986), quoting from New Boston Garden Corp. v. Assessors of Boston, 383 Mass. at 466. Here, the evidence adequately supported the board’s finding that, unlike the typical commercial property, the company’s property taxes were among the expenses recovered from the ratepayers and, therefore, did not affect the rate of return on the owner’s investment.6 See, e.g., Boston Edison Co. v. Assessors of Watertown, 387 Mass. at 304 (a potential buyer of a public utility *522would be influenced by the rate of return specified by the DPU and its effect on the buyer’s investment).

The board’s expertise in this regard is entitled to “some deference,” Koch v. Commissioner of Rev., 416 Mass. 540, 555 (1993), which we extend to “the board’s judgment concerning the feasibility and fairness of alternate proposed methods of valuation.” Massachusetts Inst. of Technology v. Assessors of Cambridge, 422 Mass. 447, 452 (1996). The board determined that the usual dilemma of estimating a commercial property’s income utilizing the property taxes actually incurred, which are based on the very assessment in dispute, was not present here. We conclude that the board adequately clarified its decision to utilize the property taxes actually paid, rather than a tax factor, to account for the effect of rate regulation on the property’s ability to generate income when estimating the property’s value to a potential purchaser.

Contrary to the company’s assertion, the cases upon which it relies do not mandate that the board use a tax factor in valuing all manner of commercial property.7 Those cases, in which the use of a tax factor was favored in the income capitalization approach, involved private income-producing properties rather than rate-regulated utility properties. See, e.g., Assessors of Lynnfield v. New England Oyster House, Inc., 362 Mass. 696, 700 n.2 (1972), citing New Brunswick v. New Jersey Div. of Tax Appeals, 39 N.J. 537, 546 (1963); Assessors of Lynn v. Shop-Lease Co., 364 Mass. 569, 573 (1974). Indeed, “the value of property for rate-making purposes, related, as it is, to assuring provision for an adequate return on a utility’s investment, may have little to do with what the property would sell for on a free and open market.” Boston Edison Co. v. Assessors of Watertown, 387 Mass. at 303.8

The assessors’ analogy to valuation of net lease properties is *523apt. See, e.g., General Elec. Co. v. Assessors of Lynn, 393 Mass. 591, 609-610 (1984) (tax factor not necessary where the tenant pays the property taxes and the landlord’s income is not reduced thereby); Pepsi-Cola Bottling Co. v. Assessors of Boston, 397 Mass. 447, 453 (1986) (affirming board’s decision not to use a tax factor where tenant was responsible for the payment of all taxes). In the net lease situation, the property taxes are paid by the tenant, while in the case of regulated utilities such as the company, property taxes are recovered from the ratepayers through rate setting. In both instances, use of the amount of the actual property taxes, rather than a tax factor, is appropriate in the income capitalization approach because the actual amount of the taxes incurred does not affect the owner’s rate of return on investment.

On appeal, the company also attempts to undercut the board’s finding that the utility’s property taxes were a recovered expense by focusing on the lapse of time between rate-setting proceedings, claiming that a significant increase in property taxes in the interim would not be factored into the rates. According to that argument, the increased property taxes would not be recovered from the ratepayers but, instead, would be borne by the company, thereby diminishing its profits. On that basis, the company maintains that the use of a tax factor was necessary to estimate the effect of such increases on the property’s income-generating capacity.

The assessors counter that an increase in property taxes between rate-setting proceedings likely would be covered by the annual inflation increase included in the rates.9 See generally Boston Gas Co. v. Department of Telecommunications & Energy, *524436 Mass. at 235. The assessors also point out that a significant or unanticipated increase in property taxes between rate-setting proceedings would entitle the company to seek recovery under the procedure available for exogenous costs.10 On this record, the company has failed to persuade us that property tax increases between rate-setting proceedings would not be recovered from the ratepayers pursuant to those safeguards.

Based on the foregoing, we conclude that the board’s findings of fact on remand concerning the company’s recovery of its property taxes through the rates set by the DPU were supported by substantial evidence and, thus, justified the use of the property taxes actually incurred, rather than a tax factor in the board’s income capitalization approach.

The company conceded at oral argument that, unless it prevailed on the tax factor issue, resolution of the remaining issues in its favor would not result in a valuation below the property’s 2004 assessed value, and so we touch on them only briefly.

2. Exclusion of 2001 EBITDA. The board excluded the 2000 and 2001 revenues from the seven-year sample of the company’s annual EBITDA figures, which were used to arrive at an estimate of the annual EBITDA attributable to the property for purposes of selecting the EBITDA multiplier. The exclusion of the 2000 EBITDA figure was not disputed. However, the Supreme Judicial Court sought clarification for the board’s decision to exclude the 2001 EBITDA figure from the seven-year sample. The board’s original rationale, that 2001 was eliminated on ac*525count of abnormal amounts of deferred income tax and amortization expenses taken that year, was deemed insufficient by the court, as those expenses would not be a factor in EBITDA. The court also rejected the rationale that a Federal tax issue discussed by the assessors’ expert provided a basis for excluding the 2001 EBITDA. See Boston Gas I, supra at 731-732.

On remand, the board explained that the anomalies it cited in its original decision were indicative of other concerns that rendered the 2001 EBITDA unreliable for inclusion in the sample. “In particular, the Board was influenced by 2001’s atypical expense ratio, its substantially negative sum relating to income taxes, and perhaps most significantly, the fact that the average EBITDA as a percentage of total operating revenue for the years presented was more than 50% higher than the percentage for 2001.” The board found that the atypical figures appeared to be connected with Keyspan’s acquisition of Eastern Enterprises in 2000. On that point, the assessors’ expert, George Sansoucy, testified that the purchase was reflected in the company’s 2001 EBITDA, specifically, the calculation of earnings as shown in the records of the company and Keyspan for 2001.11

The company maintains that the link between the anomalies in the 2001 EBITDA and the acquisition of Eastern Enterprises was not supported in the record. We view Sansoucy’s testimony as sufficient to support a reasonable inference connecting the two. In so doing, we defer to the board to decide Sansoucy’s credibility, the weight to be given his testimony, and inferences to be drawn therefrom. See generally Boston Gas I, supra at 738, citing Fisher Sch. v. Assessors of Boston, 325 Mass. 529, 534 (1950). According the board our usual deference, we are satisfied that the board’s findings of fact regarding the 2001 *526EBITDA were based on substantial evidence and adequately accounted for the elimination of 2001 from the sample.

3. EBITDA multiplier. In its original report, the board adopted Sancoucy’s EBITDA multiplier of 11.7 for use in its income capitalization approach. The EBITDA multiplier was reached by looking at six comparable sales of regulated utility property, and calculating the ratio of sales price to annual EBITDA for each. The Supreme Judicial Court described the board’s selection of the EBITDA multiplier as “the approximate average” of the ratio of sales price to annual EBITDA for those properties. Boston Gas I, 458 Mass. at 731.

One of the six sales relied upon was Keyspan’s acquisition of Eastern Enterprises in 2000. At the time of the acquisition, Colonial Gas was a subsidiary of Eastern Enterprises, having become a subsidiary in August, 1999. As described in Boston Gas I, the company maintained in its first appeal that the board erred in “calculating the sales price to EBITDA ratio for that sale by using a sale price from 2000, which included the amount paid for Colonial Gas, while using an EBITDA from the end of 1998, which did not include Colonial Gas’s contribution to EBITDA.” Id. at 733. The court remanded the issue for the board’s further consideration.

On remand, the board agreed that Sansoucy’s chosen EBITDA multiplier failed to account for the contribution to earnings of Colonial Gas. To that end, based on the corrected ratio of Eastern Enterprise’s sale price to annual EBITDA, the board took the average of the six ratios to derive a corrected EBITDA multiplier, thus reducing the EBITDA multiplier from 11.7 to 11.57.12

On appeal, the company now objects that the new EBITDA multiplier of 11.57 was not reached by the same method used by the board in selecting the original multiplier and that, pursuant to the remand order, the board was required to use the same method. According to the company, the board reached the original multiplier of 11.7 not by averaging the six ratios, but by select*527ing the figure at roughly the midpoint between the Eastern Enterprises and Colonial Gas multipliers.13 As noted, however, the Supreme Judicial Court described the board’s selection of the original EBITDA multiplier as the approximate average of the ratio of sale price to EBITDA in the six sales, and found no fault with the method itself. Because the board selected the revised EBITDA multiplier using the method implicitly approved by the court, we see no need for further remand.14

Reinstated decision of the Appellate Tax Board affirmed.