In my efforts to establish a performance baseline I ran another reference match against a strong engine. This time I picked gaviota and decided to run a few more games at a slightly longer time control to get a more accurate result.
600 games at a TC of 100 moves in 10 sec + 0.5 sec increment per move, ponder off, own book
Score of gaviota v0.85 vs iCE 0.3 v2394 : 392 - 111 - 97 [0.73] 600
ELO difference: 176
which seems about petty correct. So this is the level I want to improve my engine from.