What I read, was that for long term storage, the cells were deliberately discharged and kept shorted, there presumably being a difference in the reactions going on inside between that and self discharge. Probably something to do with the chemical properties of the internal self discharge path growing with use. Then again, thinking back, it was NiCads that I was reading about, things could well be different with NiMH. The one cell being less good than its mates is the origin of the memory myth with NiCads.
With batteries, as any other electrical component, there are just the two basic failure modes - it can either go short, or it can become disconnected, allowing that there are degrees on the way to either.
If a pack grows a disconnection, voltage measured by a smart charger will be the full charger voltage but with zero current. Just one dud connection between or within a cell will do that.
For a shorting problem, it is likely to be just one cell that discharges itself all the time, and the voltage measured becomes pack voltage minus 1.2 volts.