Comments/Ratings for a Single Item

James Zuercher wrote on 2018-12-09 EST

Have the piece values on an 8x10 board been determined for the following variant pieces? Dragon Horse, Dragon King, Crowned Rook, or Gnu. If so what are these values and how were they determined?

In the better late than never spirit, now that there's a bit of a lull, I'll give you my own roughly estimated values calculated just now for such pieces, and how I reached them (noting first that a Crowned Rook is another name for a Dragon King, as is noted in CVP's Piececlopedia):

Dragon Horse (Bishop+Wazir compound piece) I estimate as =B+Wazir+P,

where largely based on the average number of squares a B can reach I guess [single] B roughly=3.75 (note I rate B=3.5 on 8x8, and prefer a B always worth less than 4 pawns),

Wazir = (Guard-P)/2 (note Guard=Ferz+Wazir compound, which I rate as =Ferz+Wazir+P, with Ferz approx.=Wazir); Guard's value on 8x10 computed by my home formula of 32x(max. number squares guard reaches)/(number of cells on board)=32x8/80=3.2 in this case (note this home formula doesn't seem to work too badly for a large range of board sizes, but there are limits). Thus Wazir=(3.2-1)/2=1.1 on 8x10.

Thus Dragon Horse on 8x10= 3.75+1.1+1=5.85 is my estimated value.

For Dragon King (Rook+Ferz compound piece) it's rated by me =R+Ferz+P where I estimate R=5.5 (unchanged from what I give it as on 8x8).

Thus Dragon King on 8x10= 5.5+1.1+1=7.6 is my estimated value.

For Gnu (Knight+Camel compound piece) it's rated by me =N+C+P where Camel is still assumed by me=2 on 8x10 (like it is commonly given by other people, and myself, as for camel on 8x8). I rate a Knight =3.5 on 8x8, but for board sizes other than that I'd apply a home formula which is rather complex, and doesn't work for every possible board size (similar to my Guard formula given earlier). If I've got it written down right, I estimated N=3.5-(0.25x{[no. rows]-8}/2+0.25x{[no. columns]-8}/2)+(Average number of squares N can move to on empty board-5)/8 for a given board size. For 8x10, I've thus worked out an estimated value of 3.38 for a N. Note that, like for the case of a B, on 8x10 a N has a considerable number of juicy squares in terms of influence, if it reaches any of them (which in the case of a N helps compensate somewhat for the greater board size than 8x8, IMHO).

Thus Gnu on 8x10=3.38+2+1=6.38 is my estimated value.