No they won't give better results on every plasma TV. They will give DIFFERENT results. It is not possible to tell if the results will be better than using "normal" patterns until you try both methods then spend time viewing a LOT of content to decide which method produced the best final result. In many cases, you won't be able to decide which result is better because they will each have things that look better and things that look worse. But the results will be DIFFERENT -- a little different. It won't be a dramatic difference.
The problem is this... ABL will ALWAYS be present. You cannot turn it off. Other things that you CAN turn off should be turned off for calibration so you get the best result. An automatic iris is a little bit like ABL EXCEPT you are changing the amount of light AFTER the images are made so the entire image gets darker or brighter and gamma and other parameters really donot change relationships (the numbers may calculate differently, the image appearance will be consistent with an auto iris. With ABL, you are also getting non-linearities in the design of the TV included as brightness is limited in high-luminance images. You cannot stop that, it will always happen. If you calibrate WITHOUT ANY ABL, your calibration will not include any of the non-linearities. Which is why I advocate using standard window patterns where the window covers 10%-12% of the screen area with a black background. That allows for SOME ABL but not all of it. So you will at least be calibrating to eliminate SOME of the errors inherent in the TV. When you completely bypass ABL, you aren't getting ANY compensation within the calibration for those non-linearities. That's why you'll get DIFFERENT results with non-APL patterns and APL patterns. Most plasma TVs calibrate fine with standard 10%-12% windows. The few models that have problems are likely to STILL have problems with constant-APL patterns vs. non-constant-APL patterns. But it will likely take different measurements (than traditional measurements) to see those problems... and those problems may or may not be obvious just from viewing images, much depends on the acuity and training of the person as well as the type and size of the errors.
I would not expect any sort of "OMG!" reaction to calibration with different patterns though, you're likely to be disappointed that the results weren't more obvious.