It's good to experiment.
1. You may indeed be a victim of expectation bias. It is very difficult to eliminate that.
2. Arguments which claim there is no difference electrically are simple "steady state arguments", such as the (very well written) info Leseuf provides. For non steady state signals, that argument is incorrect. (edit: the steady state analysis assumes the stored energy within the crossovers, but transient signal conditions lack that).
3. Technical arguments which claim differences are below JND are flawed in that they do not consider human localization cue parametrics nor our dynamic response to changes with them. (edit: the conclusions may indeed be correct, but the methodology which arrives at that conclusion does not support it)
I cannot claim biwiring is audible, nor can I discount it. But I can recommend a test which has a far better chance of audibility success..as it eliminates the human adaptation of localization parameters that simple DBT testing ignores.
A. Use a mono signal.
B. Adjust the levels, speakers, and listening position such that the image is exactly in the center of the speakers.
C. Modify the connections on ONE speaker.
D. Listen for the focus of the soundstage, with attention paid to the exact location of all sounds. If no difference is found, the modification did nothing within human discrimination capabilities.
One day, others will understand what I speak of..
And then they can explain this goop to me...
Cheers, John
(edit: ps.. It appears that the red conductors are on opposite sides of the cable. If so, then the connection topology you have used allows inductive coupling between the cables. Personally, I'd have assigned the two reds to one section, and the two whites to the other..this of course raises a polarity issue, that has to be correct of course. But my recommendation is based on engineering/physics reasoning and NOT on an audibility argument.)
(gotta have fun, that's what life is all about.)