Frank Krygowski wrote:

This has a couple possible implications.
Here's one: sprockets (especially chainrings
on less expensive bikes) are seldom perfectly
round. Usually you can notice the change in
tension (i.e. change in visible slack) as you
turn the cranks with the bike on a stand.
I suppose it's possible that the first
estimate of chain "tightness" happened with
the chain on a high spot, the next one on
a low spot.

That's true! What spot should you choose, then?

I always put the crank vertically, the one
where I can pull the most.

I haven't thought of the possibility of
overstretching at the other end. But you have
to pull really tight for that to happen, right?

--
underground experts united
http://user.it.uu.se/~embe8573