The failure mechanism being resolved is a combined issue. First, a poor charging regimen starts accruing sulfate. This leads to a slow loss of capacity which is not usually noticed on an SLI battery. Especially one that is a bit over-sized for the application. The sulfate after some time changes chemical state to the 'hard' crystals and over more time this growth can cause dendrites which become soft shorts. While these dendrites are forming, the increasing sulfation of the battery slows the charge acceptance, further compromising the inadequate charging regime thereby speeding up the sulfation.
When the dendrites achieve a soft short (inter plate) condition the battery will appear to 'lose' charge fairly quickly, even if its residual sulfated capacity is sufficient for the application (SLI). This is usually when the user has to replace the battery.
MrAl If your battery responded to equalization then it is slowly sulfating and the loss of plate area will cause the apparent voltage drop due to internal resistance losses. Sudden voltage drop (overnight) implies a soft short.
If a battery suffers a mechanical or corrosion failure a 'hard short' or 'hard open' can occur. The hard 'short' requires a charge and load test cycle to detect while monitoring voltage, a voltage drop in multiples of 2.2V is a sure sign. The hard 'open' will see the cell gassing under load or a very hot spot (I*I*R heating and gassing) under charge. A load test will see the battery voltage collapse in a couple seconds as the electrolyte bridging the 'open' decomposes under the reverse current and emits hydrogen/oxygen and loses its ability to conduct.
A fracture in the grid or buss bars can create a tiny 'open' condition which requires a a significant load to detect. One that exceeds the electrolytes conduction capacity.
A battery left undercharged for many months will develop a LOT of hard sulfate and this will swell the plates and warp them; occasionally swelling the battery case as well in bad cases. This warping will cause both hard open and hard 'short' conditions to occur as well. Such batteries are well past recovery.
I have recovered 12V batteries with voltages as low as 3V at rest with no additives which brings up another grey area.
On the topic of additives which I have also investigated the 'recovery' mechanism requires an electroplating process. This is a double edged sword.
Early: A sulfated negative plate will have fissures in the sulfated paste which permit some capacity. Electroplating (usually cadmium sulphate additive) these fissures temporarily improves their conductivity and gives better electron access to 'lost' islands of active plate material. This causes an apparent improvement in battery capacity and cranking amps after a few 'full' charge/discharge cycles permit enough electroplating to occur.
Late: After many more SLI battery cycles which are mini charge/discharge cycles the electroplating can start to coat the plates to a point where normal plate function is compromised and the battery is truly dead.
This analysis is supported with the addition of copper sulphate instead of cadmium sulphate to a battery. The copper plating progression (-ve plates) is easily visible and progresses 'downward' from the buss bar into the plates over the charge cycles.
In the case of 'EPSOM' salts or MgSO4 , the Mg will not plate due to its high reactivity. What it does though, is it competes with the lead for sulphate ions (ion swapping) and, being very soluble, interferes with the solid lead sulfation mechanism at the plate surface. Also it increases the inherent conductivity of weak electrolytes. Thus it can help a weak cell IF it is added BEFORE dendriting occurs; how much 'help' is unknown. I don't have long term comparisons for the use of Epsom, but it seems to increase plate corrosion and water loss thus buying only a little time. This is possibly due to its electrolytic nature. Thus adding MgSO4, CdSO4 or CuSO4 et al to a new battery as 'insurance' is not useful at all in the long term.