Not a very good question, because the answer (at least for SQL Server 2012, SQL server 2014, and SQL Server 2016) depends on whether the index is partitioned or not. Partitioned indexes get statistics calculated on rebuild using the default sampling rate. This is documented clearly in BOL (including, for 2016, on the page referenced in the explanation - and on the v=110 version of that page for 2012).
I'm torn here. If I write "for a table", I'd assume this is a non-partitioned table. That's what the majority of tables are, and for most installations, partitioning isn't possible. I would always expect this to be a default choice. If the table is partitioned, then I'd mentioned "for a partitioned table".
I think writing "non-partitioned table" potentially confuses or gives something away. Would every question need to state partitioned or non-partitioned? It seems strange to me to do this, but I can see your point. If there are partitioned indexes, then that's a change.
I've added that here, but I'm torn on whether we should ask for specific versions, and specific things like non partitioned in descriptors. I'm worried we'll end up with too many specifics that confuse issues.
I suppose we could just try to ensure each question has DDL, but that seems excessive.