Generally for Lead unless your recharging immediately I've seen cells fail on the fist few cycles.
I'm speaking for the amateur application of operating in the field and coming home to recharge. I'm sure we could come up with a lot of use cases that deep discharges are a poor idea, but an afternoon of operating then recharging the next day, the battery will be fine. Yuasa for one, does provide cyclic life spec's down to 100% DOD and it's more than 100 cycles.http://www.yuasabatteries.com/pdfs/NP_7_12_DataSheet.pdf
most batteries in the last 10% show substantial voltage changes.
Right. It works out that this is a self limiting process. There aren't too many radios out there that will operate much below 11.5V much less 10.5V or lower, so you're almost guaranteed some degree of remaining capacity. In the case of many Lithium batteries, the built in monitor circuits will cut off before damage can occur. My point is that there's little advantage to quit using a battery at some arbitrary point. Run it until either the radio or battery quits.
If you abuse a battery it could fail during the cycle you most want it for.
Back to the typical amateur application - those levels of abuse don't happen here. The operating conditions to cause damage to a battery aren't conducive to successful field operations. You'd have to trying to hurt these batteries on purpose with something other than a radio for a load and a proper charger.
As to parallel batteries... poor option. Most type want to be about the same age and use history or they will self discharged into each other.
Only if left that way unattended for long periods. While actively discharging or discharging parallel batteries work fine together, even combinations of new/used and large/small.
that will just have the better charged one dumping energy in the lesser charged and that is not efficient process.
That doesn't happen. You can't charge a battery with another one of the same voltage and chemistry no matter how disparate the state of charge.
You make a few other good points though about characterizing the radio and knowing the modes and currents that will be used. How much margin one chooses above that to accommodate unanticipated situations or conditions is a separate debate. My primary point is to dispel the myth that you can only discharge a battery to 50% or some other contrived number in the interest of battery life, because in this application it will most likely be age before cycle life is a factor. You're paying for capacity and cycles, and carrying around a bigger/heavier battery than you have to only to take wasted life to the recycler.
My solution is to have half a dozen 7Ah gel cells on float, ready to go. I can grab one and use it stand alone or use up to all six in parallel, and parallel that set to a wet cell battery (SLI or marine) if I need to. Very adaptable and configurable for different scenarios.