Yes, we are going into the "it's possible" scenarios. That's what safety is all about. The rigorous safety analysis starts with a determination of the consequences of an event, then progresses to the calculate the likelihood of the event occurring. Safety, or 'safe enough' depends on the combination of the severity of the consequence and the likelihood of occurrence. The most serious consequences require the lowest probability of occurrence. And any change in the sequence of events that lowers that probability is deemed necessary.
In this case, the ultimate consequence of an unintended firearm discharge is death. Nothing is more serious than that. We must take every opportunity to reduce the likelihood of an inadvertent discharge.
So, we carry in a holster, a good, strong holster that covers and contains the gun so that, if it doesn't totally eliminate snagging the trigger and hammer it at least makes it very, very unlikely. Good.
And when the gun is out of the holster we're very diligent about not pointing it anywhere that someone could be hurt if it went off. Good.
But how about the transition? How about, if we are holstering the gun for the 10,000th time, a very familiar action that has never, ever caused a problem before, we lose focus or are distracted somehow and miss the holster, or catch the hammer on our belt or pocket? Yep, could happen. Has happened.
Now, let's look at the two scenarios: 6 loaded and the hammer down on a between-chamber safety notch or pin, and 5 loaded with the hammer down on an empty chamber. Many seem to think, and I take it from your posting that you are one of them, that these two conditions are equivalent from the safety standpoint. Unfortunately, that's not true.
6 loaded, hammer on between-chamber safety notch/pin: In this case the bolt is resting on the surface of the cylinder, between stop notches. The only thing keeping the cylinder from rotating is the hammer in the notch/on the pin. It is a positive, firm lock, but it's a single condition. If the hammer is pulled back a very few degrees it clears the notch/pin and the cylinder CAN rotate. It's not being forced to rotate by the hand, as the hammer has not moved enough, but there is nothing preventing movement caused by some outside influence, like rubbing on the holster or belt loop or some other article of clothing. From the rest, between-chambers, position the cylinder only needs to rotate 30 degrees to come to a loaded chamber, and at that point the bolt will drop into the cylinder stop notch and the cylinder will stop, in battery. There is little likelihood that the hammer will reach the half cock position as it did not move very far before the cylinder reached the battery position. All that's needed now is for the hammer to drop on the loaded chamber when the snag clears, and since it is under mainspring tension it will surely do so. Bang.
5 loaded, hammer down on empty chamber: In this case the cylinder is prevented from rotating by two things: the bolt resting in a cylinder stop notch and the hammer on the nipple. In order for the cylinder to rotate here the hammer has to move much farther back than in the between-chamber case; it needs to move far enough for the hammer cam to retract the bolt from the stop notch. In addition, the cylinder must rotate twice as far - 60 degrees - before it comes to the next chamber and the bolt drops into the next stop notch, leaving a loaded chamber in battery. The hammer, under mainspring tension, may fall on the loaded chamber, but it's also much closer to the half cock position than it is in the between-chamber scenario, so the possibility that it will enter that position is higher.
I'm not trying to argue that one condition is safe and the other unsafe; what I'm saying is that the 5 loaded, hammer on an empty chamber case is, by the physical design of the action, measurably SAFER than the 6 loaded, hammer down on between-chamber safety notch/pin. And the SAFER status, given the dire consequences is a better choice.
Is it 10% safer, or 1%, or even 0.1%? I don't know. But even at 0.1%, it's better. Is it possible to be safer by somehow eliminating any possibility of snagging the hammer? Sure it's possible. But, how sure can you be that EVERY snag possibility is TOTALLY eliminated? Seems to me it's easier to just make the safer choice in carry method.
And by the way, the "it's never happened to me or anyone I know in millions of hours of carrying a gun" argument is an illogical fallacy: the post hoc ergo proctor hoc fallacy.