There is a reason NHRA Pro Stock engine builders use iron blocks in place of aluminum. The weight savings is not worth the horsepower loss because of the loss of ring seal and block squirm. An iron block will always make more power than an aluminum block. If weight is a huge concern then you have to weigh your options.
This. There are a few racing venues where weight trumps power and the weight savings is justifiable but there will always be a power sacrifice.
Also, sleeves "dropping" isn't the issue...the problem usually comes from when the sleeves are decked even with the block, the block can expand more than the sleeves leaving them "in the hole". The other scenario is the sleeves don't get installed to their full depth and will "settle" after the engine is run. Both scenarios are very common with aluminum blocks. It's usually recommended to leave the sleeve up a couple thou on a new build, but if you have to deck the block for some reason, it's usually not real practical to remove sleeves before you do it, re-install sleeves, then deck them to a different height. At that point, the engine has run, the sleeves are settled, and zero decking them usually isn't an issue.
Which is what happened to one of my builds.
We take the new blocks out of the box, clamp a head gasket and torque plate to each side, and put them in an oven to get the sleeves to take a set, to try to eliminate that "settling" that can happen after the engine is ran. The sleeves are still usually proud a couple thou and match up perfectly with the fire ring in the head gasket, but the issue is that the block's decks aren't always perfectly square/straight/parallel with the crank, so the decks need cut.
So this particular block had the sleeves set, decks cut, and "dropped" a sleeve on the dyno, just far enough to necessitate pulling the engine down again. On an aluminum block, the deck height actually grows, so if you have a sleeve drop and then the block grows up around it even more, it's not a good scenario.
I'm not convinced that ring seal isn't always compromised, especially with a different running temperature/condition than when the block was machined. I've pulled fresh all-aluminum FE's off the dyno, and have yanked the heads to see what I would call a "shadow" in some of the cylinders, almost like the block had never been torque plate honed. I could see how these engines would automatically be down for a little oil usage right off the bat.
Complicating all of this is the fact that there are no aluminum FE block manufacturers who can push out enough quantity to justify some real hard core durability testing, like the OEM has.....or even like Dart/World would have. That's why we see porosity/flaws/etc. on larger sample sizes of blocks.
A CGI block would be where it's at, if you could take advantage of the weight reduction capabilities.