Crowding at the Top
I propose the following argument for evaluative systems in the single-winner context. I don't think I have seen this argument before, and I am surprised at how long it took to occur to me.
I assume that the main purpose of voting-system reform in the United States is to put the oligarchy out of power and empower the workers instead, with regard to choosing who can serve in high office in government at both levels.
The choose-one plurality system a. k. a. "First Past the Post" (FPtP) provides special power to monied interests, as explained by Warren D. Smith. In Smith's writing to which I refer, he doesn't explicitly mention the concept of "Prisoner's Dilemma", but let me point out that that is also a name for the mechanism he talks about whereby money wins. The decisions that seem best from each individual's viewpoint, based on a very rational analysis of the consequences of each choice the individual could make, turn out to the disadvantage of the group, as measured by any reasonable method of evaluating overall satisfaction by aggregating over the individual levels of satisfaction. Let me point out that an essential ingredient of this mechanism with FPtP is a power imbalance from voter to voter based on the voters' positions toward the candidates. The motivations affecting each individual voter, as alluded to by Smith, take into account that whoever votes for a "Nader"-type candidate loses voting power. So here we see that imbalance of power between individual voters leads to grosser imbalance of power between groups of voters, those who agree with the monied interests vs. the rest.
Consider a single-winner election having many voters, including a voter we'll call v1, ten candidates who in v1's assessment these ten candidates oppose oligarchy, and finally there are two pro-oligarchy candidates as usual. The ten anti-oligarchy candidates may differ from one another markedly on issues that voters other than v1 consider crucial. Maybe some of these candidates favor the legal prohibition of abortion, and some others oppose such prohibition. Maybe some think that private interests should continue to be allowed to own large chunks of capital equipment and buildings and land and use them as private enterprises to compete in the market, and other candidates think only representatives of the government should control the use of those large chunks of equipment, buildings, and capital as the means of production of the things people need or want. So, under conditions where the voting system encourages voters to act based on mainly their top choices, significant categories of voters are unlikely to support many of the ten anti-oligarchy candidates, because too many of them have offensive positions on abortion or capitalism or guns or whatever. Or maybe some of the candidates would seem unacceptable under those conditions because those candidates refuse to take a clear stance regarding those divisive issues.
Now let us return to the position of v1. v1 feels that the issue of oligarchic control vs. worker control trumps, for the time being, all the other differences among the ten pro-worker candidates. Maybe v1 thinks that abortion must be permitted and she is so angry at those who favor the legal prohibition of abortion that she sometimes fantasizes turning a flamethrower on them. Maybe a second voter, v2, thinks that abortion is murder and it's essential in the long term to prohibit it. But like v1, v2 thinks that for the time being, the fight between worker control and oligarchic control is more important than abortion. Maybe v3 and v4 also agree that the workers-vs-wealthy issue is for the time being the most important thing, but in the long term, they are absolutely opposed to one anthers positions on capitalism vs. socialism. And v5 and v6 are at odds with one another on gun control, but agree that the short-term issue is about popular power and ousting the oligarchs from their special power that is out of proportion to their numbers. Each one knows that his voice on gun control can be heard once the people have political power, and not before.
In a rating election, it is clear how v1, v2, v3, v4, v5, v6 and any other voters who agree with them should vote, in order best to promote their values. They will rate the pro-worker candidates at or near the top, and the pro-oligarchy candidates at the bottom. Now consider v7, who is in favor of one of the pro-oligarchy candidates, and v8, who is in favor of the other pro-oligarchy candidate. They will bullet vote for their respective favorites. In this scenario, if almost all the millions of voters agree with one or another of the voters I have mentioned, and if anti-oligarchy voters (the ones who feel so strongly about that that they are willing to risk electing people who disagree with them on specific issues) outnumber pro-oligarchy voters, the workers shall win the election!! Think what a huge break that would be from the history of the US to date. Think how important such an outcome would be. Finally, a chance for ordinary people to have a voice in the nation's politics.
Now consider by contrast the situation with the voters feeling like this but the election using ranking ballots. First, let's look at voters v7 and v8 and those who agree with their values. Their vote does not have to change. They still can and will bullet vote for their respective favorite oligarchs. The switch from evaluative voting to ranking voting has not robbed any of the power of these voters. But what about v1, v2, v3, v4, v5, v6, and the other voters who agree with them?? How the hell are they going to vote? The system prevents them from putting their ten candidates at the top, where they belong. They are spread across the top ten ranks. Even if the tally is Borda, the average of the ten candidates' scores are substantially depressed compared to where they deserve to be. IRV, with its outsized level of attention on first choices, will do even worse from the worker viewpoint. Basically IRV will put people who agree with each other at odds and split their vote.
Anticipating a Counterargument
Almost every time that I try to point out to IRV advocates an hypothetical case where IRV would fail but evaluative voting, by contrast, would confer democratic representation, they counter that my example is contrived and unrealistic. However, this example is realistic in the US for year 2022. In every State, there will be a Republican or a "Democratic" candidate, or both, for governor and for each open seat in the state house of delegates and senate (there are one or two States with unicameral legislatures, but seats in those will be up for election as well). Remember that most voting law is set at the State level, even for Federal office. If third-party advocates work hard, and there will be more incentive in that year for them to work hard than any time in the last 50 years, there may well also be a Libertarian candidate, a Green candidate, a People's-Party candidate, and maybe a candidate from the Party for Socialism and Liberation or a social-democrat or democratic socialist candidate of some party or other. So we are looking at two candidates representing the oligarchy and maybe five who don't particularly hold with it. Five is not equal to ten, so maybe my example is contrived when I say ten, but still no moral person could find justification for denying equal power, voter for voter, to the workers by depressing their average vote by two and a half ranks.
 By "workers", I mean people who don't own enough capital or land to live on rents and interest alone. This includes some people who can't work or aren't currently working. So, this is not a strictly literal use of "worker".
My argument above has been demolished on Facebook. Maybe a similar argument could be made with a different example, but I now see that with the example I gave, the Alternative Vote (IRV) gives the same electoral outcome as an evaluative system would.