That doesn't look like the right target curve for your measurement. There is no sign of roll-off in the original measurement within the 100Hz span shown, does your sub measurement include a crossover in the path? If it doesn't you shouldn't use a target that has a crossover filter applied. You also seem to have some 'house curve' applied to the target, what settings did you use for that? To tell whether the response is better overall it needs to be seen over a wider span and with a main speaker active so that the match to the main speaker can be seen.