Composition Test: Answers, Results, and Conclusions

Yesterday I offered a test of the hypothesis that in visual art, a really good composition is so well balanced that should anything be moved, added, or removed, the overall effect will be ruined. You can see the test and comments here.

Answers
The original, untouched Diebenkorn compositions are

Example 1: Left
Example 2: Left
Example 3: Bottom

Results
Today I tallied the responses. A few folks changed their choices, so it was slightly confusing, but I think I got it right now. I only counted someone's answers if they made a clear choice for all three examples. The results of the 15 people who did are below:



As noted in the chart, each correct choice is marked with an X. The one most folks got correct was Example 1. The one most folks guessed incorrectly was Example 2. Three of you got all three correct.

Conclusions
When the guessing first started I was confident my preconceptions would stand. The blogger going by training VS taste ???, who guessed first and immediately got all three correct, offered concrete, even logical reasoning for his/her choices. Such as

regarding second one, much of the drawing is about crossing of line (whether perspectival at bottom and the x or t's at the top. given that, it seems to make more sense that one also crossed the picture plane and by placing the x to the far right, the eyes are forced to cross the picture plane

But even this blogger noted "i don't think there are right or wrong answers here (or most anywhere)," a notion that ending up being supported again and again throughout the remainder of the voting.

Indeed personal taste does seem to be a factor in determining whether a composition is good or not. One blogger chose the incorrect choice for each example, and I know this blogger to be a very talented artist...perhaps one poised to change our general perceptions about compositions. Who knows?

There are a few assumptions/problems with this experiment that make it less than scientific, and most of them were highlighted in the comments. One interesting one that I hadn't considered was pointed out by heather lowe:

This is difficult because the images are so interesting beside one another and they have a definite influence on one another.The pair becomes something entirely unique, in my opinion.

Another assumption that I had considered was noted by auvi

I guess the underlying assumption is that Diebenkorn is better than Edward,
everywhere, all the time. Maybe that's not true.

I should acknowledge that I spent no time considering the composition when I made the changes. Perhaps I have some innate skills, but there was no analysis or attempt to "improve" upon the original. I merely wanted different types of changes (flipping, centering, vertical moves) to see if an analytical conclusion could be drawn. As training VS taste ??? pegged what was wrong analytically with each of my changes (almost frightfully so), I was intially convinced the hypothesis was true.

But again and again in the comments, folks prefaced their choices as a matter of preference, even training VS taste ???, suggesting that whereas some folks prefer logical compositions, others prefer compositions that defy logic and, as JT Kirkland noted, have "bit more excitement to" them.

As I also noted yesterday, perhaps a better test is to add or remove something from a composition (rather than just moving), as that's generally how the hypothesis is framed, although that opens a whole new kettle of fish regarding whether objects and their relationship to each other (i.e, narrative) plays a major role in whether we like a composition or not....but enough for today...the coffee shop awaits. Thanks to all for participating!