note eyepopslikeamosquito I just noticed a very similar Roman numeral golf recently concluded at <a href="http://golf.shinh.org/">golf.shinh.org</a>. Adjusting ySas' winning Perl solution to that game for this one produces this 58 stroker: <CODE> s/(M)|D/4x@-x5/eg,s/C4//,$\=y/LXVIC4/DCLXM/for(<>)x3;print </CODE> which has a similar form to [robin]'s early-lead 60 stroker above -- though it uses a different algorithm. Both these solutions are very beautiful, very Perlish, and quite astonishing (at least to me). I especially enjoyed ySas' ingenious use of <C>@-</C> above. I've never seen <C>@-</C> used like that in golf before and am tempted to give it a name, "ySas' device". By way of explanation, note that adding parentheses around <C>(M)</C> above adds one more element to the <C>@-</C> array when the (parenthesized) <C>M</C> matches. That is, the number of items in <C>@-</C> is used to differentiate between a matching <C>M</C> (two items in <C>@-</C>) and a matching <C>D</C> (one item in <C>@-</C>); this fits like a glove here because one and two happen to be the required multipliers. Update: By applying ideas from ySas' solution, we can reduce Robin's original 60-stroker: <CODE> y/DCLXVIM /4 DCLX9/,$\=s/( )?\d/9x($&+!$1)/egfor(<>)x4;print </CODE> to 58: <CODE> s/M2//,y/DCLXVIM/1MDCLX2/,$\=s/\d/2x$&x5/egfor(<>)x4;print </CODE> 759963 813652