Are chain-drive gears 'ridiculously' efficient even when dry of oil? Allegedly chain-drive gears are something like 99% efficient when well oiled and something like 98% efficient when without oil. Personally I do not like the noise a dry chain makes, this has a _'I must be going slow'_ psychological effect on my riding so, even if I am not going slow, I think that I am. For this reason I never drive with a dry chain. If the efficiency gains of an oiled chain are minor then that means the main reason for oiling the chain is to reduce wear/prevent rust rather than efficiency/speed. This is why I would like to know the efficiency of a chain when oiled as well as dry. Comparisons against other drive mechanisms and their efficiency, e.g. bicycle shaft drive, automatic car transmission are also of interest to put the oiled/dry chain in context.

> If the efficiency gains of an oiled chain are minor then that means the main reason for oiling the chain is to reduce wear/prevent rust rather than efficiency/speed. This is why I would like to know the efficiency of a chain when oiled as well as dry.

As explained in the IHPVA paper linked in moz's answer, apparently lubricating the chain makes little difference for efficiency, so yes, it's mainly to "reduce wear/prevent rust".

However, you should note that the IHPVA paper apparently used a new chain in mint condition. It may well be that the lubrication does not make a difference in this case. Still, an unlubricated chain will corrode _much_ faster than a lubricated one, and corrosion does make a huge difference in efficiency (as it increases internal friction).

So in practice, lubrication include efficiency, though indirectly (by preventing corrosion).