(concurring) — In virtually every recent death penalty case decided by this court, a different defini*215tion of proportionality has been promulgated. The resulting confusion is unacceptable.
Prior to 1987, proportionality was defined in very general terms. See State v. Mak, 105 Wn.2d 692, 724, 718 P.2d 407, cert. denied, 479 U.S. 995 (1986). Reversal of a death sentence was required only if the sentence had been " 'wantonly and freakishly imposed”, or if taken together similar cases did not result in a death penalty verdict. State v. Rupe, 108 Wn.2d 734, 767, 743 P.2d 210 (1987) (quoting State v. Harris, 106 Wn.2d 784, 798, 725 P.2d 975 (1986), cert. denied, 480 U.S. 940 (1987)), cert. denied, 486 U.S. 1061 (1988).
However, in In re Jeffries, 114 Wn.2d 485, 489-91, 798 P.2d 731 (1990), we began to engage in a more detailed examination of the proportionality principle. We cautioned against second-guessing prosecutors’ charging decisions and juries’ verdicts recognizing that, while two cases may appear similar at first glance, different outcomes may result from persuasive mitigating circumstances. Jeffries, at 490.
Shortly thereafter, seven members of this court (Dore, C.J., dissented on other grounds) adopted an analytical framework for proportionality in State v. Lord, 117 Wn.2d 829, 822 P.2d 177 (1991), cert. denied, 506 U.S. 856, 121 L. Ed. 2d 112, 113 S. Ct. 164 (1992). In Lord we recognized that the primary purpose behind a proportionality analysis was to eradicate a historical and systemic problem in the application of the death penalty throughout the country; namely, that the death penalty often was imposed with regards to the race of perpetrators and victims, rather than the facts of the crime. Lord, at 910. Further, we acknowledged that it was impossible to expect the facts of every case resulting in a death penalty sentence to somehow match up. Instead, we determined that it was necessary to search for a family of resemblances. In other words, rather than requiring a death penalty case to have "one characteristic or set of attributes in common” with other cases, we insisted on a "network of overlapping similarities”. Lord, at 911.
This was followed by State v. Dodd, 120 Wn.2d 1, 838 P.2d 86 (1992), which agreed generally with the Lord approach. *216However, less than 2 years later, State v. Benn, 120 Wn.2d 631, 845 P.2d 289, cert. denied, 510 U.S. 944, 126 L. Ed. 2d 331, 114 S. Ct. 382 (1993) suggested a statistically based version of proportionality.
Now, the majority suggests yet another version of proportionality. See majority, at 207-214. It is not so much that I disagree with the majority’s analysis as with the revisions in each new case.
It may be that proportionality itself depends on the consistency of a proportionality analysis. If applying a varied proportionality analysis with each case produces varying results, we thwart RCW 10.95.130(2)(b) which requires this court to determine "[w]hether the sentence of death is excessive or disproportionate to the penalty imposed in similar cases”, and ultimately we may commit a tragic disservice to the victims of these crimes. Moreover, the resulting confusion contributes to an already excessive delay in these cases.
I recommend adopting the proportionality test established in Lord because it garnered the most support and provides the most reasoned analysis.
Guy, J., and Andersen and Brachtenbach, JJ. Pro Tern., concur with Durham, C. J.