Some reloaders are very concerned about bearing surface length. I don't worry about it too much for a number of reasons. For example, every time we do a seating depth test, the amount of bullet bearing surface gripped by the neck changes by whatever seating depth [length] increment is used in the test, even if the bullets used were all essentially dimensionally identical. Nonetheless, we use the results without concern that a few thousandths' difference of bearing surface in the neck completely undermined the test results. Likewise, when we seat bullets with varying bearing surface length to a uniform seating depth, the amount of bullet shank seated in the neck also changes. If one is involved in a discipline where a few thousandths difference in bearing surface/neck wall contact can make a difference that they can actually shoot, then they probably need to be concerned about it. I freely admit that I cannot shoot the difference of a merely a few thousandths' bearing surface/neck wall contact, so I don't overly concern myself with trying to length-sort bullets by BTO. If one shoots in a discipline where it does matter, have at it, but I would point out that the OP is talking about factory second bullets here. There is a limit to how far I would go with sorting factory seconds, just because.
On the other hand, I point bullets, so length-sorting bullets by OAL is one way to facilitate obtaining more uniform points. In my hands, the bullets I use (mostly Berger) seem to have pretty uniform BTO dimensions, with most of the OAL variance residing in the nose region of the bullet. For that reason, sorting these bullets by OAL is more or less analogous to sorting them by nose length. For that reason, sorting bullets by OAL can improve seating depth consistency, which I believe to be very important. As shown in the cartoon (below), there is a relationship between what I refer to as the "two critical contact points". One of these critical contact points is farther out on the ogive toward the meplat where the seating die stem contacts and pushes the bullet into the case. The other is the point at which the caliper inserts seats on the ogive just above the bearing surface. Variance in the distance between these two critical contact points will usually mean comparable variance in seating depth as measured by cartridge base-to-ogive (CBTO).
Note that both of these critical contact points lie within the nose region of the bullet. What that means is that by sorting bullets BTO will have no effect whatsoever on length variance between the two critical contact points, as they both lie outside the BTO dimension. Thus, sorting bullets by BTO will not improve seating depth consistency. Regardless, folks can sort by bullets or anything else by whatever method they believe works best for them. What I would suggest to the OP would be to sort the factory seconds by whatever the desired method is, then use each bullet sorting group individually when reloading. In other words, sort the bullets however desired, but don't ever mix them back together when reloading and/or shooting. With a sufficient number of sorted bullets, it is usually not too difficult to use a single sorted group for a single string of fire or one match, perhaps even an entire match. One can always switch to another sorting group for additional strings or matches, but I would generally try to choose an adjacent sorting group whenever possible, just to keep variance to a minimum. I would also suggest to the OP that it might be worth testing two different bullet sorting groups, preferably two groups that actually have a reasonable length variance between them (i.e. testing two extremes), side-by-side, so that it can be determined whether the length difference is one that can actually be measured on the target. If there is no major difference on the target, perhaps it is not worth the effort to length sort the bullets.
