One ideal against which we could gauge moral progress without it being useless or self-defeating if taken to the extreme would be "Causing less suffering and death is good."

Well, the most straightforward way to judge success along this metric is to compare the amount of suffering. The problem with this metric is that the contribution of technological progress will dominate any contribution from ethical progress.

Might we be causing harm in new ways as well as ceasing to cause harm in other ways, moving backward overall? Even though Americans can't keep slaves, they do get a lot of their goods from sweatshops. The prejudice against gays may be lessening, but has the prejudice against Middle Easterners increased to the point where it cancels out that progress? Women got the right to vote, but shortly before that, children were forced into the school system.

Furthermore, it's not a priori obvious that the contribution to less suffering is what you think it is any of the examples you listed. It's possible that the people working in "sweatshops" are better off there than wherever they were before, this in fact seems likely since they chose to work there. It's possible that our modern attitude towards gender roles and sexuality is causing more unhappy marriages and children growing up in bad homes and thus increases suffering; conversely, maybe our attitudes towards gender are correct and our prejudice towards (Muslim) Middle Easterners is encouraging them to adopt it and thus our prejudice is reducing suffering on net. As for the right to vote, well there's a slight positive effect from making the women feel empowered, but the main effect is who wins elections, and whether they make better or worse decisions, which seems hard to measure.

My point is that doing these types of calculations is much harder than you seem to realize.

Edit: Also, what wedrifid said.