Unless the speakers are positioned in such a way that their response below 300Hz is identical, applying the same EQ correction to two speakers with a different response will neither make the individual nor the combined sound flatter at the listening spot. See below at 64Hz, 87Hz, 100Hz...
According to Earl Geddes you need to EQ a first sub, and then add a second Sub and derive a different EQ for the second sub while the first sub is playing. Then repeat for a third differently EQ’d sub while the first two are playing. How that translates to EQing a pair of full range speakers I don’t know, as you wouldn’t want to do something similar for mid and treble.
With apologies to @flak monkey if you feel this is derailing your thread, but let me have a go at quickly proving what I've said to try to wrap this up.
As a very quick demonstration, here are measurements of each of my main speakers individually (run full range rather than with a subwoofer):
Below is a comparison of what I get if I measure the two speakers simultaneously in black, compared to what I get by calculating the sum of the two individual measurement in red. To achieve this I used the acoustic timing reference option in REW which is good enough that the sum is very accurate for all but the highest frequencies, where frankly there is no point looking at the sum anyway. (@Camverton note that this is a demonstration for an asymmetric setup, as shown by the different responses for the two speakers.)
For the sake of demonstration I then produced an EQ filter based on the calculated sum, targeting a flat response below 200 Hz at 78 dB. Measuring the result of applying this filter to both speakers, with both speakers playing together, I got the result in purple below (with the uncorrected combined response in black for reference).
Finally, here's what the left and right speakers measured individually with the EQ applied look like. Note that the individual responses are far less flat below 200 Hz than the combined result above is.