I'm not sure what the theory would be on a bad connection. That should affect all cells. The same electrons have to go though the whole series chain. There is a one in two chance of it being that way just by chance (assuming two failed batteries).
You said you equalized them, and they were AGM. Most AGM manufacturers don't allow equalization (Lifeline is the exception, but even they have unusual requirements and say don't do it unless you really need to). I'm not familiar with USBatteries, do they recommend equalization?
I don't envy moving the L16. I have enough trouble with 6GC type, but probably have bought my last LA batteries a couple of years ago - next ones will be LiFePo4.
Generally I agree that if there was a high resistance connection it should affect all of the cells in a common 12 V battery.
I'm not so sure that when a 12 V battery is constructed from 2, 6 V batteries that the external series connection (which happens in a 12 V battery inside the box) does not play a role in the early demise of one component used in this constructed 12 V battery.
That's why I asked if anybody knows?
I also am at odds, with your odds.
When you have 4 individual batteries and 2 fail, the possible battery combinations that failed include, Battery #1 & #2, or 1 & 3, or 1 & 4, or 2 & 3, or 2 & 4, and finally 3 & 4.
Randomly, thats 1 in 6 odds of picking any 2 particular batteries, not one in two.
My understanding is that US Battery does not make the AGM batteries that they market, label and sell as their own. Whatever equalization recommendation, if any, should ultimately come from the actual manufacturer, who may well have changed over time.