It's hard to say, though I think the UPSERT capability looks useful because it simplifies handling duplicates. On the other hand it does not appear that Ares offers clustering, which is critical for large datasets.
(I work on ClickHouse and enjoyed this article when it came out.)
Thank you for the comment. I will take a look then.
It should be mentioned, to ClickHouse' credit, that they made an effort to publish relatively detailed benchmark results for a some data sets and queries, when they first came out. They even got in contact with my research group at the time (the MonetDB group at CWI) to make an effort to present the MonetDB results in a fair manner.
What databases are comparable to ClickHouse say for
Ease of single node deployment
Super fast basic analytics, filter/group/count without a lot of optimization
Fabulous compression