Sorry in advance for the long post, I got a little carried away.
As detailed, there was no statement that said replacements will begin on the 23rd. People took that as the day and ran with it (obviously including people here). One week is not a long time in these situations for a company, so yes, I'm sure there has been much evolution in the situation. As their focus is on remediation of the situation for the best of all people affected (themselves and their customers), I would certainly hope they are exploring all possibilities to keep those balls on the market, especially considering the questionable circumstances surrounding the sudden ban and the imperfect process for determining the measurement (I'll get to that in a minute).
Maybe I'm a bit more empathetic since I work in Regulatory Affairs in a regulated industry. When a bomb like this goes off with zero advanced warning, there is a ton of information gathering going on. What you think you know one day could be very different the next. In the rush to get information out to customers on the day of the event, the main focus has to be on easing customer concerns and letting them know they WILL be supported. As days go by and more and more information is gathered and analyzed, A story begins to build and strategies are altered. When the customer base is expecting an update, it is provided based on the current level of information. And that is what we saw.
I will conjecture that the core (pardon the pun) of the issue is tolerances and sensitivity of both the manufacturing process and the certification process. It is IMPOSSIBLE for every manufactured ball to have the EXACT specification (RG and diff) that is listed on the box. This is a truism for ANY manufactured product. The drugs you take are manufactured products, but the potency of the active pharmaceutical ingredient (API) is not exactly the same across every batch. Typically there is a tolerance range allowable in order for a batch to pass. This may be a range of 95%-105% potency. If the tested batch falls within this range it passes.
Similarly, I have to believe that there is a tolerance range for the manufacture of bowling balls. In EVERY manufacturing run there will be balls that are slightly over listed spec and slightly under listed spec, and of course some that are on spec. Since the USBC allows for a ball to be listed (qualified) at the maximum differential, then it must allow for a tolerance range based on manufacturing variance that will still consider a ball within spec. For the record, I don't know if that is true, but it would be idiotic for it not to be considering the nature of the manufacturing and MEASURING process. I'm just spit balling here, but let's say that purity of bowling balls is far more important than the purity of the medicine you take. So the tolerance range is 98-102% (not unreasonable). In that instance any ball of the line that had a measured differential from 0.0588 - 0.0612 would fall within spec. And remember that is plus or minus 12 / 10,000ths of an inch. For context, the thickness of a sheet of printer paper is 0.0038; more than three times higher.
Why should that be acceptable? Well lets take a look at the measurement process. The full procedure is provided here:
http://www.bowl.com/uploadedFiles/Equipment_Specs/Information/SOP-BALL-1_AsymmRG.pdfThe PSA is HAND DRAWN with a grease pencil with a mark that looks to be about a half inch in diameter (that's 5000/10,0000ths for those playing at home). The line HAND DRAWN with a grease pencil (a line that I would say is 1/8 inch wide, or roughly 1250/10,000ths) to mark the X axis point is measured with a pro-sect at 6 3/4" (or 6 and 7500/10,000ths). Same with the Z.
Then the ball is HAND PLACED into the swing as described in step 13 pasted below (Parenthetical remarks are mine):
Place the bowling ball in the RG Swing cradle with the x-axis at the very top pointing up through the string (how many 10,000ths is the string?) on the RG Swing. To assist with centering the bowling ball, align the line (that is 1250/10000ths wide) from the x-axis to the PSA with the side bar of the cradle on the RG swing.
This is done twice on each axis and the swing is set in motion by hand. There is zero chance that this method can be determined to have a validated sensitivity to the 10,000ths of an inch. You couple this insensitivity with natural variance and a tolerance range of 98-102% as described above is extremely reasonable and defensible. The human variance alone of the measuring process lends serious doubt to the voracity of USBCs claim of a sufficient number of balls out of spec.